Dissertation transfer pricing

A GROUP DECISION SUPPORT SYSTEM
FOR TRANSFER PRICING
IN THE PHARMACEUTICAL INDUSTRY

A GROUP DECISION SUPPORT SYSTEM
FOR TRANSFER PRICING
IN THE PHARMACEUTICAL INDUSTRY

PROEFSCHRIFT

ter verkrijging van de graad van doctor
aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus,
Prof. Drs. P.A. Schenck,
in het openbaar te verdedigen
ten overstaan van een commissie
door het College van Dekanen daartoe aangewezen,
op dinsdag 12 september 1989 te 16.00 uur

door

MICHAEL BARTHOLOMEUS MARIA VAN DER VEN

geboren te Goirle

TR diss
1749

Dit proefschrift is goedgekeurd door de promotor:
Prof. Dr. H.G. Sol

Promotiecommissie

Prof. Dr. H.G. Sol, Technische Universiteit Delft, promotor
Prof. Dr. A. Bosman, Rijksuniversiteit Groningen
Prof. Drs. B.K. Brussaard, Technische Universiteit Delft
Prof. Dr. H. Koppelaar, Technische Universiteit Delft
Prof. Dr. F.A. Lootsma, Technische Universiteit Delft
Prof. Dr. Ir. CA.Th. Takkenberg, Rijksuniversiteit Utrecht
Dr. H.C. Verlage, Raad van Bestuur Grasso's Koninklijke Machinefabrieken NV

"Transfer pricing in pharmaceuticals
is more of an art than a science"

Otto Nowotny

The writing of this thesis is in gratitude
dedicated to my wife Eugenie, my parents
and my parents-in-law.

Published and distributed by:

Michael van der Ven
p.o. Mosheuvel
Molenbochtstraat 76
5014 ER Tilburg
The Netherlands
Tel. +31(13)365309 or +31(13)359429

Cover:
Piet van Zwieten

Print:
Gianotten BV, Tilburg

The following names are registered trade marks:
Comshare
IBM
Lotus 1-2-3
MAUD
MS-DOS
Personal Wizard
System W

The author was supported by a grant under the Ministry of Education "National
Facility for Informatics" programme (project NFI-1).

CIP-DATA, Koninklijke Bibliotheek, The Hague
Ven, Michael Bartholomeus Maria van der
A group decision support system for transfer pricing in the pharmaceutical industry /
Michael Bartholomeus Maria van der Ven. - [S.l. : s.n.]. - 111.
Thesis Delft. - With index, ref. - With summary in Dutch.
ISBN 90-9002887-0
SISO 385.9 UDC 681.3:[338.5:334.71:615.2](043.3)
Subject heading: group decision support systems / transfer prices / pharmaceutical
industry

ISBN: 90-9002887-0

© 1989 by M.B.M. van der Ven

All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system, or transmitted, in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without written permission of the author.

Contents

1 BACKGROUND AND RESEARCH QUESTIONS 1
1.1 Introduction 1
1.2 Research questions and research approach 4
1.3 Outline of the thesis 9

2 THE EUROPEAN PHARMACEUTICAL INDUSTRY 11
2.1 Introduction 11
2.2 Some characteristics of the European pharmaceutical industry 12
2.3 Relationship with national health organizations 15
2.4 The approach of the British National Health Service 16
2.5 Conclusion 21

3 TRANSFER PRICING: A COMPLEX ISSUE 23
3.1 Introduction 23
3.2 Management-control transfer pricing 23
3.3 International transfer pricing 29
3.4 Transfer pricing in the pharmaceutical industry 38
3.5 A framework to describe transfer pricing 40
3.6 Conclusion 43

4 GROUP DECISION SUPPORT SYSTEMS 45
4.1 Introduction 45
4.2 Structuring decision processes 47
4.3 Multi-criteria decision making 54
4.4 Approaches to group decision support systems 56
4.5 Four aspects of developing group decision support systems 61
4.6 Conclusion 64

5 TRANSFER PRICING AT EUROPHAR: THE PROBLEM 67
5.1 Introduction 67
5.2 Europhar: the company 69
5.3 Decision process before system introduction 73
5.4 Descriptive empirical model 78
5.5 Descriptive conceptual model 81
5.6 Conclusion 85

6 TRANSFER PRICING AT EUROPHAR: A SOLUTION 87
6.1 Introduction 87
6.2 Prescriptive conceptual model 87
6.3 Prescriptive empirical model 107
6.4 Decision process after system introduction 116
6.5 Conclusion 135

vu

7 EVALUATION 137
7.1 Introduction 137
7.2 Questionnaire and answers 141
7.3 Overall evaluation 147
7.4 Conclusion 151

8 SUMMARY AND CONCLUSIONS 153
8.1 Summary of findings 153
8.2 Future research 155

REFERENCES 157

List of exhibits 165

Samenvatting in het Nederlands 167

Acknowledgements 171

List of contributing institutions 172

About the author 172

Index 173

viii

1

BACKGROUND AND RESEARCH QUESTIONS
1.1 Introduction
As a result of the post-war expansion of foreign direct investments, trade is
increasingly dominated by multinational companies, so called "multinationals". An
increasing part of this trade is intra-firm trade in the sense that the seller and buyer
belong to the same multinational group of companies.

According to Murray [1981, p. 2] around 40% of the exports of multinationals based
in the major Western industrialized countries were intra-group transfers and these
figures are rising. According to the report of the Independent Commission on
International Development Issues, intra-firm trade makes up over 30% of all world
trade, see Brandt et al. [1980, p. 188]. Although these figures should be handled with
care, they give a good indication. Besides the internal sales of goods, other transactions
take place extensively between the different parts of these enterprises, for example the
granting of loans, the licensing of technology and the provision of services. The price
which is set for the transfer of goods and services between two companies belonging
to the same group, is called "transfer price". For the purpose of this project, a more
specific definition will be introduced in chapter 3. "Transfer pricing" is defined as the
process of deterrnining transfer prices. It is related to decision making.

Within large multinational companies, decentralization is necessary as centralization of
decisions at headquarters would become counterproductive: it would overtax the time
and energy of top management. The background of transfer pricing can be found in
this decentralization principle in decision making, see chapter 3. Sheer size and the
resulting complexity of multinationals argue against centralized management. Secondly,
vesting the decision-making responsibility with the local manager allows for a more
flexible response to local circumstances which can vary enormously, see
Plasschaert [1979, pp. 23-24].

Decisions concerning transfer prices may reflect legitimate business concerns of a
company, but they are also capable of being used in order to shift profits from high
to low tax countries or to get around exchange or price-control regulations. Such
practices; although pursued in the best business interest of a company, may conflict
with the interest of host countries. This is not to suggest, according to Brandt et
al. [1980, p. 189], that as a class, multinationals have been using such practices.
Plasschaert [1985, p. 263] states that, with regard to over- and underpricing the
evidence is meager and he points to the dangers of undue generalizations. He protests
against the connotation that multinationals "are the villains" and points to the
sometimes negative effects of government measures concerning transfer prices.

Nevertheless, caused by the increased size of intra-company trade of multinationals and
the negative publicity on the subject in the seventies, see for example Vaitos [1974],

1

BACKGROUND AND RESEARCH QUESTIONS Ch. 1

government interference has increased. So the multinational has to operate in an
increasingly strict legal framework which has remained national. It has therefore to
conform with the varied and sometimes conflicting national laws in countries in which
it operates, see OECD [1979, p. 7]. The increased governmental interference in its turn
causes a growing interest in the various aspects of transfer pricing from business as
well as from governmental side.

In order to demonstrate the complexity of the issue, literature on international transfer
pricing often presents examples from the pharmaceutical industry, see for instance
Verlage [1975] and Plasschaert [1979]. According to Nowotny [1987] and Von
Grebmer [1987a], transfer-pricing problems are especially poignant for the
pharmaceutical industry. This is caused, amongst others, by their relatively high degree
of internationalization in combination with centralized production and research facilities.
In other words: high centralized costs need to be covered by international intra-
company sales. The position of the pharmaceutical industry will be explained in more
detail in chapter 2.

In all phases of the decision process, the problem of international transfer pricing,
especially in the pharmaceutical industry, is considered to be highly ill-structured.
Sol [1982, p. 5], discusses various definitions of ill-structuredness. One of these
definitions is taken from Uhr [1973, p. 274]: "Ill-formedness in problems seems to be
closely related to issues of vagueness, ambiguity, flexibility, and creativity about all of
which little is known." With Sol, Takkenberg [1983a, p. 10] notes that many the existing
definitions of ill-structuredness are vague. To avoid ambiguity, we use the definition of
Sol [1982, p. 5]. Sol starts with the definition of a well-structured problem: "A problem
is well-structured if the following conditions are met:
1. the set of alternative courses of actions or solutions is finite and limited,
2. the solutions are consistently derived from a model system that shows a good
correspondence,
3. the effectiveness and efficiency of the courses of action can be numerically evaluated"
Problems that do not fulfil these requirements are defined as ill-structured.

In literature on decision support, many authors have stressed the importance of
supporting decision making with regard to ill-structured problems. Attempts at
supporting decision making in ill-structured problem situations have focused on the
development of so called "decision support systems". The aim of these systems is the
use of computers to aid decision makers, see Sol [1985b] and Bosman [1987a].

During the last decade, decision support systems have gained dramatically in popularity.
Research on decision-making and decision support systems has increased likewise, see
McLean and Sol [1986]. According to Sol [1987b, p. 203], experiences with decision
support systems are not always positive: "... [We] observe that empirical evidence on
the effectiveness of decision support systems is not overwhelmingly positive, and often
conflicting", see also Sol [1987a]. On the basis of laboratory experiments, Van
Schaik [1988, p. 127] concludes that "teams using a decision support system do not
perform better than those without."

2

Sec. 1.1 INTRODUCTION

With reference to DeSanctis and Gallupe [1985] we define a decision support system
as an interactive computer-based system which facilitates solution of ill-structured
problems. Not that this definition does limit itself to computer support only: other
elements of the organization, like data, procedures and organizational members can
also be included, see also Bots and Sol [1988].

Confining ourselves to the pharmaceutical industry, we see the following: on the one
hand, we have an ill-structured problem of increasing importance to the pharmaceutical
company, namely international transfer pricing. On the other hand we have a relatively
new research area, namely decision support systems. We expect that decision support
systems are useful for determining international transfer prices within the
pharmaceutical industry.

As will be explained in more detail in chapter 2, the environment of the
pharmaceutical industry is complex, hostile and turbulent, especially with regard to
transfer pricing.

The environment is complex as many actors are involved: consumer groups, various
governmental bodies from many countries, supra-national organizations like the World
Health Organization and the European Community, health insurance companies and
last but not least the competition. The high degree of internationalization and the
division between prescribing, using and paying for drugs are amongst the causes of the
complexity of the environment.

Hostility of the environment is not only caused by the competition but also by the
groups that oppose current price levels and international price differences. According
to De Jong [1981] "the industry, which earlier contributed to the conquest of many
diseases, is no longer in favor with the general public", see also Breuk [1985]. It is
under scrutiny from health-care authorities and consumer groups. Books with titles like
"Corporate crime in the pharmaceutical industry", see Braithwaite [1984], give an
impression of the hostilities the industry has to face. In the book we just mentioned,
a special section is dedicated to transfer pricing.

The environment is turbulent as constantly new regulations are issued by the authorities
involved. The developments with regard to the common market and changing measures
with the aim to contain health care costs are examples, see Scrip [1988]. Moves of
competitors also need to be monitored closely.

According to Huber and McDaniel [1986, p. 577], complexity, hostility and turbulence
are the three main characteristics of the organizational environment in the post-
industrial society. For this kind of environment, Huber and McDaniel propose the
"decision-making paradigm". Like Simon [1973, pp. 269-270], Huber and McDaniel
consider decision making as the organization's central activity. In view of this,
improving decision making in organizations should be the focus of attention of the
organizational sciences. According to Huber [1984a, 1984b], one of the elements that
can contribute to improved decision making of post-industrial organizations are so-

3


called "group decision support systems". These are decision support systems explicitly
designed to be used by groups of decision makers.

With regard to transfer pricing, pharmaceutical companies find themselves in a
complex, hostile and turbulent post-industrial environment. Following Huber, group
decision support systems should be useful in such a setting. In other words, transfer
pricing in the pharmaceutical industry seems a "perfect" example to test the concept
of group decision support.

To put it more formally: it is our hypothesis that group decision support systems are
helpful for improving decision making with regard to the determination of transfer
prices in the pharmaceutical industry.

1.2 Research questions and research approach
In this section, we cover some issues that are related to our research approach.
According to Fitzgerald et al. [1985, p. 5], information technology is still so new, that
not one adequate research approach for scientific progress exists but that pluralism
needs to be accepted. Our basic presumption is that these fundamental issues must be
answered within the context of a particular research setting. That is to say, a research
approach is neither appropriate nor inappropriate until it is applied to a specific
research problem, see Downey and Ireland [1979, p. 630]. This holds also for the two
research questions which we try to answer. These research questions are:

A: What are the characteristics of a group decision support system designed to support
and improve organizational decision making with regard to transfer prices within the
pharmaceutical industry?

B: What are the effects of the use of this designed group decision support system on
decision making within the organization?

Following Huber and McDaniel [1986], we consider decision making as the
organization's central activity. In the literature, various approaches towards decision
making exist. One of these approaches is based upon "bounded rationality", see
Sol [1982, p. 7].

In our research approach we take the paradigm of bounded rationality as point of
departure. This implies that we do not presume that decision makers have all relevant
information and an unlimited capacity to process this information. Our point of
departure is that decision makers operate "in their own environment with own
definitions and limited scopes on accessible data", see Sol [1982, p. 7]. Simon [1976,
pp. 65-86] considers behavior bounded rational "when it is the outcome of appropriate
deliberation. [Bounded rationality] depends on the process that generated it." Discussing
bounded rationality, Takkenberg [1983a] states that "it is impossible to define
unambiguously what is rational. ... Rational is now also determined by the multitude
of circumstances that influence the decision maker".

4

Sec 1.2 RESEARCH QUESTIONS AND RESEARCH APPROACH

In order to determine the multitude of circumstances that influence a decision maker,
we need to analyze the existing situation, see Bosman [1986, p. 315] and
Takkenberg [1983b, p. 258]. On the basis of this analysis, we make a description of the
decision process, as it takes place. We define this description of the "as-is" situation
as the descriptive empirical model, see also Sol and Van der Ven [1989]. This model
is descriptive in the sense that the actual ("as-is") situation is described. No normative
elements are included. This first model is empirical in the sense that it describes an
actual situation which exists within one specific organizational group. Names of
organizational members, financial figures and other specific details will be included. On
the basis of one or more descriptive empirical models, a descriptive conceptual model
is constructed. This descriptive conceptual model is "data void", see Bosman and
Sol [1985, p. 82]. This model is still a description of the "as-is" situation, but now at
a higher level of abstraction: the essential characteristics of the various empirical
models are transformed into a possibly more generic description which provides global
insight in the existing situation. The level of abstraction is such, that this model can
be used for other organizations which are in a similar position.

On the basis of this latter descriptive model, literature research and other explorations,
a third model comes into existence: the prescriptive conceptual model. The prescriptive
conceptual is a combination of "description and design principles", see Bosman [1987b,
p. 32]. It is prescriptive in the sense that it contains normative elements, see
Takkenberg [1983a]: using this model for decision making should lead to an improved
decision process.

When elaborating on our first research question, we will determine three of the four
model categories as presented in exhibit 1-1. In this exhibit, the numbers 1-4 indicate
the order of the activities that we follow. These four activities precede the construction
of the corresponding model.

descriptive prescriptive
model model
1: 4:
empirical data collection implementation
model concerning the of the improved
existing situation model

2: 3:
conceptual abstraction analyses and
model improvement of
the model

Exhibit 1-1. Categories of models.

The prescriptive conceptual model is our answer to the first research question.

In order to elaborate on the second research question, we need to perform a test in
the form of experimental use of the system. Experimental use of the system requires

5


the set up of one or more empirical models on the basis of the prescriptive model
that we designed. This latter model is defined as the prescriptive empirical model. The
use of this empirical model in the daily practice of a pharmaceutical company provides
the desired insight on the effect of the group decision support system on decision
making. By means of analysis of the effect of the prescriptive empirical model, we
evaluate our prescriptive conceptual model.

Now we explained various categories of models, we can rephrase our two research
questions:

A': What are the characteristics of the prescriptive conceptual model?

B': What are the effects of the use of the prescriptive empirical model on organizational
decision making?

With regard to organizational decision making, Mintzberg [1979, p. 584] states the
following: "...It seems that the more deeply we probe into this field of organizations,
the more complex we can find it to be, and the more we need to fall back on so
called exploratory, as opposed to 'rigorous' research methodologies." To obtain a
prescriptive conceptual model, Mintzberg [1979, p. 587] proposes research within one
organization, sometimes even for years, as: "We need to be 'in touch'. Increasingly in
our research we are impressed by the importance of phenomena that cannot be
measured: by the impact of an organization's history and its ideology on its current
strategy, by the role that personality and intuition play in decision making. To miss this
in research is to miss the very lifeblood of the organization. And missed it is in
research what by its very design, precludes the collection of anecdotal information."
DeSanctis and Gallupe [1987, p. 593] propose longitudinal research specifically for
research in the area of group decision support systems. Our research activities should
be regarded as one of the "pilot testing" activities, see DeSanctis and Gallupe [1985,
p. 8], that are required to progress in the area of group decision support.

We explicitly choose to answer these questions by means of a real-life case study, see
Sol and Van der Ven [1988a]. This choice is based upon the following considerations,
see Benbasat et al. [1987]:
- little research exists which integrates all aspects of international transfer pricing,
- the project has a revelatory character, see Yin [1984, p. 47] as not much is known
about the way transfer-price decisions are taken in practice,
- as improvement of "real-life" decision making is one of the aims, close cooperation
with the industry is imperative.

It should be realized that "case study" as used above does not mean that the
researcher is watching the organization from the sideline. He is actively involved in the
design of the prescriptive conceptual model and should be regarded as a member of
the organization.

In order to analyze the existing situation with regard to transfer pricing and to test a
pilot system, we decide to cooperate with a large pharmaceutical company for a period

6


of five years. For the definition of "large pharmaceutical company" in this respect, see
Van der Ven [1989a]. In this thesis we name this company "Europhar". In research
terms, five years is a relatively long period of "on-site" analysis. Our research is of an
explicit exploratory nature.

On the basis of this on-site research, the descriptive empirical model and the
descriptive conceptual model will be created. In general, decision makers do not know
how they reach a decision. Usually, they only know the outcome, the decision itself.
Therefore the descriptive conceptual model will be based on long "on-site" observations,
direct involvement in organizational activities, open interviews and video-taped meetings.

The exploratory approach that we choose to follow is described by Conklin [1968, p.
172] as involving "a long period of intimate study and residence in a well-defined
community employing a wide range of observational techniques including prolonged
face-to-face contact with members of local groups, direct participation in some of the
groups activities, and a greater emphasis on intensive work with informants than on the
use of documentary or survey data".

Strong involvement in an organization, however, will not be sufficient for the
development of a prescriptive conceptual model. It is necessary to observe the
organization from a (mental) distance as it "would hardly be a fish who discovers the
existence of water", see Wolcott [1975, p. 115]. For us, the involvement in literature
research will provide this distance. Literature research will be concerned with the areas
of transfer pricing and decision support. Information on the pharmaceutical industry
will be gathered.

On the basis of literature research and the descriptive conceptual model, we will
design a prescriptive conceptual model. Based on this latter model, a prescriptive
empirical model for one subsidiary company of Europhar will be made.

This prescriptive empirical model should be regarded as an experimental system and
not as a full system, see Keen [1986, p. 51]. Our explicit aim is to demonstrate the
feasibility of the group decision support approach for international transfer pricing. The
development of the experimental system is regarded by Europhar as a low-risk research
and development venture, especially as a substantial part of the costs are borne by
various Dutch research institutions. Such an experimental "version 0" is defined by
Keen [1986, p. 50] as "an initial small system which is complete in itself, but may
include limited functional capability". The aim of the system, in our case the
prescriptive empirical model, is to establish the value of the proposed concept.

For this experimental system, we need to select one subsidiary of Europhar
International. For reasons discussed in chapter 5, Europhar UK Ltd in London will
be selected. In the research project decision making concerning Europhar UK plays
a dominant role and will be described in detail, see exhibit 1-2.

7


information description
on the transfer- decision- decision
pharmaceutical pricing support process
industry literature literature Europhar UK
(before)

descriptive
empirical
model
Europhar UK

descriptive
conceptual
model

prescriptive
conceptual
model

question A': what are the characteristics of the prescriptive
conceptual model?

question B 1 : what is the impact of the prescriptive empirical
model on organizational decision making?

information ''
on
Europhar UK prescriptive
empirical
model
Europhar UK

''
description
decision comparison
process
Europhar UK
(after)

Exhibit 1-2. Overview of the research approach.

8


It should be realized that the analysis of various other decision situations, for European
as well as for third-world countries, will contribute to the construction of the
descriptive conceptual model, although they will not be described in detail.

The "real-life" use of the experimental system (after implementation for the UK),
together with the description of the already existing decision-process (before
implementation for the UK) will form the basis for the evaluation.

1.3 Outline of the thesis
This thesis can basically be divided into two parts: the chapters 1 to 4 provide a
theoretical background. In the chapters 5 to 8, the Europhar case study is discussed.
The outline of the thesis is as follows:

In chapter 2 we discuss the pharmaceutical industry. It is explained why we limit
ourselves to the European pharmaceutical industry.

In chapter 3, we explain the research limitation to international transfer pricing. We
will introduce a framework for analyzing financing and transfer-pricing issues.

In chapter 4, backgrounds, possibilities and recent developments in the area of group
decision support systems are discussed. The possibilities are related to transfer-pricing
problems and lead to the design guidelines of the prescriptive conceptual model.

In chapter 5, we present the Europhar company and describe the descriptive empirical
and conceptual model.

In chapter 6, we discuss the design of the prescriptive conceptual and empirical model.
In this chapter, the use of the experimental system during six months is discussed.

In chapter 7, various approaches to the evaluation of information systems are discussed.
Interviews are the main basis of the evaluation of the pilot system. Outcome of the
interviews and anecdotal evidence are discussed.

Chapter 8 presents the conclusions and possible areas for future research.

The pharmaceutical industry, transfer pricing and group decision support are in itself
complex issues. In the chapters 5, 6 and 7, these three areas come together in the
descriptive models, the prescriptive models and the evaluation. For a proper
understanding of what is presented in these chapters, the background knowledge as
presented in chapters 1 to 4 of the thesis is essential.

9


10

2

THE EUROPEAN PHARMACEUTICAL INDUSTRY
2.1 Introduction
As stated in the previous chapter, we limit ourselves to transfer-pricing problems for
the following category of multinationals:

The European-based innovative pharmaceutical industry which develops, produces and
sells human ethical drugs.

All aspects, of this limitation are now described below. We shall explain the rationale
behind this limitation.

"Human ethical drugs" are drugs intended for human use. Use by the patient is only
allowed upon a prescription by a medical practitioner. This in contrast with so called
"over the counter" drugs like Paracetamol. For over the counter drugs, no prescription
is needed. By this limitation, veterinary products are also excluded. We limit ourselves
to this part of the pharmaceutical industry, as interference by governmental health
care organizations is especially intense for these kinds of products. As will be
explained in this chapter, governmental interference in this area influences transfer-
price decisions.

With "Europe" in this respect, we mean all countries that participate in the European
federation of pharmaceutical industry associations "EFPIA". Besides the current member
countries of the European Community (except Luxembourg), Austria, Finland, Norway,
Sweden and Switzerland participate in this organization, see EFPIA [1987]. By
"European-based" companies we mean those companies which have their headquarters
in Europe. As far as the pharmaceutical industry is concerned, outside Europe, only
Japan and the United States need to be considered, see IMS [1988]. We exclude these
non-European based pharmaceutical companies as decision making may be different
due to sharply different circumstances on the home market. This holds for both Japan
and the United States, see Burstall [1985, pp. 125-127] and Reinshagen [1988]. In
general, it can be said, that the Japanese and American firms have another approach
to decision making than European firms. This is due to cultural differences, see
Nath [1988]. It is possible that these cultural differences influence vital presumptions
of our approach, like team setting. (Team setting will be discussed in more detail in
the next chapter).

That we decided to include all European countries is because of the information we
received from the European-based industry. Like Burstall [1985, p. 2], we noted that
European companies are more "notable for their degree of uniformity than divergence".
Partially, information on the industry was systematically gathered, see Van der
Ven [1989a]. For another part, we received feed-back in informal contacts with industry
representatives during company visits, conferences and alike. It is interesting to note

11

THE EUROPEAN PHARMACEUTICAL INDUSTRY Ch. 2

that Nath [1988] treats European firms as one class. This in spite of the differences
that exist within Europe.

The limitation to innovative companies excludes those companies which only produce
and sell so-called "generics". Generics are drugs which are usually sold under a
chemical name instead of under a brand name. They are copies of good selling human
ethical drugs developed by innovative pharmaceutical companies. These copies are
brought on the market after the patent period is expired. As for the generic producers
the product and the market are already developed, marketing and research expenditure
are much less than for the research-based industry. In this chapter we shall explain
how high fixed marketing and research costs influence organizational decision making.
The exclusion of generic producers, however, does not imply that innovative companies
never produce and sell generics. The difference with non-innovative companies is that
generics only form a minor part of their sales. Within innovative companies sales of
generics contributes to research cost.

To give a global impression of the size of the pharmaceutical industry in EFPIA
countries, the following figures for 1986 are presented, see EFPIA [1987]:
- production: 43,000 million ECU,
- employment: 450,000 people.
(In the beginning of 1986, the value of the ECU was 2.46 Dutch guilders.)

When in the remaining part of this thesis we discuss "the European pharmaceutical
industry", we mean those companies as indicated by the description above.

2.2 Some characteristics of the European pharmaceutical industry
Lorsch and Lawrence [1970] demonstrated the interaction that exists between
organization and environment leading to the contingency (or situational) approach to
organizational design. As environmental characteristics may influence the organizational
structure of a company, market structure, governmental interference and cost structure
will now be discussed.

Market structure: separate therapeutic areas
The pharmaceutical market is in itself subdivided into many self contained therapeutic
areas, see Burstall [1985, p. 5]. For instance, in 1985 the Dutch market was divided
in 24 submarkets which each contained between 2% and 7% of the market, see
Nefarma [1985, p. 18]. Examples of these submarkets are the market for psychotropics
and the market for anti-asthmatics. Diversification within the industry is low as a
company needs to concentrate on one or a few of these therapeutic areas. Following
Eccles [1985, p. 272] we use the term diversification to refer to the number of different
businesses in which the company competes and how different these businesses are from
each other. Costs of finding and developing a new drug are extremely large, according
to Von Grebmer [1987b] about 100 Million ECU for every newly-introduced
medicament. According to Von Grebmer [1987b], there are technical and economic
limits to decentralization of these research activities. This centralized research

12

Sec. 2.2 SOME CHARACTERISTICS OF THE PHARMACEUTICAL INDUSTRY

expenditure is directed to one or a few therapeutic areas and this expenditure can only
be recovered by sales on this small part of the total pharmaceutical market. This
implies that the submarket needs to be covered worldwide to recover research costs.
Hence, strong internationalization is a common characteristic of these pharmaceutical
companies. It should be realized that size and other characteristics of a certain part
of the market may vary widely depending on the country and the therapeutic area that
are serviced. In 1982, for example, over 40% of sales from the pharmaceutical industry
of the European Community was realized outside the common market, see
Burstall [1985, p. 115].

Governmental interference
Due to the involvement of governmental authorities in health-care matters, sales of
pharmaceuticals is subject to an "unusual degree of government regulations", see
Burstall [1985, p. 14]. Despite a general tendency towards deregulation, recent trends
do not point to any significant reduction in the stringency of drug regulations, see
Hansen [1987].

Pharmaceuticals are closely linked to the issue of human health care, which is
considered to be one of the prime governmental responsibilities. Many of these
regulations are related to the drug safety and efficacy issues. Human health care is
very sensitive as far as public opinion is concerned. The issues at stake require careful
distribution and monitoring. This implies that the distribution of pharmaceutical
products cannot be delegated to a third party with the instruction to "just get the
product sold". Should something go wrong, hostile publicity and legal pursuits would
force the company involved in a difficult position. Therefore strict control on quality
and logistics is needed. Secondly, training and instruction of representatives is of vital
importance to assure proper use of new and existing products. Well trained
representatives are important to assure two way communication between the industry
and prescribing doctors, which is necessary from a marketing as well as from a
medical point of view, see ABPI [1985, p. 32]. Regulations require, for example, that
the company that develops, produces and sells a drug keeps closely in touch with the
market in order to pick up any adverse drug reactions (post-marketing surveillance).

To assure this proper two way flow of information between the parts of the company
that are in touch with the local market, and the head office, a strong "vertical
integration" is needed. Following Eccles [1985, p. 272] we use the term vertical
integration to refer to the inclusion of activities within the firm that could be obtained
externally, such as raw materials, distribution channels and staff services. So within the
pharmaceutical industry one finds that fundamental research in search for a new
product as well as the final instruction of medical professionals in a remote country
is done by people of the same multinational company within a framework of strict
control to the board of management, see also Von Grebmer [1987a]. This strict control
reinforces centralization.

Regulations do not only pertain to registration of new products or quality control
matters. Price regulations are involved as well. This will be discussed in the next
section.

13


Cost structure: high fixed costs
According to Von Grebmer [1987b, p. 233 and 234], "knowledge of the specific cost
structure of the research-based companies is the key to understanding their market
behavior when considering their pricing strategies ... On average, not more than 30
percent of the cost can be allocated directly to individual products, the remaining costs
and the profit margin have to be covered by contributions from total sale of the whole
product range." In other words, judgement of profitability of a pharmaceutical company
should take place at the aggregate, total company, level and not on the disaggregate,
individual product, level. Selling prices at the disaggregate level are unavoidably based
upon a more or less arbitrary allocation of fixed costs.

As the relatively high fixed costs are made anyway, the total financial contribution that
a local marketing subsidiary brings in, is of greater importance than its local net profit.
So it is vital that the energy of the sales force is spent on the sales of in-house
developed products and not on the sales of other products which might produce, in
the short run, better results for one local subsidiary of the company. This cost
structure contributes to a high vertical integration and centralization of decision making.
In other words, within the industry, the delegation of power and responsibility to lower
organizational levels (decentralization) is relatively limited.

The various items discussed above, make up the organizational environment for the
European pharmaceutical industry and determine the main characteristics of the
industry: strong internationalization and centralization, strong vertical integration and
low horizontal integration, see also Von Grebmer [1987b, p. 240]. In exhibit 2-1, we
present the structure of a typical pharmaceutical company. The arrows indicate the
flow of goods.
management center
parent shareholder
company supplier
research center

national
border <...other subs
companies
<' ''
subsidiary subsidiary marketing company
company company

Exhibit 2-1. Organizational structure of a typical pharmaceutical company.

The parent company takes care of research, production, financing and marketing
support. The parent company is the shareholder of the local subsidiary. The main
task of the subsidiary company is to market the drugs, which are developed and

14

Sec. 2.2 SOME CHARACTERISTICS OF THE PHARMACEUTICAL INDUSTRY

produced by the parent company, on the local market. (Other words for subsidiary
company are "local company" or "daughter company").

2.3 Relationship with national health organizations
Governmental interference in the supply of pharmaceuticals takes place in many
countries all over the world, see for example Bouee [1987] and PBPA [1988]. In this
section we will discuss the developments within Europe as an example. This is because
Europe constitutes the home market of the industry.

In the European pharmaceutical market, a three-tier demand system exists: the
physician prescribes the product, the patient uses it, and national health organizations
pay the bill.

Background of the increasing interference in pricing of pharmaceuticals is the aim to
contain health-care costs. Health-care costs are in most countries a part of collective
expenditure, which in general suffer from large deficits. This pressure from public
health services is a matter of great concern to the industry as "control over aspects of
pharmaceutical company operations do not in general result from a considered view
of the industry as a whole", see Burstall [1985, p. 14]. Governments and the European-
Community authorities are under pressure of consumer groups to force down prices,
see for example BEUC [1984]. In the annual report of the Netherlands-based
multinational Akzo [1985], this concern is expressed as follows:

"The need for constructive talks between authorities and innovative pharmaceutical
companies is growing. As we have argued on previous occasions, there is an
ever-present danger that, in their understandable desire to control expenditures, the
authorities will accept the prices charged by the producers of generics (drugs not
protected by patents and trademark registration) as the bench mark for national health
insurance payments. Were they to do so, there would be little or no margin left to
conduct time- and money- consuming research for the new, better drugs so urgently
needed. Such research, after all, is funded out of the revenues generated by the sale
of established products. While this issue especially concerns the industry, it is also a
significant public health matter. We are prepared to join with the authorities in looking
for ways to reconcile these divergent objectives".

Within Europe, various ways to control prices exist. A basic difference between these
price-regulation schemes is their approach to the issue of aggregation and
disaggregation. Aggregate level price-regulation schemes control total-company profit of
a specific subsidiary company. The disaggregate level pricing schemes control prices by
fixing the selling price of individual pharmaceutical products. For an overview of the
various approaches, see Van Cayseele [1987].

In view of the realization of the European common market, initial steps with regard
to harmonization of the pharmaceutical market are being made by the European
Commission. First aim of the Commission is to clarify the rules that are issued by

15


the various member states. As far as reimbursement systems are concerned, these
attempts stress above all the importance of transparency in the operation of national
systems so that importers can verify that objective criteria are genuinely being applied
and all relevant factors are being taken into consideration by member states. Given the
different medical traditions of the member states, however, the European Community
authorities realize that it would appear unrealistic to seek to lay down a completely
harmonized community price structure for pharmaceuticals, see Hankin [1988].

Besides the transparency in procedures followed by member states, a first step is
made to increase transparency on prices themselves. By the European Commission, a
database will be set up which, amongst others, contains ex-factory and retail prices, see
EG [1989]. Although the registration of transfer prices was considered by the
Commission, it was decided that these would not be included in the database as
transfer-pricing information was considered to be of a confidential nature.

The involvement of a multi-state framework like the European Community in
pharmaceutical pricing and reimbursement issues, poses additional problems for the
industry. The directive concerning transparency represents only the start of a much
wider debate on pricing, transfer pricing and reimbursement. The industry is, on the
one hand, still faced with individual sovereign governmental policies. On the other hand
it is confronted with a multi-state framework committed to the goal of a unified
market, see Owen [1988].

2.4 The approach of the British National Health Service
In the previous section, governmental interference with (transfer) prices has been
discussed globally. In this section, one price-regulation scheme is worked out in more
detail to provide insight in the practical consequences of such a scheme for the
companies involved. We choose the British scheme as, according to Burstall [1985, p.
152], "there is a general feeling among the major [European] drug companies that, if
prices must be controlled, then the British system is the most acceptable. It also
commands interest and respect in official circles elsewhere."

The scheme is limited to branded human ethical drugs paid for by the National Health
Service (NHS). Generics are excluded. For further details see DHSS [1986]. The main
purposes of the NHS scheme are to:
- secure the provision of medicines for the National Health Service at reasonable
prices,
- to promote a strong pharmaceutical industry capable of sustained research
expenditure.
Note that "reasonable prices" are not the only aim of the scheme. The NHS authorities
(the "Department of Health and Social Security") also act as sponsor for the industry,
on behalf of the British government. Aspects that are relevant for the authorities as
part of their "sponsor" responsibilities are the size and quality of employment in the
United Kingdom and the size of exports. According to the Association of British

16

Sec. 2.4 THE APPROACH OF THE BRITISH NATIONAL HEALTH SERVICE

Pharmaceutical Industries, for example, recent changes in the regulations encourage the
development of research and production facilities in Britain, see Scrip [1987a, p. 5].

The "Pharmaceutical Price Regulation Scheme" provides a set of regulations for the
calculation of profit (the "Annual Financial Return"). This profit calculation forms a
basis for negotiations between each individual pharmaceutical company and the NHS
authorities. Focal point is the profit that the company makes on sales to the NHS. The
NHS-profit calculation of each company needs to be completed annually and submitted
to the NHS authorities. All costs claimed are examined critically by the health service.
In principle most of the costs are allowed, but there are however some special
constraints. If the health service considers that profits of a company are excessive
under the terms of the price-regulation scheme, price reductions or cash repayments
are sought to the extent necessary to reduce profit to an acceptable level. The
negotiations between pharmaceutical companies and the authorities have been described
as "tough but fair", see Scrip [1987b, p. 2]. The target profit on which price increases
or repayments are based, are established by taking the average return on capital of
British industries in other public sector business, like aircraft-building companies. This
target profit is influenced by the state of the public finances and subject to negotiation
with each individual company.

NHS sales
-- transfer prices (landed costs)
-- manufacturing costs
-- manufacturing overhead
gross margin
-- distribution costs
-- information costs
-- promotional costs

operating performance
-- royalties
-- general overheads
profit before interest and taxation

Exhibit 2-2. First part of the NHS-profit calculation.

Usually the target profit is expressed as a return on capital. This return on capital is
the focal point during the negotiations between the authorities and the individual
company. If the actual return on capital is greater than the target, defined as an
"excess", a repayment to the NHS might be the consequence. In case the actual return
on capital is smaller than the target, defined as a "shortfall", a price increase might be
the consequence. The calculation basically exists of three parts:
- calculation of the "profit before interest and taxation", starting from the NHS sales,
- calculation of the "total adjusted profit", starting from the profit before interest and
taxation,
- calculation of the "excess/shortfall on the target return on capital", starting from the
total adjusted profit.

17


The global outline of the first part of the NHS-profit calculation is presented in exhibit
2-2.

In exhibit 2-2, the sign "--" stands for "minus". The gross margin, for example, is
calculated by subtracting transfer prices (or landed costs), manufacturing costs and
manufacturing overhead from the NHS sales. We presume that lump-sum royalties are
paid to the mother company and the mother company is the only supplier. All figures
are expressed in British pounds sterling (GBP). This set of figures is made up for
NHS sales as only these sales are controlled. Generic sales in Britain or exports are
excluded in the profit calculation. When costs are directly related to activities (like
distribution, information and promotional costs), there is no need for allocation. In
cases that costs cannot be precisely apportioned between NHS sales, generic sales and
exports, an artificial cost allocation has to be made. Possible bases for allocation are:
- net sales,
- local manufacturing costs (added value),
- costs of goods (costs of goods equal the sum of landed costs and local
manufacturing costs).
The first step of the NHS-profit calculation is to calculate the gross margin. The
distribution, information and promotional costs are subtracted from the gross margin
to obtain what is defined as the operating performance.

The scheme makes a strict distinction between allowable and non-allowable costs.
Allowable information costs include for example expenses on samples for identification
purposes and medical symposia. Allowable promotional costs include among others
literature, representatives, administration and advertising. Gifts are not allowed as a
promotional cost on the profit calculation.

Both royalties and general overheads are subtracted from the operating performance
to obtain the profit before interest and taxation. The authorities are only interested in
the profit before interest and taxation. The companies are not allowed to include
taxation costs and interest costs in the profit calculation because the different financing
structures of the companies would distort the comparison between companies.

Now follows the second part of the NHS calculation. The profit before interest and
taxation, as calculated above, is the basis for the calculation of the "total adjusted
profit" as presented in exhibit 2-3.

profit before interest and taxation
+ add-back injected costs
+ add-back promotional costs
-- export discouragement correction

total adjusted profit

Exhibit 2-3. Second part of the NHS-profit calculation.

18

Sec. 2.4 THE APPROACH OF THE BRITISH NATIONAL HEALTH SERVICE

An add-back for royalties and transfer pricing (defined as add-back injected costs) is
added to the profit before interest and taxation, when the sum of transfer prices and
royalties exceeds a certain percentage, considered acceptable by the NHS, say x%. In
this case the NHS authorities consider the "injected costs" (cost transferred from the
mother company) too high and demand a correction on profit before interest and
taxation. So the formula for the add-back is as follows:

add-back injected cost =
(allocated royalty + NHS transfer price) -- x% of NHS sales

Here we can clearly see how transfer pricing decisions and regulations of health-care
authorities interfere: the value of the add-back on injected cost equals the difference
between transfer prices added to royalties and x% of the total net sales. This x% is
a negotiated figure and differs per company. It is important to note that there exists
only an add-back for transfer prices and royalties. Where the local picture does not
satisfy the authorities, they then consider the consolidated picture. If, however, the sum
of transfer price and royalties is lower than the negotiated percentage, no action is
taken. In other words, there is no symmetry in the approach of the authorities in this
respect. They only adjust the figures where this is to their advantage.

Say, for example, that a British subsidiary of a multinational company has liquidity
problems and that a transfer-price decrease is considered. In case this company should
decide to lower the sum of transfer prices and royalties to a percentage lower than
the agreed x%, it artificially increases its profits. This high profit could ultimately result
in a repayment to the national health service. This example clearly illustrates the
interference of transfer-price decisions with health care regulations.

The NHS sets a maximal sales-promotion allowance. Not only are some costs
unacceptable, when a company spends more then this allowance, the amount overspent
(add-back promotional costs) has to be added to the profit before interest and taxation.
Also a cash repayment of the amount overspent is requested. For more details on the
background of this punishment for overspent promotional costs, see Haayer-Ruskamp
and Dukes [1986].

A strict apportionment of capital and costs between home and export markets may in
some circumstances produce an unfair result. High export sales may result in a smaller
part of capital allocated to the NHS. This is called the export discouragement (or
"disincentive"). As on the other hand, the authorities are concerned with balance of
payments and employment effects, a correction mechanism is applied. So an export-
discouragement correction is deducted from the profit before interest and taxation. A
special set of rules exists for the calculation of the correction.

19


The concepts used in the third and final part of the NHS-profit calculation are:
- negotiated target return on capital, percentage,
- average capital employed (NHS sales), in GBP,
- average capital employed (total), in GBP,
- actual return on capital, percentage,
- excess/shortfall on target, in GBP.
The first step is to calculate the actual return on capital. The size of the acceptable
return on capital earned by individual companies on NHS sales is a matter of
negotiation with the NHS authorities. The negotiated target return on capital depends
on the nature and scale of the company's relevant investment and activities and the
associated long-term risk. Eventually the negotiated target return on capital is assessed
by the authorities.

In order to obtain the actual return on capital first the average capital employed for
NHS sales has to be calculated. The sum of the net-capital employed on the opening
and the closing balance sheet is divided by 2 to obtain the total average capital
employed. The average capital employed for NHS sales is calculated by allocating the
total average capital employed according to one of the allocation methods mentioned
earlier in this section. The actual return on capital (ROC, sometimes also called "return
on investment") is a division of the total adjusted profit by the average capital employed
for NHS sales multiplied by 100. In formula:

total adjusted profit
actual ROC = * 100%
average capital employed (NHS sales)

The excess or shortfall on target can now be calculated by subtracting the negotiated
return on capital multiplied by the average capital employed for NHS sales from the
total adjusted profit. In formula:

excess/shortfall =
total adjusted profit -- [negotiated ROC * average capital employed (NHS sales)]

This final excess/shortfall calculation is not only influenced by the levels of transfer
prices. Changes in royalty payments, changes in intracompany payment terms and
intracompany cost charges also influence the excess/shortfall. This implies that of all
these financing instruments, the impact on the profit calculation needs to be
determined.

In this section we explained the British price-regulation scheme. Their approach to
pharmaceutical pricing is one amongst many, both within Europe and worldwide. In
most countries, Eke in Britain, interaction exists between transfer pricing and
reimbursement. By means of the example given above, we demonstrated what such a
relationship might look Eke. All transfer-pricing issues that relate to the relationship
with the health-care authorities, wiU be referred to as the commercial aspects of
transfer pricing.

20

Sec. 2.5 CONCLUSION

2.5 Conclusion
Three conclusions can be drawn from this chapter.

In chapter 1, we stated that the pharmaceutical industry finds itself in a complex,
hostile and turbulent environment, see Huber [1984b], certainly as far as pricing and
transfer-pricing issues are concerned. In this chapter we elaborated on this statement.

Secondly, great similarities exist in the way European pharmaceutical companies
operate. As the environment of the industry is for an important part determined by
governmental regulations and interest groups, the environment is alike for most
companies. As a consequence, many European pharmaceutical companies are structured
in a similar manner. This similarity is confirmed by the research of Burstall [1985] and
by our own experience, see Van der Ven [1989a]. This conclusion is of relevance when
possibilities for a more general use of the prescriptive conceptual model will be
discussed in chapter 8.

Thirdly, the European pharmaceutical industry can be described by the following
characteristics: a high degree of internationalization and centralization, a high degree
of vertical integration and a low degree of diversification. These characteristics are
relevant for our discussion of the various approaches to transfer pricing in the next
chapter.

21


22

3

TRANSFER PRICING: A COMPLEX ISSUE
3.1 Introduction
Transfer pricing is a subject that seems favorite amongst economists, as an abundance
of literature can be found, see for example Eden [1985], Kassicieh [1985] and
Hirschleifer [1956]. Not only scientists like the subject. Transfer pricing appears to be
a rather preferred choice of masters-degree students in economics, see for example
Mostermans [1982] and Van der Plas [1983]. It struck us that so many of these
publications seem irrelevant for the problem at hand. Although a profound ability in
mathematical modeling is demonstrated, practical relevance for our analysis is small.
As will be discussed in section 3.2, the publication of Rasch [1980] is a notable
exception. In general, scientists are aware of this limitation, see for example
Meijboom [1986].

Transfer-pricing studies that provide a picture of most aspects involved are for example
Verlage [1975], Plasschaert [1979] and Krens [1979]. The relevant aspects can be
divided into two categories. The first category deals with what is defined as
"management-control" issues, see Krens [1979]. Planning of corporate resources and
performance measurement are the two main aspects of management-control transfer
pricing. These will be discussed in the next section. The second category is of
importance in the case of international transfers and is concerned with international
financing, tax issues, duties, foreign-exchange controls and alike. These will be discussed
in section 3.3. Pharmaceutical price-regulation schemes as discussed in chapter 2 are
related to this "international" aspect. In section 3.4 we will apply the various concepts
to the situation in the pharmaceutical industry.

In the remaining part of the thesis we decide to concentrate on transfer prices for
goods. Transfer prices for services are excluded. As will be explained in section 3.3,
this is not to say that payments for services, like management fees and royalty
payments, do not relate to the transfer-pricing problem, see also OECD [1979].

3.2 Management-control transfer pricing
For a better understanding of management-control transfer pricing, we should go back
to micro-economic theory. According to Adam Smith, individuals should pursue then-
own interest and by doing so, they promote the interest of society, see Landreth [1976,
p. 39]. Presuming competitive markets, individuals are led by an invisible hand when
pursuing their own interest, according to Adam Smith. This invisible hand leads
individuals by means of prices for goods, services, capital and labor. As stated in
section 3.1, we concentrate on the prices of goods.

23

TRANSFER PRICING: A COMPLEX ISSUE Ch. 3

A relatively high price for a good that an individual produces and sells, leads to a
relatively high income and hence to relatively high spending power. In other words, the
individual is rewarded by society for offering goods that are relatively scares. Below,
we will relate this aspect to "performance measurement".

This rewarding mechanism influences the decision that the individual makes: he will
plan the use of his resources in such a way that he is able to offer those goods on
the market that are relatively scares. Below we will relate this aspect to "planning".

It should be realized that above, we only discussed external prices. External prices that
are set between independent parties operating on their own behalf on a competitive
market. We now discuss the use of a price-mechanism for use within large
organizations.

According to Chandler [1977, p. 1], modern business enterprises partially "took the
place of market mechanisms in coordinating the activities of the economy... In many
sectors of the economy the visible hand of management replaced what Adam Smith
referred to as the invisible hand of market forces". Cause of this replacement,
according to Chandler [1977, pp. 6-7], is that modern companies "... continued to grow
by setting up or purchasing business units that were theoretically able to operate as
independent enterprises - in other words, by internalizing the activities that had been
or could be carried on by several business units..." From the point of view of the
organization, a side-effect of this growth is an increase in management problems.

Allocating limited resources to alternative uses can be expressed as a mathematical
problem for which various formulations and solutions exist. In case of linear constraints
and a linear target function that expresses company objectives, linear programming can
be used, see Lee and Shim [1986]. As Bosman [1977] and Verlage [1975] explain,
detailed knowledge on all units which constitute the organization is required for
successful use of this technique. All technical parameters, for example, should be
known. If this approach is followed, an optimal solution is presented allocating the
resources to the final products. Given this solution, so called "shadow prices" of the
resources can be calculated. Both Bosman [1977, p. 241] and Verlage [1975, p. 185]
reject the use of shadow prices as a coordination mechanism for large complex
organizations.

An alternative approach to the problem of allocation of resources is divisionalization,
combined with the "right" transfer prices. Divisionalization can be defined as
decentralization plus delegation of divisional profit responsibility to lower organizational
levels at which decision making also exists, see Verlage [1975, p. 5]. Splitting up in
divisions is an attempt to combine many of the advantages of a big company with
many advantages of a small one. Such a system replaces detailed planning and
evaluation decisions by top management. Another possibility which will be discussed
later in this section, is to split up an enterprise into smaller units without strictly
adhering to the profit-center concept. Transfer pricing that is related to
divisionalization issues is defined as "management-control transfer pricing", see
Krens [1979].

24

Sec. 3.2 MANAGEMENT-CONTROL TRANSFER PRICING

Transfer-price decisions have great influence on the distribution of profit of the
divisions involved. Take for example a multinational A. X and Y are two divisions of
A, each with own profit responsibility. The intermediate product z is produced by X
and then sold to Y. When the transfer price of product z is increased, the divisional
profit of X will increase, to the expense of Y's divisional profit. Divisional management
considers the maximization of its own divisional profit as its main target. The divisions
"compete" with each other to obtain the highest divisional profit. As the divisional
management is judged (and compared) on the basis of this divisional profit, a shift of
cost is of great importance to divisional management. So a correct approach to transfer
pricing is a prerequisite to make the profit-center concept work.

According to Verlage [1975], a transfer-pricing policy should serve the same two roles
that are performed by external prices: performance measurement and planning. From
a point of view of profit measurement (ex post), transfer prices must ensure that
divisional profit, as it is measured, is the same as the controllable profit for all
divisions involved. So divisional profit should not be influenced by decisions taken by
other divisions and reflect only the impact of decisions made by the own division. A
transfer price which possesses this quality is defined as "neutral". From a planning
point of view (ex ante), transfer prices must ensure that, when divisional management
bases its decisions on these prices to maximize its own divisional profit, the profit of
the enterprise as a whole is maximized. A transfer price which possesses this quality
is denned as "optimal". Verlage [1975] concludes that a transfer-pricing policy based
on market prices is the only system that meets both requirements. A prerequisite for
such a system, however, is that a competitive market for intermediate products exists.

The concept discussed above deals with a really divisionalized enterprise. We now
turn to another approach. It is possible that decision making in the organization takes
place by means of teams and that no internal competition between divisions exists. This
approach of decision making in groups relates to the "behavioral theory of the firm",
see Verlage [1975, p. 19]. According to the behavioral theory of the firm, company
goals are the result of interaction between organizational members. Rasch [1980, p. 65]
defines a team as a group of people with a common goal. People at different positions
within the organization may have different information and hence may take different
decisions based on that information, but there is no incentive problem. There is no
conflict of interest between team members. Under the condition of team setting, the
transfer-pricing policy does not have an impact on company goals. Rasch's proof of
this statement is of a mathematical nature. Under the premise that the individual
decision maker has knowledge of the impact of his decision on the organization as a
whole, the transfer price is an internal mechanism for the firm and as such it has no
impact on the overall objectives. This result can be anticipated by the realization that
the decision makers are attempting to achieve corporate objectives which are derived
from an overall perspective based on external (market) conditions. The internal transfer
price is merely a means for liquidity transfer and does not affect the performance of the
organization as a whole. The decentralized team-decision policy is independent of
transfer pricing and the determination of a transfer-pricing policy should not be based
on its effect on the decentralized decision-maker's actions, according to Rasch.

25


Rasch [1980, pp. 100-103] even states that, given team setting, the expenses involved
in working out a transfer-pricing policy may be avoided and an increase in efficiency
might be realized.

Several approaches to planning and performance measurement exist. The
"divisionalization" approach discussed above takes divisional profit as the central item
and a guideline for divisional decision making and performance measurement. The
"team setting" approach as described by Rasch [1980], takes the concept of teamwork
as the central means to achieve company goals.

On the basis of empirical research, Eccles [1985, p. 274] relates these two approaches
to company and market characteristics, see also Lorsch and Lawrence [1970].
According to Eccles, the transfer-pricing policy depends on the type of organization
and, consequently, the environmental setting in which the organization finds itself. The
two main factors which determine the transfer-pricing policy according to Eccles, are
vertical integration and diversification.

The term vertical integration refers to the inclusion of activities within the firm that
could be obtained externally, such as production of raw materials, distribution channels
and staff services. Vertical integration creates mutual dependencies between the various
units like divisions, corporate line management and staff functions of the enterprise.

The term diversification refers to the number of different businesses (or markets) in
which the company competes and how different these businesses are from each other.
The larger the number of products and markets in which a company competes, the
greater the diversification, according to Eccles. In exhibit 3-1 three types of company
are shown.
A
high x X
(cooperative) (collaborative)

degree of
vertical
integration

low x
(competitive)
I ^.
low degree of high
diversification
Exhibit 3-1. Eccles' contingency approach to management control.
Of these three types, two are relevant for our research: the cooperative and the
competitive type. These will now be discussed in more detail. The following
descriptions are based upon Eccles [1985, pp. 276-277].

26


The cooperative organization
The cooperative organization is high on vertical integration and low on diversification
and has its focus on a small line of products. This organizational type is often found
in capital-intensive industries. The term cooperative is used to indicate that decision
making within this firm is focused on cooperation among all managers to "maximize"
total company performance. The strategy for this type of company is a strategy for the
company as a whole, with all decentralized strategies being expressed in terms of their
contribution to it. The narrow product focus results in the company being organized
as one single profit center. Management-control systems distinguish functional
responsibilities in terms of costs, revenues, and physical units. Great emphasis is placed
on using internal standards and historical results in evaluating performance. The high
interdependence across the various units (for example, manufacturing and sales)
requires substantially subjective judgment on the part of general management in
evaluating performance, since high interdependence makes divisional profit an imperfect
measure. The processes between superiors and subordinates are very much top-down
because of the focus on a narrow range of products and the interdependence across
units. Horizontal processes for conflict resolution aim at expertise sharing and finding
common interests. In other words, units seek to find a solution that is in the best
interest of the company as a whole. Top management exercises control on the actions
of lower-level managers primarily through the organizational structure. Control over
actions is necessitated by the high degree of interdependence caused by vertical
integration. It is possible to have such control because of top management's experience
in and detailed familiarity with the narrow range of businesses in which the company
is engaged. Because of the high degree of interdependence, lower-level managers are
rewarded in terms of their contribution to the performance of the company as a whole.

The competitive organization
The opposite of the cooperative organization is the competitive organization. The
competitive type is high on diversification and low on vertical integration. Its structure
is based on many different businesses that do little sharing of such resources as
manufacturing facilities and sales forces. This organizational type is exemplified by
conglomerates or holding companies. The term competitive is used to indicate that the
subunits of this organization are profit centers and compete with each other for capital
and other resources on the basis of performance. Performance is comparable across
all subunits, unlike the case of the cooperative organization. Corporate strategy in the
competitive firm is largely an aggregation of subunit strategies, such as those of
divisions. Although broad strategic guidelines might exist for the company as a whole,
they are not nearly as important in developing subunit strategies as they are in the
cooperative organization. Competitive firms employ a multi-divisional structure,
supplemented by systems that measure the profitability and return on investment of the
divisions. As a result of the high degree of decentralization of authority and
responsibility, vertical processes are bottom-up - initiated by lower-level profit-center
managers when they want assistance from top management or corporate staff. The
horizontal processes of conflict resolution are characterized by negotiation, in which
each party attempts to maximize his own share in the context of fixed-sum payoffs. In
other words, profit centers seek to maximize their own results even at the expense of

27


other profit centers. The emphasis is on using financial measures, like divisional profit,
to evaluate performance. Top management must exert control primarily through the
measures of financial outcomes, since it cannot know each of the company's many
businesses in sufficient detail to deal more directly with operating concerns. The
lower-level general managers are rewarded in terms of the performance of their
divisions.

The empirical research undertaken by Eccles [1985] shows that depending on the type
of company (competitive or cooperative) a different approach to transfer pricing exists:

The cooperative type of company often uses full-cost transfer prices. A full cost price
includes both variable and fixed-cost elements. An example is provided by means of
the following formula for the transfer price of intermediate product z:

transfer price = variable cost + x% of fixed cost

The x percent serves to allocate fixed overheads. The percentage of x is determined
by means of dividing the budgeted fixed cost for the coming period by the "normal
size" of the production, see Van der Schroeff and Groeneveld [1980] and
Bosman [1977]. Following Bosman [1977, p. 258], the term "allocation" should in this
respect be interpreted as "a more or less arbitrary apportionment of cost, which results
from planning activities". Transfer prices determined by means of full costing are not
very useful instruments in the area of management-control transfer pricing. Therefore,
little effort is made to identify responsibility for variances. The performance of each.
profit center affects, and is affected by, the performance of others. A full-cost transfer-
pricing approach is neither "neutral" nor "optimal" as defined earlier in this section.

The distinctive characteristic of cooperative kind of companies is the low level of
conflict that exists. This is because of the high level of cooperation among the
managers involved. All are focusing on objectives at a level higher than their own unit.
In all cases, subjective judgment is important in evaluating and rewarding the
performance of these managers, since divisional profit is an imperfect indication of
their contribution to the company as a whole.

For this type of company, Verlage [1975, p. 188] proposes consolidated, and not
divisional, profit measurement to determine the performance of divisional management.
Detailed budgets should be used for planning purposes. This means, in practice, that
profit responsibility is replaced by responsibility for much more detailed budgets, which
are stated in terms of volumes, cost, etc. According to Bosman [1977, pp. 241-242],
these budgets are determined by means of negotiations between top management and
lower management levels. A good management information system is a prerequisite for
detailed budgeting procedures and consolidated performance measurement.

According to Eccles [1985], the competitive type of company uses transfer prices based
upon market prices of the intermediate product that is sold between the divisions.
Inter-divisional transactions only occur when they are agreed upon by managers in both

28


the buying and selling division. The distinctive characteristic of the competitive type of
company is that there are discouragements to trade internally. This problem results
from the pressures placed on profit-center managers to achieve financial objectives and
from their awareness of the difficulties internal transfers could create for them in
inhibiting their ability and authority to achieve these objectives. Because they are
evaluated and rewarded primarily in terms of divisional profit, they have little incentive
to cooperate with other divisions.

The distinction between the cooperative- and the competitive type companies, as
developed by Eccles, is useful for our research. However, as Eccles [1985, p. 273]
states, any actual organization can be composed of several of these types. Relations
between divisions, for example, could be identified as competitive as a high degree of
diversification and a low degree of vertical integration exists. One of the divisions, say
division X, might consist of various decentralized units. Relations between the
decentralized units within division X could be identified as cooperative as a low degree
of diversification and a high degree of vertical integration exists. Consequently, the
transfer-pricing policy should be set in accordance with the different structures in the
different parts of the company. In chapter 6, the Europhar case will illustrate this
issue.

3.3 International transfer pricing
In the previous section the management-control aspects of transfer pricing have been
discussed. In a multinational enterprise, however, another aspect is also of great
importance.

As the transfer price is the price for which the product crosses the national border,
the height of the transfer price has, among other things, consequences for the
international distribution of funds and profits, corporate taxes paid in various countries
and import duties. We define transfer pricing based on these considerations as
"international transfer pricing" to distinguish it from transfer pricing for management-
control purposes. Note that this definition has a wider scope than the term "fiscal
transfer pricing".

While discussing multinationals, the OECD report on transfer pricing [1979, p. 7]
states: "... Such an international economic entity generally operates in a legal framework
which has remained national; it has therefore to conform with the varied and
sometimes conflicting national laws [and other regulations] of the countries in which
it operates." A group of Dutch transfer-price experts concludes that fiscal and other
authorities should be reticent in their requirements towards the multinational because
in the area of transfer pricing it is difficult to qualify prices as "right" or "wrong", see
Vereniging van Belasting Wetenschappen [1987, p. 65]. As an illustration, we now
present some of the issues that may be at stake, see also Vereniging van Belasting
Wetenschappen [1987], OECD [1984], Verlage [1980] and Plasschaert [1979]:

29


- A company in country A bought from its subsidiary in country B at a price which
was too low, according to the customs authorities of country A, and therefore the
price was increased. This increase, however, was not accepted by the corporation-
tax authorities in country A. The company took this decision to court, but the court
decided that the decision of the customs authorities was not relevant to tax on
profit.

- In dealing with the customs authorities of various countries, multinationals may
adapt price setting to the customs regulations. This can, however, lead to different
transfer prices for different countries which in itself can be a source of criticism for
fiscal or other authorities.

- Minority shareholders and/or profit-sharing personnel may require a low transfer-
price level with the aim to maximize own income.

- In countries where, as a result of government interference, loans and other financial
means are hard to obtain, transfer prices might be changed to increase the cash
flow of the division in need. This in order to satisfy already existing local money
suppliers.

- Constraining foreign exchange control regulations may influence transfer prices as
sales and/or purchase of goods might be the single method allowed to transfer
money from one country to another.

- Governments usually try to capture the largest possible slice of the taxable cake. All
in all, the fiscal authorities tend to look critically at business practices which reduce
their portion of taxable revenue but show no interest in questioning the payments
that increase their revenues. Their attitude is asymmetrical which may result in
double taxation.

- In some industries, governments use data about transfer pricing to reduce, say, the
final prices to the consumers and may induce other countries to take similar action
(so called "domino effect") although the market circumstances may widely diverge.
In this process the least favored position for the industry tends to spread amongst
countries.

- Government interference on transfer prices sometimes imposes an inordinate amount
of reporting and red tape on the multinational. Data produced are occasionally used
by civil service bureaucrats without proper understanding of business mechanics.

The different approaches to transfer pricing from various types of authorities (for
example tax and customs) may cause conflicting requirements for the multinational.
As the first example given above showed, problems concerning these regulations may
be difficult, or nearly impossible, to resolve. In the requirements that are to be met
by the multinational, an important distinction can be made: some requirements are set
at the aggregate level and other at the disaggregate level, see also Sol [1985a].

30

Sec. 3.3 INTERNATIONAL TRANSFER PRICING

Corporation-tax authorities might judge a company on total company figures (aggregate
level). Custom authorities might judge prices for individual articles (disaggregate level).

The situation concerning transfer pricing is not only complex for a multinational.
Interpretation of the situation is also difficult for the authorities as "it is an essential
feature of the problem that it is always necessary to have regard for the particular
facts in each case", see OECD [1979, p. 18]. This might be the reason that most
disputes between authorities and multinationals concerning transfer prices are settled
before they reach court. This in itself can be regarded as a proof that transfer pricing
in practice is very difficult, see Verlage [1975, p. 194].

According to Plasschaert [1979, p. 34], it may well turn out that the most acceptable
system is one that minimizes conflicts with external groups, such as governments. Also
Verlage [1975, p. 194] underlines the importance of international transfer pricing. He
states that the situation around transfer pricing creates room for manoeuvering not only
for the multinational, but for fiscal authorities as well. Since in practice, the authorities
will try to use this room for their own benefit, it is necessary for the company to
analyze ways and means of defending its interest.

It will be clear that the problems around international transfer pricing are completely
different from the management-control problems as discussed in the previous chapter.
Transfer prices established for management-control purposes are intended to produce
internal goal congruence between the different parts of the company. International
transfer pricing aims at fitting the transfer-pricing policy within an external framework
of rules and constraints established by legislation and administration, see Quirin [1985,
pp. 132-133].

Over the last two decades, the awareness has increased that an international and well-
defined approach to international transfer pricing is needed. The "arm's length
principle" is one of the concepts which resulted from this approach. The arm's length
concept has originally been developed for fiscal purposes. It will now be discussed
briefly below.

The OECD committee on fiscal affairs [1979, p. 7] developed the arm's length concept
for fiscal purposes. The arm's length price is in their report defined as the price
"which would have been agreed upon between unrelated parties engaged in the same
or similar transactions under the same or similar conditions in the open market". Note
that the definition states: "which would have been". In other words, the concept is in
fact based upon a fiction. The Vereniging van Belasting Wetenschappen [1987, p. 65]
states that "it is more realistic ... to approach transfer-price problems from the point
of view of the multinational as a whole, in which each part has its proper business
function".

In order to assess whether prices are at arm's length or not, the following four
approaches have been established, see Verdoner [1988] and IBFD [1987]: r

31


Comparable uncontrolled price method
The comparable uncontrolled price method offers the most direct way of determining
an arm's length price. The transfer price is set by reference to comparable transactions
under the same or similar conditions between a buyer and a seller which are not
associated.

Resale-price method
There will be many cases where no useful evidence of uncontrolled transactions will
be available, because, for example, the goods which are supplied are so special to
the group that there is no open market for them. This is the case for patented drugs.
In such circumstances it will often be necessary, in order to establish an arm's length
price, to use the resale method. The resale (or sales-minus) method starts from the
fined selling price and subtracts the cost and profit mark-up of the last division
involved. The resale method is useful where it is applied to marketing operations of
finished or semi-finished products, like for a local marketing subsidiary of an
international pharmaceutical company.

Cost-plus method
In cases where neither the comparable uncontrolled price method, nor the resale
method are applicable, the cost-plus method can be used. The cost plus method starts
from the direct cost of providing the goods or services and adds whatever costs and
profit mark-up that are appropriate.

Other methods
The complexity of real-life business situations may put many conceptual and practical
difficulties to the application of the methods referred to above. A mixture of these
methods may sometimes therefore have to be used, sometimes in combination with
other methods. Another such method might be to look at the yield on capital involved
or to determine the contribution of the various units involved ("functional method").

Above, we limited ourselves to prices concerning the transfers of goods. Such transfer
prices, however, are only one of the instruments used for the international transfer of
funds within a company, see Krens [1979]. We explicitly consider international transfer
pricing only one aspect of the problem of international financing. As explained in the
previous chapter, pharmaceutical multinationals are in general highly centralized. Hence
various other means for transferring funds between parent and subsidiary exist. These
are now briefly described below, see Ammer and Ammer [1977].

Transfer prices
Transfer-price payments are the payments from the local subsidiary to the parent
company for goods bought from the parent company by the subsidiary. (As stated in
section 3.1, we limit ourselves to the prices of goods.) These payments are related to
the parent company's role of production center and supplier. (In section 3.5, a
distinction will be made between transfer prices and internal selling prices.)

32


Royalty
Royalty payments are payments from the local company to the parent company in
exchange for the right to use proprietary trade marks, patents or other intangible
assets from the parent company. These payments are related to the parent company's
role of research center and patent holder.

Dividend/share capital
Dividend is the part of a local company's profit that is distributed among the
shareholders in proportion to their share of ownership. In our example, it is a payment
from the local company to the parent company. Share capital is also related to the
ownership of a local subsidiary. It is a payment from the parent company to the
subsidiary company. Payments of share capital and dividend are related to the parent
company's role of shareholder.

Current account
Another financial tie is the so called "current account". The current account is the
ongoing trading-debt position of the subsidiary company due to the purchase of goods.
It is possible to change payment terms for intracompany sales. By decreasing the
creditor's payment term from the parent company, debts have to be paid off earlier
and the creditors account will decrease. Hence the local company's liquidity decreases.
Increasing the payment term would have the opposite effect. Transfers due to changes
in the current account are related to the parent company's role of supplier.

Intracompany loan
A loan is a sum of money, borrowed at interest, for a period of time. The local
company can lend out to the parent company and vice versa. Payments are related to
the parent company's role of internal bank for the multinational group as a whole.

Cost transfer
A cost transfer is a payment from the parent company to the local company or vice
versa. It is based on a specific agreement between the two. An example is a
promotional allowance to compensate for the extra marketing effort of the local
company related to the introduction of a new product. Payment of allowances are
usually related to the parent company's role of management center. In exceptional
cases, the parent company may decide to compensate the legally existing loss of the
subsidiary company. So it is a payment from the parent company to the local
subsidiary. We classify these payments also as cost transfers. They are related to the
parent company's role of shareholder.

Like transfer prices, the size and conditions of the other financing instruments
described above, should be at arm's length, see IBFD [1984]. In the decision process
on local company financing, arm's length considerations play an important role.

The direction of the transfer of funds depends mainly upon the liquidity situation of
the local company. A liquidity surplus is defined as an excess of funds resulting from
operations. A liquidity shortage is defined as a shortage of funds resulting from

33


operations. Some of the alternatives discussed above have an impact only on the
liquidity position of the local company (for example an intracompany loan or capital
increase) and others have an impact on the liquidity position ay well as on the
profitability of the local company (royalty or transfer-price increase).

From the point of view of the subsidiary company, a change in liquidity position can
be due to changes in the profit and loss account or it can be due to changes in
balance-sheet items. As transfer prices are the subject of our research, we shall focus
on those problems for which transfer-price changes are considered a possible solution.
In general these are the situations in which liquidity problems (shortage or surplus) are
caused by changes in the profit and loss account. So we focus on two problem areas:
- a combination of profit and liquidity surplus,
- a combination of loss and liquidity shortage.
When we discuss a liquidity surplus in the remainder of this thesis, a profit is usually
implied. When we discuss a liquidity shortage, in general, a loss is implied. By means
of this limitation, financing of investment programs, for example, is excluded.

Costs of the various ways to transfer funds are of importance. Tax and duty rates
may have a significant impact on costs. To compare the cost of various ways to
transfer funds, a small standard scheme is used. Point of departure of this scheme is
a fixed sum of liquidity that needs to be transferred. This transfer can be either to or
from the subsidiary company. The scheme that we use is quite simple. It is also used
by others with the aim to compute the total net effect of the various tax savings and
tax costs, see for example Comptax [1987].

To illustrate this scheme, we use the example of the company that will be described
in chapter 5. A parent company in the Netherlands, called Europhar International
BV owns a subsidiary company in the United Kingdom (abbreviated as "UK"), called
Europhar UK Ltd. Let us presume a liquidity shortage for the British company.

In setting up these schemes, the central question is: In order to increase the liquidity
position of the UK subsidiary with GBP 100.-, what is the amount in NLG that needs
to be transferred by the parent company ("NLG" stands for Dutch guilder, "GBP"
stands for the British pound sterling)? So we calculate the marginal effect of a net
liquidity increase of GBP 100.--. We presume a fiscal profit or adequate fiscal
compensation possibilities for all companies involved. We define this marginal effect
as the "financial impact" of the financing alternative under consideration. Calculations
are rounded off at two decimals. In exhibit 3-2, the calculation of the financial impact
of a transfer-price decrease is shown ("NL" stands for Netherlands).

34


transfer-price decrease - supplying company is Europhar Int BV:
input parameters:
corporate tax rate UK: 35 X
corporate tax rate NL: 42 I
exchange rate: 3.66 NLG/GBP
calculation of the marginal effect of a net liquidity increase of GBP 100.
liquidity increase: GBP 100.00
UK corporate tax: + GBP 53.85
transfer-price decrease in GBP: GBP 153.85
transfer-price decrease in NLG: NLG 563.08
NL corporate tax: -- NLG 236.50

net decrease of NL liquidity: NLG 326.58

Exhibit 3-2. Financial impact of a transfer-price decrease.

In order to obtain a liquidity increase of GBP 100.--, it is necessary to decrease
transfer prices with GBP 153.85 as lower costs increase profit and hence corporate
tax in the UK. Given an exchange rate of, say, 3.66 NLG/GBP, transfer prices need
to be lowered with NLG 563.08. Due to lower revenues of the parent company,
corporate tax decreases. The net effect is NLG 326.58. This is less than the GBP 100.-
- required as GBP 100.-- equals NLG 366.-.

In exhibit 3-3, the financial impact of an increase in payment term is presented as a
second example.

increase of the credit current-account position with supplier Europhar Int BV:
(payment term is increased)
input parameter:
calculation of the marginal effect of a net liquidity increase of GBP 100. — :

amount of current-account increase: GBP 100.00
net decrease of NL liquidity : NLG 366.00

Exhibit 3-3. Financial impact of an increase in current account.

As the simple scheme shows, no tax effects exists. A third example that we want to
present is the intracompany loan, see exhibit 3-4.

35


intracompany loan from Europhar Int BV:
input parameters:
corporate tax rate UK: 35 %
corporate tax rate NL: 42 Z
interest rate intracompany loan: 10.3 I
of a net liquidity in
interest paid: + GBP 11.04
tax on interest: GBP 3.86
--
amount of intracompany loan (GBP): GBP 107.18
amount of intracompany loan (NLG): NLG 392.26
interest received: -- NLG 40.40
tax on interest: + NLG 16.97
net decrease of NL liquidity: NLG 368.83

Exhibit 3-4. Financial impact of an intracompany loan.

In this example, the amount of the intracompany loan needs to be such that the
interest and tax effects for the first year of the loan are included. In the UK, the
payment of interest causes lower corporate tax. In the Netherlands, the interest
payments cause higher corporate tax. Note that the loan is slightly disadvantageous
as NLG 368.83 needs to be paid to obtain the counter value of NLG 366.--. In this
example, interest payment of one year is included, as within Europhar financing
decisions are reviewed every year. It is possible, of course, to include interest payments
for a longer or shorter period. The interest to be paid by the subsidiary is calculated
by means of the formulas shown in exhibit 3-5.

i - i * (too + i - - T) i * 100
T - t * i 1 -- i * (1 -- t)

1: interest payment
T: tax on interest
i: interest rate
t: tax rate

Exhibit 3-5. Calculation of the interest payments for an intracompany loan.

The tax effects influence the liquidity position with a considerable time lag. This causes
second-order interest effects which alter the financial impact. Also the time of interest
payment causes a second-order interest effect. Under normal circumstances (no hyper
inflation, for example), these net second order effects are relatively small. So for the
determination of the financial impact, we decided to exclude these effects: we presume

36


that all payments of a certain calendar year take place on the first of January of that
calendar year. For the liquidity planning, these delays should be included, of course.
Exhibit 3-6 presents a summary of the results.

summary of financial impact results in case of a liquidity shortage
input parameters:
corporate tax rate UK: 35 X
corporate tax rate NL: 42 X
interest rate intracompany loan: 10.3 %
alternative : net decrease in NLG :
transfer-price decrease Europhar Int BV : NLG 326.58
current account : NLG 366.00
intracompany loan : NLG 368.83

Exhibit 3-6. Summary of the financial-impact results.

This example shows that of the three alternatives, the transfer-price decrease causes
the lowest liquidity decrease at Europhar International BV. The financial impact is one
of the elements which are relevant to select a financing alternative. The three examples
are relatively simple. The only intention is to demonstrate the principle of the financial
impact calculation. For an explanation of other more complex alternatives like dividend
payment, see Van der Ven [1988b].

The financing alternatives described above, all relate to intracompany transfers. One
other alternative, however, is usually evaluated simultaneously with intracompany
transfers. This is the local borrowing/lending alternative:

Local borrowing/lending
Like the intracompany loan, the external loan is a sum of money, borrowed at interest,
for a period of time. This financing alternative is not an internal transfer of funds like
the other alternatives but results from an agreement with an external local bank.

In case of Uquidity shortage in the UK, for example, it might be better to borrow
locally instead of transferring funds from the parent company. The net interest
differences between the Netherlands and the UK need to be added to the schemes
presented above in order to make them comparable with the local borrowing
alternative. By means of the intracompany transfer a real flow of funds takes place.
At the end of a period, the financial impact can be measured exactly at the parent
company as the amount of money has actually decreased (or increased) with a certain
amount of guilders. In the case of using the local borrowing alternative, however, no
real transfer takes place. In other words, a comparison needs to be made by means
of an accounting exchange rate. Although important, we consider a discussion of this
matter outside the scope of this thesis. For more details we refer to Van der
Ven [1988b].

37


Above we discussed various financing alternatives. It should be realized, however, that
a decrease in cost or an increase in sales by means of a temporary rebate can solve
an existing problem in the financing area. All alternatives which primarily have an
external impact, like the two mentioned above, are defined as operational alternatives.
It will be clear that making decisions with regard to these operational alternatives is
the core of the business.

Financing alternatives can be defined as those alternatives which only cause an internal
flow of funds. Local borrowing/lending is also considered a financing alternative. No
flow of goods or operational activities result from implementing a financing alternative.
Financing is a secondary activity within the multinational. The focus is on the
operational activities.

Besides operational and financing alternatives a third category exists: relocation
alternatives. Relocation alternatives cause a shift in activities within the multinational.
So their primary focus is internal. An example of a relocation alternative is the shift
of production or packaging activities from one group company to another. Like
financing activities, relocation activities should support the smooth running of the
business. Hence it is a secondary activity. It will be clear that activities in the
relocation area, like the operational activities, influence the financing situation. We
discuss this distinction between operational, relocation and financing alternatives to
clearly define our "problem space": financing problems only.

On the basis of literature study we categorized the financing alternatives into eight
categories, see Ammer and Ammer [1977]. It should be realized that the number of
existing financing alternatives for a subsidiary company is much larger. From a
theoretical point of view, the number is near infinite. The cost transfer alternative
alone, for example, offers already an extremely large number of possibilities as in
theory each cost item or combination of cost items can be considered for transfer to
or from the parent company.

In this section, we globally described the issues at stake in international transfer
pricing. The main conclusion is that international transfer pricing is a part of the
problem of international financing, so it cannot be considered in isolation.

3.4 Transfer pricing in the pharmaceutical industry
In chapter 2, we concluded that high vertical integration and low diversification are
dominating characteristics of the pharmaceutical industry. As explained in section 3.2,
this indicates that we deal with a cooperative type of company in which decision
makers act as a team with a common goal, see exhibit 3-7.

38

Sec. 3.4 TRANSFER PRICING IN THE PHARMACEUTICAL INDUSTRY

A
high x
pharmaceutical companies:
cooperative type

degree of
vertical
integration

low

low degree of high
diversification

Exhibit 3-7. The place of pharmaceutical companies in Eccles' contingency approach
to management control.

Consequently, no transfer-pricing policy is needed from a management-control point
of view. Other mechanisms, like detailed budgets and consolidated performance
measurement are used as a steering mechanism (ex ante) and for performance
measurement (ex post).

According to Rasch [1980], team setting implies that a transfer-pricing policy, whatever
the kind, does not influence corporate performance as decisions are in line with
company goals anyway on the bases of other steering and control mechanisms. As
explained, a transfer-pricing policy is not useful and even inefficient in this case,
according to Rasch.

Although many, if not all, pharmaceutical companies belong to the cooperative type,
the conclusion of Rasch would be completely false. As explained in the previous
section, the international aspects result in high "compliance costs" of all kinds of
regulations, see Quirin [1985, p. 148]. The interference of health-care authorities only
adds to this point. Although there is no need for transfer pricing for management-
control purposes, there is a great need for a transfer-pricing policy to respond to
sometimes contrasting and continually changing regulations. Not only regulations from
fiscal and customs authorities, but also regulations from central banks, health-care
authorities and other institutions are involved. In his analysis of a large Swiss-based
pharmaceutical company, Von Grebmer [1987a], reaches the same conclusion. He
stresses the problems resulting from cost containment measures as described in the
previous chapter. As stated in chapter 2, all aspects relating to pharmaceutical-pricing
and cost-containment issues will be referred to as the "commercial aspects".
International transfer pricing deals with the financing, fiscal and commercial aspects of
transfer pricing.

In summary the position of the pharmaceutical industry can be presented in exhibit
3-8. Other branches of industry might find themselves in other positions of the scheme,
as the example of "company A" illustrates.

39


Let us take a competitive type of company, named "company A", whose divisions are
located in various countries. Company A might score very high on the "management-
control axis" as the company is of the competitive type. It might also score high on
the "international axis" as the internal transfers between divisions cross national borders.
That company A scores less on the "international axis" than a typical pharmaceutical
company, might be caused by the absence of commercial aspects.

A
very
important x
pharmaceutical companies

international x
transfer company A
pricing

not
important
I ».
not management-control very
important transfer pricing important

Exhibit 3-8. The importance of various kinds of transfer pricing for the
pharmaceutical industry.

So far we reached this conclusion by studying the existing transfer-pricing literature and
publications concerning the pharmaceutical industry. In addition to this we conducted
a survey among 12 transfer-pricing experts within the European pharmaceutical industry
itself, see Van der Ven [1989a]. The outcome of this survey was as follows:
- management-control aspects of transfer pricing are not of main importance to
pharmaceutical companies,
- international aspects of transfer pricing are of great importance to the industry and
they are an area of major concern.
Our point of departure is supported by the point of view of these experts. Based on
these findings, we will now concentrate on international transfer pricing.

3.5 A framework to describe transfer pricing
In order to facilitate fast communication and better understanding of the problem,
we developed the following framework to describe a multinational:

Our point of departure is the so-called object company. The analysis of the financial
ties between the parent company and one of its subsidiary companies is the focus of
attention. This subsidiary company is defined as the object company. Besides the
financing area, these financial relationships have, among others, an impact in the fiscal
and commercial area. Various categories of financing instruments, like transfer prices,
royalties and loans, need to be considered simultaneously. Decision making requires the

40

Sec. 3.5 A FRAMEWORK TO DESCRIBE TRANSFER PRICING

involvement of local and central management and fiscal and financial experts. In the
decision process, we try to solve the financing problems (liquidity surplus or liquidity
shortage) of the object company.

This object company purchases its products from a source company. By definition, a
source company is a company which belongs to the same multinational and which
sells goods to the object company. As shown in exhibit 3-9, the parent company is a
supplier of the goods and hence a source company. The sales of these goods are
defined as "transfers".

parent
company source company

tran sfers
'
subsidiary
company object company

external .sales
'
local
wholesaler destination

Exhibit 3-9. The basic pattern between source company, object company and
destination.

Destinations are all companies that purchase from the object companies. Usually the
destinations are wholesalers from the country in which the object company is situated.
In most cases, the wholesalers are taking care of further distribution of the articles and
assist in the local marketing activities.

As shown in the exhibits 2-1 and 3-7, strong vertical integration between parent and
subsidiary is our point of departure. Conceptually, this remains so. In practice,
however, the parent company may not be the only source company. In addition it is
possible that the object company sells to destinations belonging to the same
multinational. Although amounts sold and purchased may be of minor importance, say
10% of the amount bought from the parent and sold in the local market, these
additional ties may seriously complicate the situation. The relationships within the
multinational are shown in exhibit 3-10.

41


external sale of
destination article destination
third
(external selling price)

destination

transfer of internal sale of
source source article object destination article destination
company company company
(transfer price) (internal selling price)

reference article reference
reference price)

Exhibit 3-10. Relationships within a multinational.

The price that the object company must pay for a source article purchased from a
source company is defined as a transfer price. The object company can resell the
articles to a destination within the multinational or to a destination outside the
multinational group (a "third party"). A destination within the multinational group is
defined as a destination company. The price a destination company has to pay for a
destination article is called an internal selling price. When the destination is a third
party, a destination third, this price is defined as an external selling price. Note the
distinction we make between transfer prices and internal selling prices. Usually, these
two words are synonymous. Within our conceptual framework, however, the transfer
price is the price that the object company pays to the source company and the internal
selling price is the price that the object company receives from a destination company.

Although we concentrate on one object company, its financing problems cannot be
considered in isolation. The authorities of the country of the object company determine
regulations with regard to transfer prices. These authorities influence, or are influenced
by, authorities of other countries. These countries are defined as reference countries.
Related companies situated in the reference countries are defined as reference
companies. Articles sold to these companies are defined as reference articles. Transfer
prices for these reference articles are defined as reference prices. All issues which relate
to these cross-border impacts of transfer prices, are referred to as reference effects.

Decision making with regard to international financing is a secondary activity. It has
to react to actions that take place in the operational and relocation area. Therefore
the time span cannot be too long: frequently the international-financing situation needs
to be adjusted to changes which take place in the operational or relocation area. In

42

Sec. 3.5 A FRAMEWORK TO DESCRIBE TRANSFER PRICING

international transfer pricing, chief specialists and general management, just under the
top level, are usually involved. Following the three levels of organizational decision
making as discerned by Bosman [1983], transfer pricing is not classified as a strategic
top-level problem but as an issue that is handled at the administrative - intermediate -
level.

Solving the financing problems of an object company is a multi-disciplinary process
as, amongst others, financial, fiscal and commercial aspects are relevant. Reference
effects are of importance as well. Within the multinational, various organizational
members need to be involved to provide the necessary expertise. In many companies
we find that the following units are represented: local management, central
management, price coordination, fiscal affairs and financial affairs. As within this
industry no management-control aspects are involved in decision making, no transfer-
price negotiations need to take place. Given the strong outside pressure and the
hostility of the environment, we presume that this group acts as a team. This point of
departure of team setting is supported by the small survey that we held within the
industry, see Van der Ven [1989a].

3.6 Conclusion
Based on the previous sections, we conclude the following:

Transfer pricing in the pharmaceutical industry focusses very much on international
aspects. Management-control aspects are of minor importance as other instruments, like
budgets and consolidated profit measurement, are used for planning and performance
measurement.

International transfer pricing is an integrated part of the financing activities of the
multinational. Financing of one subsidiary company, the object company, is focus of our
attention. From a theoretical point of view the number of financing alternatives is
infinite. These financing alternatives can be classified into the following eight categories:
- transfer prices,
- internal selling prices,
- royalty payments,
- cost transfers,
- capital/dividend payments,
- intracompany loans,
- current account changes,
- local borrowing/lending.
These instruments have different characteristics. Consolidated cost, for example, may
differ per alternative due to different tax effects. The consolidated cost are defined as
the "financial impact". Usually a combination of various instruments from the eight
categories mentioned above is used to solve the financing problems of a local
subsidiary company.

43


Like relocation activities, financing is a secondary activity within the multinational.
Operational activities are the core of the business.

Selection of the financing instruments is based on fiscal, commercial and financing
considerations both on the aggregate and disaggregate level. The arm's length concept
plays a dominant role in the decision process. In the multi-disciplinary decision process,
the following organizational groups are represented: local management, central
management, price coordination, fiscal affairs and financial affairs. It is realistic to
presume that this group of decision makers acts as a team.

Before turning to the next chapter, we discuss the contribution of our approach to
transfer-price research.

Verlage [1975] is one of the first authors who defines transfer pricing explicitly as a
multi-disciplinary problem. Even though we concentrate on the international aspects
alone we are still confronted with the multi-disciplinary character of transfer pricing:
a transfer price should be acceptable from a commercial, fiscal and financial point of
view. The rapport of Vereniging van Belasting Wetenschappen [1987, pp. 16 and 20]
states that "the enterprise should choose a pragmatic approach" and "transfer prices
used in reality need to be a compromise" of all aspects involved. In the final chapter
of his dissertation, Verlage proposes the set up of a headquarters staff department
which concentrates on integration of the aspects mentioned above. This department
should managed by a so-called "price coordinator". Concerning this department,
Verlage [1975, p. 197] states the following: "It is obvious that this department cannot
operate without an adequate knowledge of the following subjects:
- the accounting system of the enterprise ...
- customs regulations and fiscal law ...
- financial policy
- commercial policy ...
It is clear that this knowledge is most likely to be found in financial management and
transfer pricing should be a part of the authority of the financial director."

In his proposal, Verlage states that the integration of the various aspects and the
search for a "pragmatic compromise" is the responsibility of the price coordinator. He
does not work out, however, how the integration of these aspects should take place.
It is in this area that we intend to contribute to transfer-price research: on the basis
of developments in the realm of decision support, as will be discussed in the next
chapter, we will propose a systematic integration of all aspects involved.

44

4

GROUP DECISION SUPPORT SYSTEMS
4.1 Introduction
As explained in the two previous chapters, the pharmaceutical industry operates in a
hostile, complex, and turbulent environment, especially as far as transfer pricing is
concerned. In chapter 1, we explained that the decision-making paradigm is appropriate
when studying and designing organizations which operate under such "post-industrial"
conditions. Huber [1984b, pp. 934-939] reviews various aspects of the decision-making
paradigm. The requirement that decision making in post-industrial organizations needs
to be more frequent, faster, and more complex creates decision-tasks which are
qualitatively greater than those of the past. Awareness of this fact provokes and
promotes developments both within organizations themselves and within the
organizational-science community. Huber expects these efforts to be successful and that,
in order to enhance their decision processes, post-industrial organizations will adopt
on a widespread basis three design features: advanced computing technologies,
improved decision-group technologies and decision-process management.

Advanced computing and communication technologies
Huber expects that communication technologies, electronic mail systems and other
devices will become extremely user friendly. The consequent increased accessibility to
people, increased efficiency of communication, and increased timeliness of
communication (all much more important in the faster-paced environment of
post-industrial organizations) will cause post-industrial communications to be adopted
on a scale not greatly different from that of the telephone today, according to
Huber [1984b].

With regard to computerized aids in decision making, Sol [1987a] and Koppelaar [1988]
expect the addition of artificial intelligence techniques to decision support systems and
office automation. According to Boot and Koppelaar [1982], artificial intelligence can
enable computers to process natural language. Sprague and McNurlih [1986, pp.
363-385] expect extreme user friendliness to become a common characteristic of
computerized systems. They mention the extensive use of graphics, speech recognition
and pull down-menus.

Improved decision-group technologies
Increased environmental complexity leads to the need for more information exchange,
see Galbraith [1977]. Often this exchange takes place at meetings between staff experts
and line management. An important point to note is that the organizational members
meeting may not make the ultimate decision. They may be creating and/or reviewing
alternatives to be submitted as a short list to the next level in the organizational
hierarchy, see Gray [1986, p. 158]. Given higher levels of environmental complexity,
there will be pressure for the number of meetings in post-industrial organizations to
be greater. This pressure will be resisted strongly, however, since the managerial time

45

GROUP DECISION SUPPORT SYSTEMS Ch. 4

available for meetings may be approaching its limits and because meetings are widely
regarded as less than optimal uses of time. What is to be the resolution of this
situation?

As one approach, Huber [1984b] expects behavioral scientists and individual
organizations to develop and implement additional technologies for increasing the
efficiency and effectiveness of decision-oriented meetings. Even though a good deal of
development work has already been done, and some adoption of decision-group
technology is emerging, in the post-industrial organizations the increased need to
exchange information leads to much higher density of application of such technologies.

As a second approach, Huber expects that an increase in decision-group effectiveness
will be achieved by integrating communication and computing technologies into
decision-group technologies. This has already occurred in a rudimentary way in the
form of teleconferencing and video-conferencing. Face-to-face groups are made more
effective with marriages of communication and computing technologies and
behavioral-based technologies. Huber presents the nominal group technique as an
example: each participant writes his or her ideas on an electronic pad and then
transfers them to a "public screen" with the touch of a "send" button. Ratings are
forwarded and compiled electronically and displayed so that discussion can be directed
in more fruitful channels, according to Huber.

Huber concentrates on the cognitive processes and information exchange within decision-
making groups. This includes collecting and evaluation of information, forging
alternative courses of action and selecting one as preferred. Huber's focus is not on
the social-psychological dynamics of group behavior, see also Guzzo [1982, p. 4] and
Cook and Hammond [1982, pp. 13-39].

In summary, given the apparent need for more decision-group meetings and at the
same time a considerable resistance to them, we expect post-industrial organizations
to seek and adapt on a widespread basis more sophisticated group technologies and
that, as a result, the effectiveness of decision groups will increase.

Decision-process management
According to Huber, the demand for faster decisions will cause logistically-related
delays in decision processes to be considerably less tolerable in post-industrial
organizations. In addition, the need to make more complex decisions will require
organizations to ensure participation from a wider variety of experts and managers.
Together these needs will motivate post-industrial organizations to adopt more
structured approaches to "decision-process management".

Although the implementation of more formal decision processes is new, it is so similar
to the concepts of project management, that Huber [1984b] expects it to be readily in
operation. Both Bosman [1987a] and Sol [1985b] consider decision-process structuring
a prerequisite for providing effective decision support. According to Huber, it is likely
that as decision making receives more explicit recognition, decisions will increasingly
be viewed as ad-hoc "projects". This can be explained as follows: As situations change

46


rapidly, it is impossible to install fixed bureaucratic procedures to react to stimuli from
outside the organization. On the other hand, to assure the quality of the decision
process, the involvement of several organizational members, like managers and experts,
is required. The timeliness of the decision process requires this involvement to be
without delay. This will lead to the adoption of project-management technologies for
managing the production of lower-level decisions. For example, at the planning stage
the use of planning networks would facilitate identification of:
- needed activities such as obtaining particular information,
- needed resources such as experts, computer programs, authorizations from top
managers, etc.,
- precedence and temporal relationships among these activities, thereby contributing
to the scheduling of meetings, people and tasks.
According to Huber, it appears that process-management technologies such as these
could be readily transferred to the organizational-decision context. Future demands
could lead to the development and use of technologies specifically designed for
decision-process management. This does not mean, however, that decision making
should take place in a more formal atmosphere: in some situations, the decision
process may be managed and structured to create an informal, adaptable and flexible
way of decision making, see Huber and McDaniel [1986, p. 583].

To summarize, we can state that post-industrial organizations should include, among
other design features, advanced communication and computing technologies, improved
decision-group facilities and decision-process management as responses to the
post-industrial environment's demand for more frequent, faster, and more complex
decisions.

We agree with Huber however, that we do not expect any one particular feature, any
one particular technology, structure, or process, to be lasting. Opportunities and
problems, and means for dealing with them, will change. The appropriateness of any
particular organizational-design feature will change, often quickly. So innovation in
organizational design and organizational decision support is, and will be, called for.

And innovation is the aim of our research. In this section we developed the following
broad design guidelines, related to the decision-making paradigm:
- the use of advanced communication and computing technologies,
- improved decision-group technologies,
- decision-process management.
These three features form the basis for designing the prescriptive conceptual model.
The issue of decision-process management is closely related to structuring decision
processes. This will be discussed in the next section.

4.2 Structuring decision processes
According to Ginzberg and Stohr [1982, p. 20] "a good decision support system brings
as much structure as possible to the problem". Bosman underlines the importance of

47


structuring decision processes. When discussing the design of information systems,
Bosman [1986, p. 315] states that the distinction into "phases of the decision process
is for both the description and the design of vital importance". If we are to support
a decision process, it is essential to structure it in terms of tasks, decisions and
information needs. Hence the need for an approach that accommodates for the
apparent paradox of structuring an ill-structured problem.

In this context, the approach suggested by Bots and Sol [1987] is very promising.
Their method of task analysis provides both the structure and the flexibility required
for the modeling of our problem. Its basic concept is task. The actual structurmg of
the decision process is accomplished using tasks. A task consists either of decisions
only or of decisions and other more detailed tasks (subtasks). In the latter case, the
decisions have a coordinating function in the sense that their outcome determines
whether a specific subtask has to be performed or repeated. Hence the term
"coordinating decision". At the most detailed level, tasks are specified in terms of
decisions and information, where each decision requires information (input) and
generates information when it is made (output).

According to Simon [1977, p. 41], problem solving comprises four principle phases:
intelligence, design, choice and review. These four phases are defined as follows:
1. intelligence: searching the environment for conditions calling for decision,
2. design: inventing, developing, and analyzing possible courses of action,
3. choice: selecting a particular course of action from those available,
4. review: assessing past choices.

Huber and McDaniel [1986, p. 576] discern the following phases in problem solving.
(A set of related problems is regarded as one problem):
1. sensing the problem,
2. exploration of the problem,
3. definition of the problem,
4. generation of alternative solutions,
5. evaluation of alternative solutions,
6. selection of alternative solutions.

Various other approaches exist. For a more detailed review of decision-making
literature, see Van Schaik [1988]. In this section, we limit ourselves to the approaches
of Simon and Huber and McDaniel as presented above. For the most recent
developments, see Simon et al. [1987]. We will use the word "alternative" in the
meaning of "alternative solutions" and "particular course of action".

Neither Simon nor Huber make a distinction between "problem solving" and "decision
making". Both activities are aimed at removing the decision maker's dissatisfaction with
his present state, the problem, see Sol [1982], where he can identify:
- one or more outcomes he desires,
- two or more different courses of action,
- an environment containing factors that affect the outcomes.
Note that opportunities are included in this definition of a problem.

48

Sec. 4.2 STRUCTURING DECISION PROCESSES

Before the problem is solved, however, the solution needs to be implemented.
According to Simon [1977, p. 43] executing decisions is again a decision-making
activity, but now at a more detailed level. In order to make a distinction between these
two levels of detail, we define "problem solving" as encompassing the following tasks:
1. scan task,
2. decision-making task,
3. authorization task,
4. implementation task.
Following Bots and Sol [1987], we prefer the term "task" instead of the word "phase"
as the research of Witte [1972] shows that although decision processes can be
delineated in distinct phases, no simple sequential relationship exists between them.
Simon [1977, p. 43] puts it as follows: "The cycle of phases is, however, far more
complex than this sequence suggests".

The relationship between these tasks and the phases as defined by Simon and Huber
will now be explained.

Scan task
We define the scan task as the ongoing search for occasions for making a decision.
It is similar to the intelligence phase of Simon and the sensing of problems of Huber.
During this task, problems are noticed and identified.

Decision-making task
The decision-making task comprises all steps between identification of the problem
and making a decision. In the previous section, a distinction was made between making
a decision and authorizing a decision. We now define a decision as a, yet unauthorized,
concrete plan for action with the aim to remove the problem. The decision-making task
comprises all phases as described by Huber, except sensing the problem. Included are:
exploring the problem, definition of the problem and generation, evaluation and
selection of alternative solutions. Below we will come back on the tasks that constitute
decision making.

Authorization task
We define the authorization task as the formal acceptance of the decision by the
organizational members who are ultimately responsible for finding and implementing
a solution.

Background of the inclusion of a separate authorization task, is our focus on group
decision support. An individual who takes the responsibility for his own decisions,
does not discern a separate authorization task. In the case of an individual, making a
decision implies at the same time authorizing that decision. In case of organizational
decision making, however, it is possible that a group of experts and managers decides,
in the sense as defined above, but that only one of the group, the general manager
for example, has the power to authorize. The position of the authorizer might also be
at a higher level in the organization.

49


We would put the authorization task after the last phase (selection of solutions) of
Huber. So authorization is not included in Huber's definition of problem solving, see
also Huber [1984a, p. 196]. Following the example of Simon [1977, p. 42] on the
decision to introduce computers in an organization, we would conclude that according
to Simon, authorization is a part of the choice phase.

Implementation task
We define the implementation task as the execution of the authorized decision with
the aim to remove the problem. Like authorization, implementation is not included
in Huber's definition of problem solving. As said above, Simon considers
implementation as a new decision-making task, so it is not included in his definition
of problem solving and/or decision making.

Above, the four tasks which constitute problem solving have been briefly discussed.
According to Huber [1984a], the decision-making task should be the focusx of attention
as this is the main element of the problem-solving task. If we consider the problem
of transfer pricing as discussed in the previous chapter, we note that the scan,
authorization and implementation tasks do not give rise to substantial problems. This
is in contrast to the decision-making task. Following Huber, we will mainly concentrate
on the decision-making task, see also Bui and Jarke [1986, p. 84]. Within the decision-
making task, the following tasks are defined:
1. preparatory design,
2. integrative design,
3. choice.
These tasks are now described in more detail.

Preparatory-design task
The preparatory-design task resembles the scan task as the main aim is to gather
information. The distinction between the preparatory-design task and the scan task is
that preparatory design is focussed on a specific problem. During a limited period of
time, detailed information is obtained on the specific problem at hand and on the
solutions available. This specific problem may still lack accurate description. In contrast
to the scan task, the preparatory design task is not an ongoing process aimed at
spotting relevant problems. We define the preparatory design task as the analysis of
the relevant aspects of a specific problem and solutions related to this problem during
a limited period of time.

This task can be regarded as a part of the design phase as defined by Simon.
Following Huber and McDaniel, this task encompasses the exploration and definition
of the problem and generation of alternatives. Focus during this task is on exploration
of the problem. Definition of the problem takes place as far as required in order to
determine the focus of the preparatory design.

The limited period of time set apart for the preparatory design may be a number of
weeks. Time spent depends mainly on the urgency of the problem at hand and the
capacity available. As explained in chapter 3, from a theoretical point of view, the

50


number of alternative solutions is near infinite. In chapter 1 we explained that decision
makers do not possess unlimited cognitive abilities. Therefore the restricted period of
time available limits the number of alternatives generated.

Integrative-design task
On the basis of the information gathered during the preparatory design task, integrative
design takes place. It encompasses the same tasks as defined by Simon and Huber:
exploration and definition of the problem and generation of alternatives. Focus is on
the definition of the problem and generation of alternatives. We define the integrative
design task as:
- definition of the problem and the integration of all aspects as determined during
the preparatory analyses,
- generation of alternatives on the basis of the preparatory analyses.
Note that this definition of design is more limited than the definition of Simon.
Simon's definition of the design phase includes the preparatory design as described
above. The background of splitting up the design phase as defined by Simon in two
different tasks is our focus on group decision support: it might be more efficient to
perform the preparatory design task individually. Integrative design, however, requires
the involvement of all decision makers. For the integrative design task we can use the
term "integral decision making" as all decision makers consider the relevant aspects
simultaneously.

Choice task
Following Huber and McDaniel, we define the choice task as the evaluation and
selection of alternatives. Note that our definition of choice is more limited than
Simon's definition, as he includes the authorization task. Ways to structure the choice
task, like multi-criteria techniques, are discussed later on in this chapter.

Above we described the various tasks. Using the technique of Bots and Sol [1987]
we now present the visualization of this task structure. As explained above, a task
consists either of decisions alone or of decisions and other tasks. In the latter case,
the decisions have a coordinating function in the sense that their outcome determines
whether a specific more detailed task has to be performed or repeated. Hence their
definition: "coordinating decision". At the most detailed level, tasks are specified in
terms of decisions and information (input and output). Note the distinction with regular
flow diagrams. In this thesis, we do not present task-structures at the most detailed
level. Our research served as a case study in testing the method of task analyses. This
study is reported in detail by Bots [1989], who presents task structures at the most
detailed level.

In the exhibits presented below, coordinating decisions are depicted as small circles
and tasks as rectangles. The arrows indicate the order of the tasks. In case tasks of
a detailed level (subtasks) are depicted, the "path" is indicated. The path shows the
hierarchy of tasks. As an example we take task A. Task A is an element of task B
and in it's turn, task B is an element of task C. The path of task A is indicated as
follows: "path: C B A" (Personal computer users will note that the path indication

51


resembles the directory-path indication of MS-DOS). A task finishes whenever one of
its subtasks finishes that has no outgoing arrows, or when one of its coordinating
decisions is made in such a way that none of its outgoing arrows can be followed.
Unlike regular flow diagrams, task structures have no explicit stop symbols. When a
task is finished, one has to look at its parent task to determine what should be done
next. The name of the parent task can be derived from the path of the subtask.

O problem worth
analyzing?

yes v

decision
making

authorization

O alternative
authorized?

yes ,,

implementation

path: problem solving

Exhibit 4-1. Proposed structure of the problem-solving task.

52


The technique is used to describe both the decision-making task as well as the
problem-solving task. As explained above, we focus on the decision-making task. The
problem-solving task provides a reference for the decision-making task. Using this
technique, the problem-solving task looks as presented in exhibit 4-1.

.preparatory
design

yes f
O further
analysis
required?

Integrative
design

yes f
O further
analysis
required?

choice

path: problem solving decision making

Exhibit 4-2. Proposed structure of the decision-making task.

Note the two coordinating decisions in exhibit 4-1. The first one determines whether
it is justified to start the decision-making task. The decision-making task requires a lot
of time of the participants. Often, many problems which all require decision making
will come up simultaneously. During this coordinating decision, priorities are set. The
second coordinating decision checks whether the higher management levels indeed
authorized the decision, as requested. If not, a new iteration of (a part of) the
decision-making task needs to take place.

53


For the task structure of the decision-making task, see exhibit 4-2. As the decision-
making task constitutes a part of the problem-solving task, the path is indicated. The
two coordinating decisions check whether all necessary information is available.

4.3 Multi-criteria decision making
Another issue related to structuring decision processes is "multi-criteria decision
making". These techniques contribute to structuring the choice task. In cases of
expertise sharing, like in the case of international transfer pricing, these techniques may
especially be useful, see Bui and Jarke [1984]. Aim of these multi-criteria techniques
is to combine the following two elements, see Lootsma [1988]:
- an objective evaluation of the alternatives with regard to the relevant criteria,
- a subjective assessment of the importance of the criteria used in the evaluation.
It will be clear that the evaluation of alternatives can only be completely objective in
some exceptional cases. The main aim of the multi-criteria techniques is to improve the
decision-process structure. Many different approaches, simple and sophisticated, exist,
see Roy [1985], and Zeleny [1982].

In the introduction of the previous chapter, we discussed the gap that exists between
sophisticated mathematical models and every-day practice in the area of transfer
pricing. Like in the area of transfer pricing, also in the area of multi-criteria
techniques, such a gap exists. Fortuin and Lootsma [1985] define this as the
"practicality gap". To avoid this practicality gap, we set the following three criteria for
selecting a technique for the choice task:
- The complete technique should be understandable for all decision makers, even if
they are non-mathematicians. This implies, for example, that logarithmic scales may
be considered undesirable.
- The technique should be simple to use. This is in order to avoid long cumbersome
procedures during the meetings.
- The technique should offer a good point of departure for discussion within a group,
see Lootsma et al. [1986].
By setting these requirements, we will avoid the choice task being choked by the
technique. The three criteria presented above determine some aspects of the design
guidelines that will be used to construct the prescriptive conceptual model.

The technique that we have selected is now described below by means of an example.
It is a simple approach based on some elements as used in "Maud". Maud is a
computer package for multi-criteria decision making as designed by the University of
London, see DAU [1986a; 1986b].

Let us presume that we need to choose one out of three financing alternatives. These
financing alternatives are: a transfer-price increase, a loan and a dividend payment. The
criteria that we consider relevant are "fiscal effect", "financial impact" (cost) and
"flexibility". We define the "score" of an alternative on a specific criterion as the degree
that this criterion is met by the alternative. These scores are expressed by means of
figures in the range from one to seven. The meaning of these figures ranges from "very

54

Sec. 4.3 MULTI-CRITERIA DECISION MAKING

unfavorable" via "indifferent" to "very favorable". So the higher the score, the better the
alternative.

The second part of our multi-criteria technique is the determination of the weights.
For the three criteria, the relative importance of each criterion needs to be
determmed. This is done by attaching a number to each of the criteria. For
convenience sake, these weights are normalized, see exhibit 4-3.

weight (C(i))
normalized weight (C(i)) - * 100%
n
£ weight (C(i))
1-1
n - number of criteria
C(i) E criteria, 1 s i s .n

Exhibit 4-3. Formula for normalized weight.

Say that we weigh "fiscal effect" with 15, "financial impact" with 5 and "flexibility" also
with 15. The next step is to integrate the weights and scores into one total score per
alternative. The total score per alternative is defined by means of the formula as
presented in exhibit 4-4.

total score A - £ {score (A, C(i)) * normalized weight (C(i))}
i-1
n - number of criteria
A 6 alternatives
C(i) E criteria, 1 s i s n

Exhibit 4-4. Formula for total score.

This formula calculates the weighted average score of the alternative on each criterion.
For the complete table with weights and scores, see exhibit 4-5.

alternatives:
weight normalized transfer loan dividend
criteria: weight price
fiscal effect 15 43% 1 6 A
financial impact 5 14% 6 4 5
flexibility 15 43% 7 3 5
35 100% 4.3 4.4 4.6

Exhibit 4-5. Table with total scores.

55


In exhibit 4-5, the total scores are rounded off at one decimal. As we can see, the
dividend has the highest total score and hence the dividend alternative is selected.
However, the method should be used with the utmost care as marginal differences can
cause a different outcome. A score on "flexibility" for the dividend of 4 instead of 5,
would make the loan the favorite alternative. For more details on the effect of the
technique on the choice of the best solution, see Claessens and Vogt [1988]. In then-
comparison of various multi-criteria techniques, they used real-life scores from our
research project in order to demonstrate the differences.

We realize that the multi-criteria technique that we decide to use is very simple
indeed. Other techniques that we studied, like the method of Israels and Keller, see
Albers [1987], however, would not contribute to communication within the group due
to their complexity. It is our impression that decision makers without a mathematical
background will be put off by "dominance matrices" and "s-scores". These technicalities
would distract the attention from the problem at hand and the intragroup
communication. We realize that these complex approaches may be considered more
"elegant" and theoretically superior to the technique that we use. In our case, however,
we consider a simple technique the most appropriate.

The use of multi-criteria techniques has a significant impact on the way that decisions
are made. For one thing, evaluating and choosing alternatives becomes much more
explicit. All combinations of criteria and alternatives get a visible score. In case of
group decision making, such an approach leaves little room for so called "hidden
agendas" (or "hyper games") of the participants, see Lootsma [1988]. In case these
techniques are included in a decision support system, such a system changes the way
decisions are made. So it does more than "supporting" decisions. In fact, the decision
process changes. For this reason, Lootsma [1988] proposes the term "decision reform
systems" instead of decision support systems.

4.4 Approaches to group decision support systems
As explained in the previous section, international transfer pricing is a group decision
process. Davis [1969, p. 2] defines small groups as groups with a dozen or fewer
people. As less than twelve organizational members are involved in international
transfer pricing, we now discuss group decision support for small groups. Like Cook
and Hammond [1982] and DeSanctis and Gallupe [1987, p. 591] we abstract from
psychological interaction within the group and concentrate on information exchange and
problem understanding. This point of departure is related to the presumption of team
setting as discussed in the previous chapter. We focus on "expertise sharing groups"
We exclude negotiation support systems, see for example Jarke et al. [1987] and Jarke
and Jelassi [1986]. Expertise sharing involves the resolution of conflicting viewpoints,
so called cognitive conflict tasks, see DeSanctis and Gallupe [1987]. Resolution of
conflicting motives or interests, however, is excluded.

We focus on information exchange and decision making within the group. In other
words, the focus is very much on the problem. This problem-oriented approach of

56

Sec. 4.4 APPROACHES TO GROUP DECISION SUPPORT SYSTEMS

participants is best helped by means of an informal atmosphere, see Schnelle [1977].
It is our impression that this informal atmosphere is more important than the
"executive feel" with "plush carpeting" and "quality furnishings" that decision rooms
should have, according to Gray [1986]. For more details on motivational aspects of
decision making, see March and Simon [1958] or Davis and Olson [1985].

Following DeSanctis and Gallupe [1985] we define group decision support systems as
follows:

An interactive computer-based system which facilitates solution of ill-structured problems
by a set of decision makers, working together as a team.

To facilitate the communications and the exchange of information between members,
a group decision support system might be a useful tool. Jarke [1986] and DeSanctis
and Gallupe [1985; 1987] discern four kinds of group decision support systems, see
exhibit 4-6.

duration

limited ongoing

close decision local decision
proximity room network.
place

dispersed teleconferencing remote decision
facilities network

Exhibit 4-6. A basic taxonomy of group decision support systems.

These four categories will now be described in more detail:

Decision room
Huber [1984a] describes a decision room as follows:
- a personal screen and input device, for example a keyboard, for each group
member,
- a public display screen large enough to be seen by all participants,
- computing and communication capabilities that: (a) allow each participant to link
his input device to its respective personal screen, to the group leader's personal
screen or to the public screen and (b) allows any participant to retrieve from
predetermined databases,
- software that provides (a) state-of-the-art word-processing capabilities to each
terminal; (b) computing capabilities with a particular focus on drawing
simultaneously on data from several or all terminals or the predetermined data
bases; (c) state-of-the-art graphics capabilities for example capabilities for
constructing and altering work sheets, bar graphs, decision trees and the like and
(d) anonymity of the individual participants when this is desired.

57


We define a computer supported meeting held in a decision room as a group session.
According to DeSanctis and Gallupe [1985], the combination of one input/output
device, one computer processor and one public screen constitutes a less sophisticated
example of a decision room. A computer for each group member is not imperative.

Teleconferencing facilities
For teleconferencing mainly telecommunications are involved. By transmitting images
and voice of the participants of the meeting to the locations involved, an artificial
meeting room is created. This meeting room could be extended with the facilities
described above to become a decision room.

Local decision network '
Office automation offers the possibility to communicate "memos" by means of a local-
area network. In case one of the participants addressed by a memo is not at his desk,
these memos can be stored in order to be retrieved after his return. Reactions of
people, involved in the decision process can be sent to all involved who, at a
convenient time for them, can voice their opinion via the system. Participants can
access common and private databases or decision-support software as required and
view a distant "public screen" on their personal screen as needed. Decision makers are
engaged in their day to day work and set up distant "group meetings" or conferences
as needed. This approach offers greater flexibility than the decision room, but there
is the important disadvantage of removing the face to face communication.

Remote decision network
A remote decision network offers the same facilities as a local decision network, only
extended to "far-away" locations.

The four approaches to group decision support, as presented above, will now be
discussed with the problem of international transfer pricing in mind.

A decision room could be very helpful in providing support as the various experts
and managers involved need to integrate the different aspects of the problem:
commercial, financing and fiscal. These aspects need to be considered from both a
local as well as a central point of view. The decision room will be discussed in more
detail later on.

Local management will only occasionally be able to participate in decision-room
meetings due to geographical distance. This makes another option of the exhibit 4-6
interesting: teleconferencing. Teleconferencing, however, is not something to start with
as the experimental situation is relatively complicated. In comparison to the decision
room, for example, teleconferencing requires more hardware facilities. During a
teleconferencing session, it will be harder to create the informal task-oriented
atmosphere that we are aiming at. Teleconferencing might be an enhancement of the
decision room when the latter is properly developed.

58


A local decision network for transfer pricing can be installed in the head office. As
transfer pricing is explicitly an ongoing process and not a problem that can be solved
during one group session, a local decision network needs our attention. Such a network
has the disadvantage, however, of removing face to face contact and therefore
communication becomes less efficient, according to the research as described by
DeSanctis and Gallupe [1987, p. 598]. As with regard to transfer-pricing decision
making communication is the focal point, we consider the local decision network an
enhancement when the decision room is properly developed.

As in their very essence, transfer prices cross borders, remote decision networks can be
used to involve local management when time differences prohibit direct communication.
We agree with Desanctis and Gallupe [1985], however, that remote decision making is
not a form of decision support to start with as it is very complicated.

In their "foundation for the study of group decision support systems", DeSanctis and
Gallupe [1987, p. 602] propose the concentration on one category only as "the most
appropriate tactic for researchers" as the issue of group decision support is very
complicated. On the basis of the comments made above, it will be clear that we
consider the decision room the most appropriate approach to start with.

As we are designing a system for "real-life" situations and not for a laboratory setting,
we have to accept the constraint that, at the current time, not all decision makers
know how to work with a computer. This is described as "end-user technophobia", see
Suchan et al. [1987, p. 441]. It should be accepted that some decision makers, from
now on defined as "participants", are "computer illiterate" and wish to remain that way.
Therefore not every participant should be expected to provide (some) of the input for
the meeting via a terminal. This implies that the participants need to be supported
during the group session by one or more "analysts" who take care of handling the
computer.

Besides the role of the analyst, a second role is of great importance: the role of
facilitator, see Huber [1984a, p. 198]. As explained in the previous chapter, decision-
process structuring is one of the dominant issues in providing decision support. This
implies that one of the members of the support staff needs to take care of structuring
the meeting. He should help the participants to put their point of view forward when
appropriate and not just as it comes to mind. In his role of facilitator, the support-
staff member who takes this role should not so much concentrate on the outcome of
the meeting, but more on the structure of the decision process itself, see also
Phillips [1985].

In designing group-decision support, the problem of "minimum frequency of use" should
be addressed: each system needs a minimum frequency of use to avoid that it is totally
abandoned, due to lack of experience of its users. This issue is usually approached by
means of widening the scope of the system: the design should be such that group
decision making related to various organizational areas can be supported. Consequently,
the group decision support system should be of a general nature, see Huber [1984a]
and Suchan et al. [1987, p. 443].

59


We think, however, that another approach is also possible, see Sol and Van der
Ven [1988b]. We decide to focus on integration of group decision support in the
organization. This approach can be regarded as one of the main elements of our
design.

By using a facile decision support system generator, see Sprague and Carlson [1982],
frequency of use should be guaranteed by employment of the developed application
domain support by individual decision makers, see Bosman and Sol [1985, p. 90]. So
in our case, for example, computerized support should be available for all those who
are concerned with determining transfer prices, whether working individually or in a
group. It should be easy to adapt the application to future needs as knowledge on how
to operate the generator is already available within the organization. When using this
approach, careful attention should be paid to the choice of the decision support
generator. Frequency of use of this generator should be guaranteed by its use for other
applications as well. In case the application domain is the focus of attention and not
group support itself, we get a group decision support system that differs from the ones
described by Gray [1986] and Huber [1984a].

Huber [1984a] stresses the usefulness of general group decision support especially for
strategic top-level decisions. As explained in chapter 3, transfer pricing is a problem
of translating this top-level policy into financing decisions for the coming one or two
years. It is hence an administrative-based problem. We consider the approach of
integration of group decision support in the organization especially applicable for
problems on this administrative level. The reason is that "hard" company data are more
relevant on this intermediate level than in the case of strategic top-level decisions.

In the case of application-domain support, we would not so much concentrate on
computerized voting tools, agendas, decision trees and alike, see DeSanctis and
Gallupe [1987]. The focus is much more on "what-if" tools on the basis of spreadsheets
and databases. Both the "what-if1 tools and the' databases are specific for the
application domain. Our application-domain support can be considered to be a specific
decision support system existing of databases, various models and a user interface, see
Sprague and Carlson [1982]. Application-domain support facilitates the integration of
group decision support in the organization. Usually, computer-supported group meetings
take place in an isolated setting, see for example Phillips [1986] and Vogel et al.
[1988]. The group meets in a decision room outside their normal working environment.
In these cases, company databases and company specific models are not available
during the group session. In the case of application-domain support, however,
preparation of the group meeting and the meeting itself take place on-site.

In Gray [1986], some examples of existing decision rooms are presented. On the basis
of the design characteristics discussed in this chapter, exhibit 4-7 shows a possible
layout of an ad-hoc "decision room" which focusses on application-domain support and
integration in the organization.

60


T l
l
public screen public screen I
I
personal personal
computer
printer

□ facilitator □
□
table
with
participants

□ □
Exhibit 4-7. Possible set up of an ad-hoc decision room.

The group session should be held in a normal company meeting room and not in a
specially-designed sophisticated decision room. Hardware used can be two ordinary
personal computers. These might be used by two analysts, for example. Such a simple
set-up implies that for a group session, no special hardware is required. The only
hardware specifically needed for this ad-hoc meeting room is two public screens,
attached to the two personal computers. Various possibilities for such a public screen
exist: a video gun, a liquid crystal display or a large monitor could be used. The two
computers should be connected to a low-noise printer (like a laser printer) in order
to provide hard copies when required. Besides the two computers other means like flip
chart or white board can be used for intra-group communication. In comparison with
a sophisticated decision room, this ad-hoc set up is very cheap.

4.5 Four aspects of developing group decision support systems
Developing group decision support systems should be guided by paying attention to
the following four aspects:
- problem specification,
- decision-process structure,
- roles of organizational members,
- computerized support.
The use of these four aspects can be regarded as the basic design guideline for the
prescriptive conceptual model.

PROBLEM SPECIFICATION

Most decision-room sessions described in the literature start with problem specification
during the group session itself, see for example Phillips [1986] and Vogel et al. [1988].
As explained in the previous section, we propose integration of the system in the

61


organization. The aim of our decision support system is to support decision making
with regard to a specific problem domain, transfer pricing, inside as well as outside
the group session. This requires that the system is designed for "a specific category of
ill-structured problems". In order to be able to offer permanent support, a clear
delimitation of the problem is required. Without a problem specification, for example,
it is unknown which databases and "what-if' tools are required. Aspects of the problem
specification that are relevant are:
- time perspective (long or short term),
- level of decision making (strategic level, administrative level),
- kind of decision (operational, relocation or financing),
- categories of possible solutions (like transfer prices or cost transfers)
A clear problem specification requires an important deal of application-domain
knowledge. So a clear problem specification is not only a prerequisite for the system,
it is a part of the support as it forms the basis for a structured approach to decision
making. Or as Larichev [1983] puts it: "It is an art to set a problem correctly 'in a
solvable manner"1. Application-domain knowledge is conveyed to the user by means of
the problem specification. Structuring the decision process is facilitated by offering a
clear problem specification. In this sense, the problem specification and the decision-
process structure are related.

Although a problem specification aids the participants in finding alternative solutions,
the quality and quantity of the solutions that are generated remains dependent upon
the creativity of the participants.

DECISION-PROCESS STRUCTURE

As described in section 4.2, structuring of problem-solving tasks is a vital element in
providing decision support, see also Van Schaik [1987]. The basic structure that we
propose for problem-solving tasks is the following:
1. scan,
2. decision making,
3. authorization,
4. implementation.
For the decision-making task we propose the following task structure:
1. preparatory design,
2. integrative design,
3. choice.

For the description of these tasks, we use the technique as developed by Bots and
Sol [1987]. For the structuring of the choice task, we use a multi-criteria technique.
The execution of the decision-making tasks in this structure take place both inside as
well as outside the group session. As the definition states, the support system focusses
on the support of the decision-making task only, see the previous section.

The tasks presented above constitute a basic approach for structuring group decision
processes at the administrative level. This approach needs to be adapted to each

62

Sec. 4.5 FOUR ASPECTS' OF DEVELOPING GROUP DECISION SUPPORT

specific decision situation. The decision-process structure should facilitate decision
making on both the aggregate and disaggregate level.

ROLES OF ORGANIZATIONAL MEMBERS

In the previous section, we have already discussed the roles of the support staff. We
consider the services of the support staff an integrated part of the prescriptive
conceptual model, see also Phillips [1984]. A description of the role of the system users
is also presented as this clarifies the aim and use of the system.

Participant role
The decision makers involved are defined as "participants". As a group the participants
make the decisions. We presume that all participants work together as a team in the
best interests of the company. Team setting implies that the participants feel free to
submit any item they consider relevant. In other words, in the selection of criteria, no
normative elements are included. We presume that the participants do not pursue
"hidden agendas" and do not play "hyper games". Each participant possesses unique
knowledge which is of importance for the decision. "Expertise sharing" is the most
appropriate description of the interaction amongst participants. Besides bringing in
know-how, all participants should be involved in evaluating alternatives during a group
session. The decision support system is designed for their use. System design should
be such that the system can be used by participants at any time, both inside and
outside the group session.

Authorizer role
The authorizer is one of the participants and performs all participant activities that are
required. Besides this he has three additional tasks. Firstly he is official chairman of
the meeting. This implies that he can aid and correct the facilitator when necessary.
The facilitator is mainly concerned with the decision process. In those cases in which
no common consent by the group is reached, it is the authorizer who cuts the Gordian
knot. The second role of the authorizer is that he authorizes the final decisions made
by the group, if he is allowed to do so. In cases in which higher organizational levels
need to be involved, the authorizer submits the decision for authorization. The third
task of the authorizer is to start the decision process. This is shown by the first
coordinating decision in exhibit 4-1.

Facilitator role
The facilitator's main task is to structure the decision process, both inside as well as
outside the group session. The facilitator is explicitly a member of the support staff.
He should be independent from the group. He is not and cannot be a participant as
he might otherwise mix up the two responsibihties: that of facilitator and that of
participant. Another facilitator task is to be an intermediary between the participants
and the analysts, when necessary. He should know when the various software modules
can be used. A third task for the facilitator is to stimulate an informal, problem-
focussed atmosphere within the group, both inside and outside the group session.

63


Analyst role
The analyst operates the computer both inside and outside the group session. The
analyst should also be able to adapt the software when requested by the participants.
During the preparatory-design task, the analyst is actively involved in assisting the
participants.

COMPUTERIZED SUPPORT

As Van Schaik [1988] demonstrated, computer support alone does not guarantee better
decisions. In combination with the three aspects defined earlier, however, we consider
fast "what-if' tools, databases and a computerized multi-criteria technique extremely
useful. As decision support is only used when needed, we should guarantee a minimal
frequency of use in order to avoid disuse of the system. Therefore, an easy to use
software tool should be used as a "decision support system generator", see Sprague and
Carlson [1982]. This generator can be used for other applications as well. No fixed
links between the various modules exists. The decision-process structure provides
guidelines for software use but ultimately the user decides upon the way the various
modules interact.

With regard to the development of group decision support systems, Von
Winterfeldt [1983] comments that participants, not being experts in decision analysis,
do not want "overly complex modeling [or] too extensive use of incomprehensible or
inaccessible machinery". Phillips [1985] described the necessary characteristic of the
software as "simple but not simplistic". Little [1971, p. 483] puts it as follows: "...The
biggest bottleneck in the managerial use of models is not their development but getting
them used. I claim that the model builder should try to design his models to be given
away."

4.6 Conclusion
In this chapter, we developed possible design characteristics of decision support systems
for small face-to-face groups. In summary, the main design characteristics are:

- Our group decision support system can be described by means of the following
four aspects: problem specification, decision-process structuring, roles of
organizational members and computerized support. Decision support is considered
to be much more than computerized support alone.
- Our system is designed to aid participants cooperating in small expertise-sharing
groups. A specific "application domain" is the focus of attention and the support can
be used both inside as well as outside group sessions. In other words, individual use
is supported.
- The decision-making task and not the entire problem-solving task is supported.
During the choice task, a multi-criteria technique is used.
- For group sessions, an ad-hoc decision room is set up. The "simple and sufficient"
approach which is followed in the selection of the hardware is followed in the other

64

Sec. 4.6 CONCLUSION

areas of the design as well. Participants are not forced to operate the computer
themselves, neither inside nor outside the group sessions.
- Frequency of use of the computerized support should be guaranteed by means of
the use of a facile decision support system generator.

Aiming at improving decision making, we would like to describe our approach to
group decision support as simple and sufficient. In this phrase, "simple" is the opposite
of "complex". "Sufficient" is the opposite of "perfect" in the sense that a better decision
process, and not perfection is the aim of the prescriptive conceptual model, see also
Keen and Scott Morton [1978, p. 74].

65


66

5

TRANSFER PRICING AT EUROPHAR: THE PROBLEM
5.1 Introduction
In the previous four chapters, we provided a theoretical background for our research.
In the remaining part of the book implementation and experimental use of a group
decision support system are the focus of attention.

As explained in chapter 1, we decided to cooperate with one of the large, innovative,
European-based, pharmaceutical companies for the period of five years. Analysis and
experimental use of the system took place at this company. At the request of the
company, figures, names, and some minor details have been changed. In this chapter
we concentrate on the description of an existing situation within this particular
pharmaceutical company. The name of the company we cooperated with is "Europhar
International BV". Below the decision process at Europhar International BV
(abbreviated as "Europhar Int BV" ór "Europhar Int") is described.

The five years of longitudinal research which took place at Europhar Int are indicated
as "Year 1" to "Year 5". In the period Year 1-Year 2 the author worked as a financial
analyst in the fiscal & financial department of Europhar Int BV. One of his
responsibilities was to analyze the financing effects of transfer-price decisions. It will
be clear that this part of the author's job contributed to the insight in the decision
process before the introduction of the system that we designed. It also contributed to
the choice of Europhar Int BV as the experimental company.

The sheer size and complexity of the transfer-pricing problem requires that a detailed
description of the existing situation and the implementation of the experimental system
are to be limited to transfers between the parent company and one subsidiary company
only. The subsidiary that is chosen is Europhar UK Ltd which is situated in London.

One of the reasons for this choice is that, with regard to the financing situation of this
subsidiary, the same people are involved in decision making since Year 1. The eight
people involved are the two top managers who are ultimately responsible, and the six
participants. It is our impression that it is rather unique to have a group of eight
organizational members in a dynamic organization like a multinational company who
have worked in the same positions on the same problems for five years. This offers
the possibility for a better comparison of the situation "before" and "after" introduction.
Another reason for the choice of the British subsidiary of Europhar is the nonexistence
of language, time and distance problems. The main reason however, is the following:

Within Europhar, problems exist with regard to the British situation for a number of
years. As explained in chapter 2, the situation with regard to the United Kingdom
(UK) is difficult for many companies, not only for Europhar. But for Europhar,
however, the situation is even more complicated because of the rather intricate

67

TRANSFER PRICING AT EUROPHAR: THE PROBLEM Ch. 5

relationship with the various source and destination companies. It is the opinion of top
management that if the experimental system can aid decision makers in the UK case,
then, as far as Europhar is concerned, it has proved its feasibility. For a good insight
in the research project, we now describe the organizational set up.

In the beginning of Year 3, the start of the project is authorized by means of an
agreement between Europhar and the research institutions involved. A joint steering
committee is formed. In this joint steering committee, the following people participate:
- Prof. Dr. H.G. Sol, Delft University of Technology, chairman,
- Mr. R. Marcus, financial' director, Europhar Int BV,
- Mr. D. Pieterse, head information management department, Europhar Int BV,
- Mr. F. Terwolde, head price coordination department, Europhar Int BV,
- Dr. H.C. Verlage, transfer-pricing expert.
For an organizational chart of Europhar and its holding company Eurochem, see
exhibit 5-1. The six progress meetings of this committee take place in June Year 3,
November Year 3, July Year 4, December Year 4, March Year 5 and June Year 5.

For system design, a working group is formed. In this working group, the following
people participate:
- Mr. H. Berkers, head fiscal & financial department, Europhar Int BV,
- Mr. N. Katwijk, manager European region, Europhar Int BV,
- Mr. K. Nijboer, staff member, fiscal & financial department, Europhar Int BV,
- Mr. F. Terwolde, see above,
- the author.
On ad-hoc basis, T. Derecks, financial controller of Europhar UK Ltd and J.
Oostbeek, staff member of the fiscal & financial department of Eurochem Holding NV
participate in working-group meetings. Working group meetings take place about once
every fortnight during the whole of Year 4. Regularly, discussion drafts of the software,
as developed by the author, are presented at these meetings.

After the design, development of the experimental system is done by Derecks, Nijboer
and the author. This takes place in the period January-March Year 5. Various students
from Delft University of Technology, two programmers from a software house and two
software experts from Europhar assist in the development activities.

The time table for the research project appears as follows:
- Year 1-Year 2 (two years): same participants and top management, no transfer-
pricing support system. Until the end of Year 2, the author is Europhar employee.
- Year 3-Year 4 (two years): same participants and top management, analysis of the
existing situation, design of the system.
- January-March Year 5 (3 months): same participants and top management,
development of the system.
- April-September Year 5 (6 months): same participants and top management,
experimental system use.
- October Year 5 (1 month): same participants and top management, evaluation of
system use.

68

Sec. 5.2 EUROPHAR: THE COMPANY

5.2 Europhar: the company
We start with the description of the Eurochem group. Then the two divisions which
are relevant for our research (Europhar and Eurobitec) are described. We also
describe Europhar's financial management information system and its information-system
policy.

THE EUROCHEM GROUP OF COMPANIES

Eurochem is the overall holding company of all group companies and divisions. The
Eurochem (EUROpean CHEMical) group is a multinational chemical company of Dutch
origin. Its turnover in Year 5 is about NLG 25 billion. Eurochem Holding shares are
traded on the Amsterdam stock exchange and other main financial centers all over
the world. The Eurochem head office is situated in Rotterdam, the Netherlands.

Eurochem produces various kinds of chemical products such as plastics, fibers and
human-ethical drugs. Special divisions exist for these mam classes of products. Each
division has its own profit responsibility. All Eurochem activities are as much
decentralized as possible. This is necessary, because of the variety of markets serviced
and the great differences in products sold. So each division is responsible for and
carries out its own research, production and marketing activities.

Centralized activities relate to strategic policy, personnel management, management
development and finance and administration. Finance and administration, among others,
sets detailed guidelines for the financial MIS (management information system) and
consolidation of figures for external reporting purposes. Monthly reporting of the
complete profit and loss account, balance sheet and source and application of funds
per division is required at the Eurochem head office within twelve working days after
the end of the month. To make this possible, figures from the local subsidiary
companies have to arrive at the divisional head office within ten working days after the
end of the month. Two working days are used for divisional consolidation of these
figures before they are sent to Eurochem Holding.

There is a separate fiscal unit in most countries in which Eurochem daughters are
situated. All Eurochem daughters from all divisions in, for example the UK, form one
fiscal unit. A similar fiscal unit exists in the Netherlands.

The information-system policy for the finance and administration area is a completely
centralized activity. Within all divisions Comshare's System W modeling and database
software is used for MIS and external reporting activities. For quick ad-hoc financial
modeling, Lotus 1-2-3 spreadsheets are used throughout the company. For word
processing, however, no standards exist within Eurochem. The hardware standard for
personal computers is IBM. Standard operating system for these computers is MS-
DOS.

69


Decentralized fiscal and financial activities are coordinated by the central fiscal &
financial department of the group. This Eurochem headquarter department cooperates
closely with the fiscal & financial departments within each division.

THE EUROBITEC DIVISION

Eurobitec (EUROptan BIoTEChnology) is a newly established division of Eurochem.
It's operations started in Year 1. All biotechnological products are produced at
Eurobitec. Although Eurobitec is still small, it is growing fast. Turnover in Year 5 is
about 500 million Dutch guilders (NLG 500 mln). One subsidiary company of
Eurobitec International BV is located in Oxford (Eurobitec UK Ltd). This company
has its own production facilities. Biosulfine is one the main products that is
manufactured at the Oxford production plant. The marketing of Biosulfine under the
name of Sofanaat is done by Europhar, the pharmaceutical division of Eurochem. The
price of Biosulfine (or Sofanaat) influences the profit for both Europhar and Eurobitec.
As both divisions have their own profit responsibility, see chapter 3, inter-divisional
transfer prices may be subject of dispute between the two divisions. Transfer pricing
between Europhar and Eurobitec should be categorized as "management-control
transfer pricing". Because of this some tension exists between Eurobitec and Europhar,
see chapter 3.

THE EUROPHAR DIVISION

The pharmaceutical division is called Europhar (EUROpean P/£4i?maceutical).
Europhar has over fifty daughter companies in countries all over the world, see the
description of a typical pharmaceutical company in chapter 2. Europhar's marketing
activities are divided into five regions. Only two of them are important in view of our
research project: Europe and Far East. Europhar's turnover in Year 5 is about NLG 3
billion. The head office of Europhar is situated in Amsterdam, Netherlands where the
main production facility and research center are located. Some new laboratories have
been built to comply with "good laboratory practice" regulations. Of course, these
laboratories cause an extra need for liquidity.

As explained in chapter 3, budgets and consoUdated profit measurement are used for
planning and control purposes. This requires extensive budgeting procedures. The
budget comprises sales in units per destination per article, expected prices, detailed
cost breakdown, detailed opening balance sheet, detailed closing balance sheet and a
source and application of funds. Each autumn, regional management and the general
manager of the local company make a budget proposal for the coming calendar year.
This usually takes a number of meetings. Discussions during these meetings may be
tough. Once the budget figures are determined, however, they are accepted within the
complete Europhar organization. This holds also for exchange rates and other external
financial parameters that are included. Budgets are only made for one year. In the
final determination of budget figures, the financial director and the marketing director
are involved. The expected external financial parameters, like exchange and interest
rates, are determined by the Eurochem fiscal & financial department.

70


Europhar UK Ltd in London is one of Europhar's subsidiary companies. Europhar UK
Ltd buys from both Europhar Int BV (the Dutch parent company) and Eurobitec UK
Ltd (mainly Biosulfine) and it sells mainly to British National Health Service (NHS),
but to Europhar Australia and Taiwan as well. Europhar UK exports to non-group
customers and, in total, sales are approximately 50 million British pounds sterling
(GBP 50 mln) in Year 5.

With regard to the UK, all other European countries are treated as reference
countries. One of the main products of Europhar is Biosulfme/Sofanaat (produced by
Eurobitec). Because Biosulfine is a new product, extra promotional costs to market this
drug are allowed by the NHS authorities.

EUR0CHEM HOLDING NV (Rotterdam)

executive board

fiscal & financial
department ■— other
[Oostbeek]

other Europhar Int BV Eurobitec Int BV
(Amsterdam) (Rotterdam)
I
board of directors board of directors

-other
| Eurobitec UK Ltd |
financial director marketing director (Oxford) other
[Marcus] [Gisbers]

price fiscal & other European region Far East region other
coordination financial
department department
[Terwolde] [Berkers] regional manager regional manager
[Nijboer] [Katwijk]
information
management Europhar
department Australia Pty Ltd other
[Pieterse]
[France»]
Europhar Taiwan Ltd
Europhar UK Ltd
(London)
I
general manager
[0'Brian] Europhar Belgium NV
Europhar France SARL
Europhar Germany GmbH
Europhar Italy SRL
financial marketing other Europhar Netherlands BV
controller manager Europhar Sweden AB
[Derecks] Europhar Switzerland GmbH

Exhibit 5-1. Organizational scheme of the Eurochem group.

71


The organizational structure of Eurochem, as far as relevant for our case, is displayed
in exhibit 5-1. The links between the various companies express 100% ownership. The
dotted line between the two fiscal & financial departments indicates a functional
relationship. Names of the people involved are indicated by square brackets.

The group of decision makers with regard to financing of Europhar UK Ltd consists
of the following organizational members:
- Berkers, head fiscal & financial department Europhar Int BV, Amsterdam,
- Derecks, financial controller Europhar UK Ltd, London,
- Katwijk, regional manager Europe (authorizer of the decision process),
- O'Brian, general manager Europhar UK Ltd, London,
- Oostbeek, fiscal & financial department, Eurochem Holding NV, Rotterdam,
- Terwolde, head price coordination department Europhar Int BV, Amsterdam.
This group is supported by Nijboer who is staff member of the fiscal & financial
department of Europhar Int BV, Amsterdam. Nijboer takes the role of analyst. When
Katwijk is not allowed to authorize certain decisions, normally the involvement of the
Europhar directors Gisbers and Marcus is sufficient. In exceptional cases (for example
in case of a capital increase, see chapter 3), the involvement of the Eurochem
executive board is required. The six participants and two top managers mentioned
above are working in their current position since Year 1 or earlier.

We are now able to discuss the companies and their relations within the Europhar
case. In chapter 3, we displayed a conceptual framework to categorize different kinds
of companies that exist. Now, we fill in the Europhar companies for those definitions.
We distinguish the following classes of companies:
- object company:
- Europhar UK Ltd, London.
- source companies:
- Europhar Int BV, Amsterdam,
- Eurobitec UK Ltd, Oxford.
- destination companies:
- Europhar Australia Pty Ltd, Sidney,
- Europhar Taiwan Ltd, Taipei.
- destinations third:
- Generic sales (UK),
- NHS human-ethical sales (UK),
- Export third parties China,
- Export third parties Nigeria.
- reference companies:
- Europhar Belgium NV, Brussels,
- Europhar France SARL, Paris,
- Europhar Germany GmbH, Munich,
- Europhar Italy SRL, Milan,
- Europhar Netherlands BV, Utrecht,
- Europhar Sweden AB, Stockholm,
- Europhar Switzerland GmbH, Basel.

72


In chapter 3, we explained the relationship between these different kinds of companies.
Filling in the names of the Eurochem companies gives a scheme like exhibit 5-2.

export third
parties China
external sale of export third
destination article parties Nigeria
(external selling price) generic sales (UK)
NHS human-ethical
sales (UK)

destination
Eurobitec
UK Ltd

transfer of internal sale of Europhar
source article Europhar destination article Australia Pty Ltd
UK Ltd
(transfer price) (internal selling price) Europhar
Taiwan Ltd
Europhar
Int BV

Europhar Belgium NV
Europhar France SARL
reference article Europhar Germany GmbH
Europhar Italy SRL
(reference price) Europhar Netherlands BV
Europhar Sweden AB
Europhar Switzerland GmbH

Exhibit 5-2. Relationships of object company Europhar UK Ltd.

5.3 Decision process before system introduction
In Year 1 and Year 2, before the start of our research project, the first attempts are
made to come to a systematic approach with regard to financing and transfer pricing.
It is felt that the transfer-price policy is too much based on ad-hoc decisions and that
a consistent policy for the short term (the budget year) is lacking. Long-term
commercial, fiscal and financing policy is relatively well worked out, but the translation
of the long-term policy into a short-term financing and transfer-price policy is lacking.
In view of external pressure from mainly fiscal and health-care authorities such a policy
is required. The initiative to improve the situation is taken by marketing director
Gisbers. A small working group is formed. The following people participate in this
working group:
- Berkers, head fiscal & financial department, Europhar Int BV,
- Oostbeek, fiscal & financial department, Eurochem holding NV,
- Terwolde, head price coordination, Europhar Int BV,

73


- the author, at that time working at the fiscal & financial department of Europhar
Int BV.
The set up of this working group is rather informal. The main aim is twofold: firstly
to interpret the long-term fiscal and commercial policy and translate the Europhar
policy into practical guidelines to determine transfer prices and secondly to set up
computerized support to aid decision makers.

With regard to the first aim, the long-term transfer-pricing policy is translated in a
gross margin of the local company on a sales-minus basis. Let us say that an overall
local gross margin of 45% is considered to be the long-term aim. In case of changes
which disturb this gross margin, transfer prices might be reviewed to bring the gross
margin back on target. Signals of a changing gross margin are picked up by the
Europhar fiscal & financial department . from the monthly MIS reports. The
computerized support that is set up by the working group provides an overview of the
current situation on the basis of MIS profit and loss figures over the past three years.
The package used is Lotus 1-2-3. Specific details concerning the financing situation and
regulations are entered in text mode, in the same Lotus file.

In January Year 3, the research project starts with an analysis of the existing situation
in order to construct a descriptive empirical model. As the interviews with the
participants do not reveal much more than that "the situation is very complex" and that
"everything relates to everything", it is decided that the analysis of the existing situation
will concentrate on the situation with regard to one local company: Europhar UK Ltd
(Europhar UK Ltd is sometimes referred to as "London". Europhar Int BV is
sometimes referred to as "Amsterdam").

All documents with regard to London from the beginning of Year 2 onwards are
closely examined. These documents include telex messages between Amsterdam and
London, internal memos (memo is an abbreviation for "memorandum"), internal reports,
minutes of meetings in Amsterdam as well as in London, correspondence with the
NHS, fiscal authorities and external consultants.

The analysis of the documents reveals the decisions that are taken, but not the
decision process itself. It again makes clear that "everything relates to everything". As
this approach does not reveal much of the way decisions are taken, the meetings
during which transfer-price and financing decisions concerning the UK and other
companies are made, are analyzed. As explained in the previous section, we now
concentrate on the UK situation. One of the meetings concerning the UK is video
taped. In this meeting which takes place on July 14th Year 4, Katwijk, Terwolde and
Derecks participate. The meeting takes three hours.

The problem at hand is a liquidity surplus that comes up due to the unexpected
success of Biosulfine. The budget for Year 4 is made in the autumn of Year 3.
Biosulfine is launched in January Year 4. Biosulfine sales are above budget which
results in a an expected extra gross margin of about GBP 1.4 mln for Year 4. The
expected liquidity surplus is also GBP 1.4 mln. In Year 4, new research laboratories

74

Sec. 5.3 DECISION PROCESS BEFORE SYSTEM INTRODUCTION

are planned to be built in Amsterdam. This is one of the reasons why a transfer of
funds to Amsterdam needs to be considered.

The MIS figures for May Year 4 indicate clearly the liquidity surplus. Problems are
for the first time discussed in June Year 4. As regional manager Katwijk is heavily
involved in solving problems concerning the marketing approach for Briconcyl, one of
the Europhar products, he decides that decision making should be postponed to July.
Organizational members involved are Berkers, Terwolde, Derecks, Katwijk, Oostbeek
and O'Brian. Katwijk asks financial controller Derecks to prepare figures for the
meeting that will take place in July. Analyst Nijboer works out a consolidated profit
statement per article.

Katwijk, Derecks, price coordinator Terwolde and analyst Nijboer participate in the
meeting. O'Brian and Berkers are on holiday and Oostbeek is tied up with fiscal
problems of another division. The agenda for the meeting is as follows:
1. discussion of input papers,
2. conclusion with regard to commercial and fiscal aspects,
3. action steps.

In chapter 4 we stressed the importance of an informal, task-oriented attitude within
the group. The atmosphere during this meeting is also task oriented, informal and
open. Relief of tension takes place by means of a relaxed and humorous yet
businesslike approach.

The meeting is chaired by regional manager Katwijk. As input the following papers
are served:
- Revised profit and loss budget figures, as estimated by financial controller Derecks,
for Year 4. Per month detailed figures are provided and not only a statement of
annual figures. Comments on the estimated results, as made by Derecks, are
included.
- Estimated NHS-profit calculation for Year 4, provided by Derecks.
- Summary of the expected source and application of funds, expected balance sheet
and expected profit and loss account for the years Year 5-Year 9. Exchange rates,
sales developments, etc. are based upon Derecks' personal expectations.
- Correspondence with the various authorities, internal memos, etc.

Total number of input pages is 42. These 42 pages are provided as loose sheets of
paper in different sizes without clear headings. Most reports are computer printouts
of Lotus 1-2-3. Each participant of the meeting has a copy of these 42 pages. Input
papers are sent to the participants a few days in advance of the meeting.

Besides the items mentioned above, analyst Nijboer has brought with him detailed
information on sales in units, selling prices, transfer prices, local added value and
consolidated financial contribution per article. This information consists also of
computer printouts produced by means of Lotus 1-2-3. The number of pages is about
30. Analyst Nijboer is in possession of the only copy available during the meeting.

75


Now the first part of this meeting is discussed: Katwijk opens the meeting and sets
the agenda. Discussion of the input papers is the first item. Derecks proposes not to
transfer all the surplus to Amsterdam, but to use GBP 240,000 in order to set up a
marketing conference for the general practitioners in Britain. This conference was
already planned a long time ago. The plans have been discussed with O'Brian before
Derecks left London for the meeting.

Katwijk proposes to increase royalties. For eight years Europhar UK Ltd has been
paying a lump-sum royalty payment to Amsterdam. Sales have increased significantly
in that period. Also inflation has occurred. Therefore it would be reasonable to ask
a larger contribution from London. This could also be considered to be at arm's length
(see chapter 3). Current royalties are NLG 11 mln annually. Katwijk proposes to
change the annual royalty payment to GBP 4 mln. Using an exchange rate of 3.30
NLG/GBP, the original amount of NLG 11 mln is GBP 3,330,000 so a royalty payment
of GBP 4 mln would be an increase of GBP 670,000. According to Katwijk, the
change in currency is an advantage for London. The consequences of exchange rate
fluctuations do no longer have an impact on the Europhar UK figures, as Amsterdam
now takes the complete exchange risk, explains Katwijk.

Terwolde puts forward, that as royalty is a lump-sum payment, it needs to be allocated
for NHS purposes. Allocation is currently done on the basis of sales. According to
Terwolde, it would also be reasonable to allocate on the basis of cost of goods (cost
of goods equals the transfer price plus the local added value). By means of some
calculations, using a pocket calculator and a piece of paper, Derecks determines that
an allocation on the basis of cost of goods would not make too much difference. The
exact calculations are not shared with the group, but the conclusion is accepted.

A second option, proposed by Terwolde, is a cost transfer of transportation cost from
Amsterdam to London. As part of a Europhar Int contract with "Channel Transport
Company", a part of the Europhar UK transportation costs are borne by Amsterdam.
The distribution for the south part of the UK is cared for by "Channel Transport".
This is a part of an experimental contract for the entire Eurochem group. It would
be reasonable to transfer these cost, which are now paid by Amsterdam, to London.
Given budget sales for Year 4, these costs would be GBP 200,000, according to
Terwolde. Katwijk, however, does not agree. Biosulfine, the product that causes the
liquidity surplus, is not the only product that deviates from budget. Katwijk expects
also higher Colcipar sales as he recently has spoken to one of the marketing people
of London. As sales of Colcipar are up, extra transportation cost should be expected.
This is because of the requirement that Colcipar ampules should be stored at low
temperatures during transportation. On a piece of paper the estimated impact of extra
Colcipar sales on transportation costs is calculated. These are GBP 50,000.

This meeting continues in a similar manner. For a detailed account, see Van der
Ven [1989b]. At the end of the meeting it is decided that the total amount of the
liquidity surplus will be spent as presented in exhibit 5-3.

76

Sec. 5.3 DECISION PROCESS BEFORE SYSTEM INTRODUCTION

royalty increase: GBP 670,000
marketing conference.' GBP 240,000
transportation cost: GBP 200,000
transportation Colcipar: GBP 50,000
Retravir transfer price increase: GBP 240,000
total: GBP 1.4 mln

Exhibit 5-3. Summary of financing alternatives decided upon at the end of the meeting
in July Year 4.

Katwijk thanks all participants and closes the meeting.

After the meeting Derecks flies back to London and discusses the options with general
manager O'Brian. O'Brian agrees on all points. When working out the transportation
cost option, however, Derecks notes that a calculation mistake has been made during
the meeting in July. Colcipar transportation cost are GBP 140,000 instead of the
calculated GBP 50,000. This is communicated to Katwijk and Terwolde by means of
telex. They agree with a new total transfer for transportation cost of GBP 340,000
(GBP 200,000 + GBP 140,000) instead of the expected GBP 250,000 (GBP 200,000
+ GBP 50,000). Katwijk authorizes the transfer of transportation cost and the transfer-
price increase for Retravir. This is done by means of an internal memorandum.

In August, marketing director Gisbers is informed by Katwijk about the plans to hold
the marketing conference and to transfer the remaining amount to Amsterdam. As
Gisbers does not agree with the marketing conference, this option is disregarded. He
proposes that the entire amount will be transferred to Amsterdam. So additional means
to transfer funds need to be found. Gisbers authorizes the creation of a new royalty
contract by means of an internal memorandum.

In the meantime, the actual MIS figures of July show that the expected liquidity
surplus by the end of the year will be GBP 1.5 mln instead of GBP 1.4 mln.

In August, Katwijk and Terwolde decide to increase the transfer price of Ogmentin
from NLG 112,-- per unit to NLG 148,- per unit. This will result, according to budget
Year 4, in an extra transfer of GBP 250,000. This is the amount that is needed. For
the new picture, see exhibit 5-4.

77


royalty increase: GBP 670,000
transportation cost: GBP 200,000
transportation Colcipar: GBP 140,000
Retravir transfer-price increase: GBP 240,000
Ogmentin transfer-price increase: GBP 250,000

total: GBP 1.5 mln

Exhibit 5-4. Summary of financing alternatives decided upon in August Year 4.

Terwolde informs Oostbeek by telephone about the decisions made. Oostbeek does not
object against any of the measures that are proposed. This especially as the entire
liquidity surplus is transferred to the Netherlands. Oostbeek considers this a good
thing in view of the demands made by the Dutch fiscal authorities. Berkers is informed
by Terwolde and does not object. So Katwijk authorizes the transfer-price increase for
Ogmentin in his role of regional manager. The end of August, financial director
Marcus authorizes the transportation-cost transfer by means of an internal
memorandum.

The decision process of which some parts are described above will be used for
construction of the descriptive empirical model.

5.4 Descriptive empirical model
To describe the decision situation we use the four aspects for describing decision
situations as presented in chapter 4: problem specification, decision-process structure,
roles of organizational members and computerized support. The descriptive empirical
model as presented in this section is based upon a systematic analysis of the entire
meeting which took place on July 14th Year 4, see Van der Ven [1989b]. In the
previous section, a part of this meeting has been described.


The problem at hand is a liquidity surplus. Not only financing alternatives are
considered during the meeting. Besides the four categories of financing alternatives
(dividend payment, cost transfer, royalty increase and transfer-price increase),
operational and relocation alternatives are also discussed. The operational alternatives
discussed are: the marketing congress for general practitioners and the withdrawal of
Prefanid and Apticram from the market. The relocation alternative discussed is: the
transfer of the packing activities for Cotergenine to Amsterdam.

Two different time perspectives are considered simultaneously. The Year 4 situation
is discussed on the basis of budget figures. When this is considered relevant, figures
are adapted during the meeting. Long-term plans are merely based upon the
expectations of the individual that makes up the figures, in our case Derecks. During

78

Sec. 5.4 DESCRIPTIVE EMPIRICAL MODEL

the meeting, two time perspectives (budget year and long term) are considered
simultaneously. No distinction is made between the relatively "hard" budget figures and
the much "softer" long-term expectations.

Not all eight categories of financing alternatives, as discussed in chapter 3, are
considered. The four options that are not discussed during this round of decision
making are: local lending, internal selling-price decrease, current account change,
intracompany loan. It does not become clear why these alternatives are not considered.

The meeting does show that transfer-price problems are closely linked with other
financing alternatives. It also becomes clear that fiscal, commercial and financing
problems are considered simultaneously.


Not all decision makers are directly involved in the decision process. Instead, decisions
are made in subgroups and by means of the use of telecommunications (telex,
telephone and telefax). The meeting that we partially described above is a subgroup
meeting. Some decision makers are put into a position similar to the position of the
higher management levels that finally authorize the decision. Their only contribution
is their "yes" or "no" with regard to the solution proposed. Their expertise is not used
in the decision process itself. This is illustrated in the analysis of the dividend
alternative, see Van der Ven [1989b]. No single argument pro or contra is provided.
Merely the opinion of Oostbeek is put forward. Just before the final decisions, not
concerning dividend, are authorized, the opinion of Oostbeek is asked.

No clear distinction is made between the aggregate and the disaggregate-level aspects
of the problem. This is illustrated as only a transfer-price increase of Retravir is
considered. On an ad-hoc basis, a lot of information at disaggregate level is provided
by analyst Nijboer, but this information is not accessible for the participants, and
therefore it is not used. In total this happens four times. After the meeting, when
marketing director Gisbers objects to holding the marketing conference, Katwijk and
Terwolde decide to increase the transfer price for Ogmentin. This decision is taken
without any systematic search at the disaggregate level.

When the decision process is analyzed, various actions can be discerned, both during
the main meeting in July Year 4 and during the entire period of decision making
(June-August Year 4). The frequency of relevant actions taking place during the
meeting is presented in exhibit 5-5.

During the entire period (June-August Year 4), four times partial decision making
(in subgroups) takes place.

The meeting is not very well-structured: nine times a change of subject takes place
without a proper conclusion. This is realized by the participants. During three
occasions, problems arise as the budget is not properly actualized. The use of
computerized support is not very well organized. Two times, it takes the participants

79


a long time to find the relevant sheet amongst the 42 sheets of computer output. As
the input papers are related to action steps and the approach vis-a-vis the authorities,
the agenda for the meeting cannot be pursued. No structure is detected in the
meeting, except that all items discussed relate in one way or another to financing
problems. Excluding the break, the meeting takes three hours. No systematic analyses
of all alternatives (on aggregate or disaggregate level) takes place. Systematic evaluation
of possible options is lacking completely. When the participants get a transcript of the
meeting, they feel uncomfortable about the lack of structure. One of the participants
remarks that he "did not realize that it was that bad".

relevant actions with frequency
information available but not shared with other participants
considerable loss of time due to search for information
number of times that participants are missing
need for properly actualized budget
information on alternatives lacking
number of operational or relocation alternatives evaluated
number of alternatives evaluated (at aggregate level)
general complaints about the complexity of the issue
disaggregation without systematic selection of articles
quantitative insight in the consequences of actions impossible when needed
sensitivity analyses with regard to external parameters impossible when needed
no proper introduction or résumé when relevant
change of subject without proper conclusion
manual calculation error
choice task mixed with the integrative-design task

Exhibit 5-5. Frequency distribution of actions taking place during the meeting of July
Year 4.

A last point that we make is that the decision of August Year 4 to increase the
Ogmentin price is a good example of a "partial" decision: only one alternative is
considered and decided upon by a small subgroup (Terwolde and Katwijk).


Derecks does not defend a specific local interest in view of the transfers of funds
under consideration. All participants who are involved, share their expertise in order
to find the best solution available.

Katwijk takes the authorizer role and chairs the meeting. He initiates the decision
process and authorizes the outcome of the meeting. In the case described above, no
lasting disagreement between group members occurs. When this might happen, Katwijk
takes the final decision. From analyses of other meetings and the interviews, we would
like to put forward, however, that these cases are extremely rare.

Nijboer takes the role of analyst in the sense that he prepares the reports for the
meeting.

80

Sec. 5.4 DESCRIPTIVE EMPIRICAL MODEL


Computer output is provided by Derecks and analyst Nijboer. Nijboer provides
information on the financial contribution per article (disaggregate level, see chapter 2).
Derecks provides information on the long-term expectancies for the Europhar UK
situation.

Information is provided in the form of reports. Neither the layout, nor the contents
of the reports are discussed in advance with the participants. The relevance of
Nijboer's input is not properly understood by the participants and the discussion of
Derecks long-term planning gets stuck as the participants do not agree on the
premises. As no on-line computer faculties are available, it is impossible to perform
"what-if" calculations during the meeting. Calculations are made by hand, which in the
Colcipar case causes a substantial calculation error.

Layout of the reports is not explained to the participants before the meeting. This
causes irritation and loss of time during the meeting itself.

5.5 Descriptive conceptual model
In this section, a descriptive conceptual model is presented. The aim of this model
is to describe the basic pattern of this, and other decision situations concerning
transfer pricing. The model is not only based upon the empirical model as presented
in the previous section but it is also based upon other decision situations that we
encountered at Europhar in the period Year 1-Year 3. These decision situations
included both European and third-world countries. For the description of the
conceptual model, we again use the four aspects as presented in chapter 4: problem
specification, decision-process structuring, roles of organizational members and
computerized support.


The problem specification is vague. The only thing that is clear is that a financing
problem exists: This problem is defined as a liquidity surplus or a liquidity shortage.
Management-control issues are not considered to be a part of the problem.

The myriad of problems addressed in the decision process includes the following
aspects:
- short and long-term consequences,
- strategic and administrative-level decisions,
- operational, relocation and financing alternatives,
- various categories of financing alternatives.
Financing, commercial and fiscal consequences are considered simultaneously.

81


"

i 1 scan

no

analyzing?

yes w

decision
making

alternatives for
which authorization at
higher level is required

submission formal
for agreement
authorization by
at higher authorizer
level

Ó determine authorization
level required
-o- alternatives for which
alternatives yes
authorized? authorizer involvement
is sufficient

implementation


Exhibit 5-6. Structure of the problem-solving task in the descriptive conceptual model.

In chapter 3, we discerned eight categories of financing alternatives. All these eight
categories are known and used in various decision situations. Not all eight categories
of financing alternatives, however, are explicitly considered as possible solutions for one

82

Sec. 5.5 DESCRIPTIVE CONCEPTUAL MODEL

specific liquidity problem which requests a solution. Only the alternatives that come to
mind are discussed.


Using the process structure for problem solving as described in chapter 4, we get the
picture as presented in exhibit 5-6. The authorize task as presented in this exhibit
differs from the authorize task as presented in the previous chapter. The authorize task
is split into two different tasks. Within the decision-making group, one participant takes
the role of authorizer. His involvement is shown by the task formal agreement by
authorizer. It is possible that the authorizer's formal agreement is sufficient for
implementation of the group decision. This depends upon the authorization regulations
that exist within the company. In case a higher level within the organization needs to
be involved, the authorizer asks permission for implementation at the higher level. This
is shown by the coordinating decision determine authorization level required and the task
submission for authorization at higher level. Note that in general, relocation and
operational alternatives are also considered in the decision process and therefore also
need to be authorized.

The environment is constantly scanned by all participants for signals which might
require the start of the decision process. Whether the decision process will really be
started depends on the importance of the other problems at hand and the time
available. It might be that more urgent matters require the attention of the participants.
In exhibit 5-7, the scan task is shown in more detail.

<'
global
analysis
1 of the
environment

no
}J changet
policy required?
yes ,r

priority
evaluation

path: problem solving scan

Exhibit 5-7. Structure of the scan task.

83


The global analysis of the environment is performed by all participants. On the basis
of MIS-reporting, "hearsay" and any other possible information source, participants
decide whether further analyses of the financing situation is desirable.

The decision makers determine whether changes in financing policy are required during
the coordinating decision changes in financing policy required?. In case the difference
between the expected commercial, fiscal figures and the desired figures is considered
to be unacceptably high, priorities are evaluated.

Where the priority evaluation task needs to be performed, it is necessary to determine
whether there is any other business which needs attention. Time available from both
support staff and participants is considered. On the basis of this information the
priority of the various projects is determined by the authorizer. The output of this task
is the input of the coordinating decision problem worth analyzing?, see exhibit 5-6.

Decision making can explicitly be described as "negative selection": any solution that
is acceptable for all participants is selected. A decision-process structure which
stimulates generation of alternatives or a systematic choice amongst the available
options, lacks completely. The decision-making task is performed by means of a
negative selection process. Any solution that might contribute to solving the problem
is decided upon under the condition that none of the participants opposes. The
decision-making task, which is the focus of our attention, in shown in exhibit 5-8.

''

propose
i partial
solution

no 1'

all participants?

path: problem solving decision making

Exhibit 5-8. Structure of the decision-making task in the descriptive conceptual model.

During the task propose partial solution, a solution is proposed. During the coordinating
decision acceptable to all participants? is determined whether any participant objects.
If not, the alternative is selected. The analysis does not disclose more detail of the
decision-making task.

Only a very limited set of options is discussed. In chapter 4, the decision-making task
was divided into preparatory design, integrative design and choice. Within the
descriptive conceptual model, none of these tasks can be discerned. Activities which
should constitute the preparatory design task, for example, are performed as a last

84

Sec. 5.5 DESCRIPTIVE CONCEPTUAL MODEL

check before the proposed solution is implemented. Integrative design and choice are
lacking completely.

"Partial decision making" constitutes the decision-making task: a part of a problem is
solved by one isolated solution and not considered in relation to other causes of
financing problems or other solutions. Between the various partial decisions, a
considerable amount of time elapses. Partial decision making is defined as the opposite
of "integral decision making", see chapter 4. "Partial" is this respect can be interpreted
in two ways: not all participants are involved in the decision process simultaneously
and not all aspects of the problem are considered simultaneously.

Aggregation and disaggregation are not explicitly separated. Participants may put
forward arguments on the aggregate and the disaggregate level simultaneously.


No management-control issues are addressed. The meeting is in line with our point
of departure of team setting. The participants work together as a team in the best
interest of the company. Expertise sharing is the dominant characteristic of the
interaction between the participants. All participants are involved in selecting
alternatives. The authorizer starts the decision process, chairs the meetings and
authorizes the decisions made. Where he is not allowed to authorize the decisions
made, he submits the decision for authorization at higher management levels. Support
staff provides input in the form of computerized reports.


Computerized support is provided in the form of reports and not via the telephone
during the meetings itself. These reports are made by means of ad-hoc modeling using
two-dimensional spreadsheets. Layout and contents is not discussed in advance with the
participants. "What-if" analysis during the meeting itself is impossible.

5.6 Conclusion
From the description of the descriptive conceptual model it is clear that the situation
with regard to transfer pricing is unsatisfactory when compared to the guidelines as
described in the previous chapter. This feeling of dissatisfaction is also encountered
when discussing the situation with the organizational members involved.

Main points calling for improvement concern problem specification, decision-process
structure and computerized support. Support should be provided for both inside as
well as outside group sessions.

Except for the facilitator role, the roles of organizational members are already similar
to those described in the previous chapter. The point of departure of group-decision
making and team setting proves to be realistic.

85


A last point to make is that the problem-solving task described in this chapter is
similar to the problem-solving task described in the previous chapter. In both cases the
following four tasks are discerned: scan, decision making, authorization and
implementation. In other words, as the design guidelines as presented in chapter 3
already match the existing situation, we do not see much room for improvement. This
in contrast to the decision-making task.

The decision making-task described in this chapter lacks structure and is based upon
negative selection. According to the design guidelines as presented in the previous
chapter, however, decision making should constitute preparatory design, integrative
design and choice. This implies that the decision-making task is an area that explicitly
calls for improvement.

86

6

TRANSFER PRICING AT EUROPHAR: A SOLUTION
6.1 Introduction
In chapter 1 our research approach has been explained: when analyzing the existing
situation (the descriptive models) we go from the empirical model to the conceptual
model. In this chapter the prescriptive models are described.

In section 6.2, a prescriptive conceptual model is described. This model forms the
answer on our first research question (A') defined in chapter 1:

What are the characteristics of the prescriptive conceptual model?

As explained in chapter 1, this model is based upon information on the pharmaceutical
industry (chapter 2), transfer-pricing literature (chapter 3), group decision support
literature (chapter 4) and the descriptive conceptual model (chapter 5). The concepts
used in the current chapter are explained in these earlier parts of the book. Without
these explanations, this chapter will be difficult to read.

Although the design of the descriptive conceptual model is based upon various decision
situations, the existing situation has been described in detail for only one local
company of the Europhar group. The implementation of the concept is done for the
same local company: Europhar UK Ltd. The prescriptive empirical model is presented
in section 6.3. Decision making on the basis of this model during a period of six
months is presented in section 6.4.

6.2 Prescriptive conceptual model
In chapter 4 we defined four aspects to describe decision situations and group decision
support systems: problem specification, decision-process structure, roles of organizational
members and computerized support. These four aspects will now be used to describe
our prescriptive models in this chapter. This in order to facilitate the comparison
between the old and the new situation. The prescriptive conceptual model combines
the "description" as presented in the previous chapter and the "design guidelines" as
presented in chapter 4. For a better understanding of the prescriptive conceptual
model, the areas of change are indicated in exhibit 6-1. It should be realized that the
classification into one of the three categories is sometimes arbitrary. The main aim of
exhibit 6-1 is to indicate the major difference between the prescriptive conceptual
model and the descriptive conceptual model: the structure of the decision-making task.

87

TRANSFER PRICING AT EUROPHAR: A SOLUTION Ch. 6

the prescriptive conceptual model compared to the descriptive conceptual model
problem specification
distinction between liquidity shortage and liquidity surplus
distinction between short and long term
distinction between operational, relocation and financing problems
explicit consideration of the eight categories of financing instruments
decision-process structure
scan task
authorization task
implementation task
preparatory-design task: actualize and review
integrative-design task: feasible-solution search / disaggregation
choice task: preference ordering
group session
roles of organizational members
authorizer role
participant role
analyst role
facilitator role

computerized support
accounting module (component 1)
regulation database (component 1)
preference-ordering module (component 2)
financial-impact module (component 2)
disaggregation module (component 3)
reference database (component 3)

legend:
-: no change
c: change
*: completely new

Exhibit 6-1. The prescriptive conceptual model compared to the descriptive conceptual
model.


Once a financing problem arises, it needs to be defined as a liquidity-shortage or as
a liquidity-surplus problem, see chapter 3. This is determined directly after the start
of the decision process. With regard to this aspect, no difference with the old situation
exists.

Although operational and relocation problems are of great importance to a company,
they should be excluded from the financing-decision process, see chapter 3. Including
operational and relocation aspects in the financing decision process creates a myriad
of related problems, see the descriptive conceptual model. Decision making should take
place in clear iterations. Firstly, it is determined what is going to happen in the
operational and relocation area. Once that has been determined, the financing problems
are reviewed.

It can happen that financing problems cannot be solved. This implies that a solution
must be found in the operational or relocation area. The fact that these areas are
related, however, does not imply that decision making should take place simultaneously.
Among other things operational and relocation problems require other organizational

Sec. 6.2 PRESCRIPTIVE CONCEPTUAL MODEL

members to be involved. After feedback from the operational and relocation area,
financing decision making takes place.

Related to the limitation of the problem to the financing area is the distinction of two
time perspectives: short term and long term. Financing decisions should be made for
the period for which a detailed forecast is available: the short term. These figures are
usually available for the next one or two years. With regard to the long term, the set
up of forecasts for financing problems alone is not a viable approach. The reason is
that, in order to create reliable figures, many organizational members need to be
involved, usually of a high organizational level. It is unrealistic to expect that the heavy
procedures, which are necessary to produce these figures, will be set up in the
organization, only to solve the problems in the financing area. In other words, financing
decision making should take the "infrastructure" which already exists into account. If
no general accepted forecasts exist for the long term, financing decision making should
do without. This is not to say that the expectations of the participants should not be
taken into account. But it should not be tried, for example, to perform sophisticated
"what-if' analysis for the long term.

All eight categories of financing alternatives, (for instance transfer price, royalty, and
current account, see chapter 3) should be considered explicitly. It will be clear that not
all categories are relevant in each case, but none of the alternatives should be
disregarded in advance. In other words, the preparatory-design task should include the
analysis of all eight categories of financing instruments.

With regard to the criteria against which the alternatives should be judged, we follow
the two points of departure as explained in chapter 4: all criteria are welcomed and
a checklist of possible criteria is made in advance. That we welcome all criteria implies
that no criteria will be considered unacceptable in advance: during the selecting of the
criteria, any criterion can be brought up. The use of the checklist should avoid that
some of the criteria are forgotten, see chapter 4.


Structuring the problem-solving task is the core of the prescriptive conceptual model.
As explained in chapter 4, we use the following general process structure for problem
solving:
1. scan task
2. decision-making task
3. authorization task
4. implementation task
With regard to the structure of problem solving, see exhibit 6-2, the prescriptive
conceptual model is similar to both the descriptive conceptual model as presented in
the previous chapter and the design guidelines as presented in chapter 4. As the
existing situation already matches the desired situation as included in the design
guidelines, there is no need for change.

89


<j problem worth
analyzing?

yes f

decision
making

alternatives for
which authorization at determine ♦ problem pare for
higher level is required solved I O- which no
unsolved problem part for solution is
pare which a solution found
is found

solution
submission formal search in the
for agreement operational
authorization by or
at higher authorizer relocation
level area

Ó determine authorization
V level required

-o-
alternatives
authorized?
yes alternatives for which
authorizer involvement
is sufficient

implementation


Exhibit 6-2. Structure of the problem-solving task in the prescriptive conceptual model.

The scan task is identical to the situation as described in the previous chapter: the
global analysis of the environment is performed by all participants. On the basis of
MIS reporting, "hearsay" and other possible information sources, participants decide

90


whether further analysis of the financing situation is desirable. MIS reporting is one
of the main sources of information as it provides profit and loss figures, balance sheet
and the source and application of funds on a monthly basis.

If the problem is considered to be worth analyzing, the decision-making task is started.
If there is no need to start this task, the scan process continues. With regard to the
coordinating decision problem worth analyzing?, the situation is identical to the situation
as described in the previous chapter. (For the explanation of the term "coordinating
decision", see chapter 4.) The priority evaluation is the input for this decision.

It is important to note that the coordinating decision determine solved/unsolved part
differs from the one described in the previous chapter. One of our design guidelines
is that a distinction needs to be made between decision making with regard to the
financing, relocation and operational area. For the unsolved part of the problem, by
definition a solution needs to be found in the operational or relocation area. Other
organizational members, usually of a higher hierarchial level, need to be involved. On
the basis of the outcome of decision making in the operational and relocation area,
a new iteration of financing decision making might be initiated via the scan task. Input
and output of the task solution search in the operational or relocation area is the
unsolved part of the financing problem, which consists of the target financing,
commercial and fiscal figures in combination with expectedfigures(after implementation
of the financing alternatives). Solutions which are found on the basis of this search
may influence, in their turn, financing decisions via the scan process. As described in
chapter 5, in the descriptive conceptual model, no explicit distinction was made between
operational, relocation and financing decisions.

Note that, with regard to authorization and implementation procedures, (represented
by three tasks and two coordinating decisions) the prescriptive conceptual model is
similar to the descriptive conceptual model as presented in chapter 5 as the situation
does not call for improvement.

The decision-making task is our focus of attention. In chapter 4, we discerned the
following three tasks within decision making as our design guidelines:
1. preparatory-design task,
2. integrative-design task,
3. choice task.
As explained in chapter 4, these three tasks do not yet describe the decide tasks in
detail. They are merely the application of existing decision-process structures to group
decision processes. Detail description depends on the class of problems at hand. We
now apply this task structure to financing problems, see exhibit 6-3.

The main differences in process structure between the prescriptive and descriptive
model takes place within the decision-making task. The reader will notice the big
difference: in the descriptive model, as described in the previous chapter, the decision-
making task is based upon "negative selection" whereas in exhibit 6-3 a distinct
sequence of steps is shown. The explicit breakdown of the decision process in various
steps brings structure to the "myriad of problems" as described in chapter 5.

91


preparatory actualize
design and
review

(re)set
financing
commercial
and
fiscal
targets

integrative
design required
to be
I
target rese
but proved
impossible? *
no

feasible-
solution
search

<
) prefer
no orderii
requir id?
yes ,

' preference
ordering i

choice
''
budget
budget I target
recalculation and not all
alternatives
considered

analyze budget '
and alter latives (b *—
budget # target and

alternatives exhausted
# means: does not equal; path: problem solving decision making

Exhibit 6-3. Structure of the decision-making task in the prescriptive conceptual model.

92


In the preparatory-design task the review of financing alternatives and the actualization
of budget figures takes place. Hence the name we choose for this task: actualize and
review. We name the primary integrative-design task feasible-solution search as the main
aim of this task is to determine the solutions that are feasible in view of the
regulations that are issued by the various authorities. The main choice task is named
preference ordering as the choice between the alternatives is based upon their order of
preference. In theory, each task should be performed during each iteration of the
decision process. As exhibit 6-3 shows, a new iteration takes place after the budget
recalculation except in case the budget figures equal the target. In practice, however,
tasks can be skipped when the outcome of an earlier iteration can still be applied.

During the actualize and review task the actualization of budget figures and a detailed
review of all eight categories of financing alternatives takes place. The input of the
actualize task is the existing budget. The output is the actualized budget, which can
be regarded as a short term forecast. All participants are involved in the actualize task.
Usually, this preparatory task is performed individually or in small groups. The
participants from the marketing discipline, for instance, take care of determining the
expected sales and the financing people within the group of participants review, for
example, expected inventories and the credit side of the balance sheet. People from
outside the group, like support staff, are usually involved.

When the actualization of the budget has been performed, all eight categories of
financing alternatives, see chapter 3, come under review. Participants contact other
organizational members and external consultants for relevant information. The fiscal
expert, for instance, might contact the fiscal consultant to obtain necessary information
with regard to the fiscal unit of which the object company is a part. If transfer-price
changes of various source companies are to be considered, these source companies
need to be contacted. The review of the alternatives "transfer-price change" or "internal
selling-price change" takes place on both the aggregate and the disaggregate level. As
explained in chapter 4, the number of alternatives generated during this task depends
upon time available and the creativity of the participants. In summary, input of the
review activity is any information from any possible source. Output is a piece of
integrated and updated background information per reviewed alternative. This output
serves as input for the group session.

The second task in exhibit 6-3 is the determination of the financing, commercial and
fiscal targets. This task should take place during a group session. We define this first
part of the integrative-design task as: (re)set financing, commercial and fiscal targets.
Output of this task is the (adapted) financing, commercial and fiscal targets. Input is
the actualized budget figures, the long term financing, fiscal and commercial policy,
regulations and any other relevant consideration (like the position of other parts of the
fiscal unit).

When a reset of targets is required because the targets cannot be reached by using
the financing alternatives, it may be possible that a target reset is considered
undesirable in view of the long term financial, commercial and fiscal policy. In this

93


case, a solution needs to be found outside the financing area. This is described by the
coordinating decision: target reset required but proved to be impossible?

The third task in exhibit 6-3 is feasible-solution search. Various iterations of this task
take place. To start with, a part of the problem is focussed on. For this part of the
problem, all categories of alternatives are considered. For each alternative, the maximal
amount of funds that may be transferred is determined (the maximal remittance).
Constraints are the long-term financial, commercial and fiscal company policy.
Regulations on both aggregate and disaggregate level are interpreted. Depending on
the specific problem at hand, it is determined which gross margin range is acceptable.
On the disaggregate level, reference prices are considered. For all financing instruments
under consideration, the limits are determined. In summary, input for this task is the
targets, regulations, the budget, the reference prices and background information per
alternative. Output of this task for each iteration is the total remittance which should
be obtained and the maximal remittance per financing alternative.

We now turn to the various parts of the choice task. When all feasible options
together provide more possibilities than are needed for the transfer of funds,
preference ordering of these feasible solutions is required. This is indicated by the
coordinating decision: preference ordering required?. If no preference ordering is
required, the budget recalculation can take place directly.

During the preference-ordering task, all financing alternatives of the current iteration are
considered simultaneously. Only one of the aspects that is considered in detail now,
is the cost of the various alternatives on a consolidated basis (the "financial impact",
see chapter 3). Once these costs have been determined, a choice is made by means
of a multi-criteria technique. Criteria concerning the alternatives, on the basis of which
a selection is made, can be taken from a criteria checklist, if available. Input for this
task is the total remittance which should be obtained during the current iteration, the
maximal remittance per alternative, regulations, financial parameters and the criteria
checklist. Output of this task is the actual remittance per financing alternative for this
iteration.

With the chosen alternatives included in the budget, the task budget recalculation is
performed. This recalculation offers the possibility to determine interaction effects
between the alternatives. The input of this task is the financing alternatives and the
budget. The budget (after recalculation) is also the output.

In case the target has been reached, the decision-making task is finished. (As explained
in chapter 4, in the representation technique that we use, this exit of a task is not
indicated by an arrow.) In exhibit 6-3 this is shown by the coordinating decision
analyze budget and alternatives. For a new iteration two possibilities exist:
- The target has not been reached yet and not all reviewed alternatives are
considered. The next iteration starts with feasible-solution search concerning a new
set of reviewed alternatives.
- The target is not reached and the feasible alternatives are already exhausted. In this
case a reset of the targets is required.

94


Before turning to the roles of the organizational members, two tasks are worked out
in more detail: feasible-solution search and preference ordering.

select
subset of
alternatives
and set
total
remittance

select
alternative
and determine
disaggregation maximal
remittance at
aggregate
level

yes
O transfer prices/
internal selling
prices considered and
and disaggregation
still to be done F

all alternatives of Q
subset considered?
yes X

summarize
information
on subset of
alternatives

path: problem solving decision making feasible-solution search

Exhibit 6-4. Structure of the feasible-solution search task.

The feasible-solution search task, see exhibit 6-4, starts with the task select subset of
alternatives and set total remittance. These alternatives focus on a part of the problem.
It is possible to make a distinction between the financing instruments that have a direct
impact on profit before interest and taxation (like transfer pricing and cost transfers)
and the financing instruments that are mainly capital transfers (like a capital increase
or a change in the current account position). Another possibility is to focus on one
of the targets at the time. When financial problems are urgent (for instance in case
of a large liquidity shortage), the group might start with solving the financing problem.
In a second and third iteration, the group might then try to find compensation for
undesired commercial and fiscal effect that may be caused by solving the financing
problem. During the select subset of alternatives task, regulations and background

95


information on the various financing alternatives are considered. This background
information is gathered during the actualize and review task. Input for the select subset
of alternatives task is background information on the alternatives of the subset,
regulations and the financing, commercial and fiscal target. Output is the subset of
alternatives and the total remittance for the subset. It will be clear that after the last
iteration of decision making, the target variables as calculated during the budget
recalculation should be equal to the target as determined during the task (re)set
financing, commercial andfiscaltargets.

Once the total remittance has been set, the task select alternative and determine
maximal remittance at aggregate level is performed. For each alternative the maximal
remittance (output) is determined on the basis of the following input: budget,
regulations, background information per alternative and the targets. In case the answer
on the coordinating decision transfer prices or internal selling prices considered and
disaggregation still to be done? is "yes", the disaggregation task has to take place.

During the disaggregation task, the aggregate-level alternative, reference prices and
regulations (input) are considered in order to translate the aggregate-level alternative
into one or more price changes of individual articles (output). The aggregate and
disaggregate level cannot be strictly separated. As is shown in exhibit 6-4, effects are
considered on both the aggregate and the disaggregate level. In a systematic way the
effect of price changes on (aggregate-level) financial, commercial and fiscal figures is
determined. To obtain a desired aggregate-level effect, disaggregation takes place in a
systematic way. As computerized support is indispensable to perform the disaggregation
task, this subject is covered in more detail later on.

In case there is no need to perform the disaggregation task, or in case this task is
already performed, the decision process continues with the coordinating decision all
alternatives of the subset considered?. In case the answer is "no", the next alternative is
taken. In case the answer is "yes", the task summarize information on subset of
alternatives follows. During this small task, the maximal remittances per alternative of
the subset (input) are summarized. The summarized information constitutes the output.
During the feasible-solution task, all participants are involved in group discussion.

Another task that is discussed in more detail is the preference-ordering task, see
exhibit 6-5. This task starts with the task select relevant criteria. The criteria which
determine the preference ordering are the output of this task. One of the criteria
which is discussed here is the financial impact, see chapter 3. On the basis of the
criteria checklist, a first selection of criteria is made. On the basis of regulations and
background information, new criteria, which were not yet included in the checklist, may
be added. Regulations, background information on the alternatives, the criteria checklist
and the financial impact is the input of this task. Output is a list with relevant criteria.

96


select
relevant
criteria

determine
actual
remittance
per
participant

outcome for each
participant identical?

discuss
differences
and make
group
decision

path: problem solving decision making preference ordering

Exhibit 6-5. Structure of the preference-ordering task.

The second task in exhibit 6-5 is determine actual remittances per participant. This is the
only group-session task that is performed individually. For the performance of this task,
a multi-criteria technique is used, see chapter 4. The participants determine a weight
for each criterion that is considered relevant. Per alternative/criterion combination, each
participant determines a score. On the basis of weights and scores, the preference
order of the alternatives is calculated. This preference order in combination with the
maximal remittances, determine the selected actual remittances per participant. Input
for this task is the background information per alternative, regulations and the hst with
relevant criteria. The output is a list of actual remittances per participant.

In case differences in actual remittances exist, a discussion amongst participants needs
to take place. During the coordinating decision outcome for each participant identical?
it is decided whether this group discussion needs to take place.

When the outcome on this coordinating decision is "no", the next task of preference
ordering is discuss differences and make a decision. Differences in weights and scores
are spotted and discussed. On the basis of a group discussion, a group decision is
made. Input for this task is the actual remittances, weights and scores per participant.
The output is the group decision, including the actual remittances for the subset.

97


So far, we have mainly discussed the task structure of the decision process. A second
aspect which needs attention is how decision making takes place. In the chapters 4 and
5, we explained the difference between partial and integral decision making. As far as
feasible-solution search and preference ordering are concerned, integral decision making
should take place. In other words, these tasks should be performed during a group
session. The actualize and review task can be performed individually or in subgroups.


As explained in chapter 4, we take team setting and expertise sharing as our point of
departure.

Most roles of organizational members do not differ from those of the descriptive
conceptual model: the authorizer, the participants and the analysts perform similar
tasks. For the analysts only one task is added: to provide on-line computer support
during the group session.

What is new with regard to the existing roles, is that decision making should not only
take place during sub-group meetings, but that a group session should also take place.
As explained in chapter 4, all participants are explicitly defined as members of this
decision-making group and in this group, integral decision making is of importance.
During such a group session all participants are actively involved in the decision
process.

With regard to most roles, no great distinction exists between the descriptive and the
prescriptive situation as the existing pattern is considered to be already in accordance
with transfer-pricing theory and decision-support theory. The difference is that all
participants are defined as group members. The decision-process structure makes clear
what is expected from the participants.

In the prescriptive conceptual model, however, one role is completely new. This is the
role of the facilitator. As explained in chapter 4, the facilitator is a member of the
support staff. He is the actual leader of the decision process, gives a short conclusion
of the different points of view and integrates propositions of possible solutions. The
decision process is structured by the facilitator, who is a sort of interface between the
decision group and the analysts. He asks the analysts for specific output and can
explain, if desirable, the outcomes achieved. We choose to include this role in our
prescriptive conceptual design as this role, when properly performed, will guarantee
that the participants stick to the designed decision-process structure. It contributes to
proper use of the designed software. For the set-up of the ad-hoc decision room we
follow the design guidelines as developed earlier. Two personal computers, two public
screens and a printer form the main elements, see chapter 4.


The fourth and last aspect of the prescriptive conceptual model is the specification of
the designed computer support. For the following three tasks, three specific decision

98


support components, see chapter 4, are designed: feasible-solution search, preference
ordering and disaggregation. This is not to say, however, that the use of the support is
limited to these tasks only, but for the system design, one of these tasks is kept in
mind. Each support component should exist of two modules: a "what-if' tool and a
database (When we mention "what-if' spreadsheet software, target seeking is included).
Menus and a help facility are part of the human interface. Per component, the design
is described below. In this description, we limit ourselves to the functional
specifications. Menu layouts will be used to illustrate these specifications. Using the
approach as described in chapter 4, the computerized support is presented in exhibit
6-6.
databases models
-
database component 1 module
nil

module component 2 ordering module
8-8 f E

36 component 3 module

ik 1i

user interface
menus
help facility
logging

Exhibit 6-6. Global outline of the computerized support.

Note the inclusion of a logging facility in the design. The logging facility is not a part
of the computerized support. It is designed and used for research purposes only.
Together with audio tapes, it will be used to describe the steps of the decision process
and the use of the various modules.

Regulation database (component 1)
As explained above, the regulations are of vital importance in the decision process. The
regulation database should offer detailed information on regulations concerning
financing alternatives. This information is generally filed per authority by means of legal
textbooks, correspondence, etc. When considering all aspects (financing, commercial and
fiscal) it is not easy to find the right piece of information. A second problem is that
legal texts are very hard to read for non experts. In these texts, concrete
interpretations are lacking. For the regulation database, two design characteristics are
of importance:
- The regulations should be stored per alternative and not per authority. This means
that all regulations concerning one financing instrument should be stored under the
same heading. A distinction is made between the use of a certain instrument in a

99


liquidity-shortage or in a liquidity-surplus situation as the approach of the various
authorities is influenced by their financial interest. So the direction of the cash flow
is considered, see chapter 3.
- Interpretations of the regulations, as made by participants, other organizational
members and external consultants should be included. The text should be
understandable for anyone involved in intracompany money transfers. Following the
approach taken by the International Bureau of Fiscal Documentation, see
IBFD [1986], information is stored in text mode.

Accounting module (component 1)
The accounting module is designed to perform "what-if calculations at the aggregate
level. Various sets of figures should be related to each other. The first set represents
the MIS (management information system) figures. On the basis of these figures, the
local company activities are judged by local and central management. As explained
in chapter 3, the MIS is an internal reporting system. These figures should be
translated into the statutory figures. These statutory figures form the basis for all
external reporting. Various kinds of external reporting may exist. Fiscal reporting is
usually one of them. Another kind of external reporting may be for health-care
authorities (pharmaceutical price control), central banks (foreign-exchange control), or
capital-base requirements (avoidance of capital export).

The accounting module offers the possibility to gain insight in financing, commercial
and fiscal consequences, see chapter 3, when MIS figures are changed. It offers
extensive facilities for "scenario analysis" and "sensitivity analysis". The sensitivity analysis
is mainly concerned with determining the impact of financial parameters like exchange
interest and inflation rates. Changes in sales volume and inventory period can be
considered. The scenario-analysis option offers the possibility to determine the impact
of the various (combinations of) financing alternatives on key figures. These key figures
may reflect the financing, commercial, fiscal or currency situation.

We divide the variables of this module into four categories, see exhibit 6-7.

instrument variable

exogenous variable •
parameter
variable

endogenous variable

Exhibit 6-7. Four types of variables.

The most important endogenous variables are the key figures which reflect, amongst
others, the financing and fiscal situation. These key figures are defined as "target
variables". The exogenous variables can be categorized into "instrument variables" and
"parameters". All exogenous variables of which the size is decided upon by the
decision-making group are defined as "instrument variables". These instrument variables

100


(the financing alternatives) are related to the scenario analysis. The exogenous variables
that are not determined by the decision-making group are defined as "parameters".
Note that according to this definition the interest rate for an intracompany loan, for
example, is an instrument variable. The interest rate for an external bank loan,
however, is a parameter.

In exhibit 6-8, the main menu of the accounting module is presented.


main menu:

1) MIS profit and loss account
2) MIS balance start, January 1st
3) MIS balance close, December 31st
A) MIS source and application of funds
5) fiscal profit calculation
6) other calculations required
7) summary key figures
8) data entry financing alternatives, liquidity shortage
9) data entry financing alternatives, liquidity surplus
10) data entry sensitivity analysis
11) help message
12) exit

enter your choice (1-12) :

Exhibit 6-8. Main menu of the accounting module.

The options 1 through 4 of this menu should give extensive information on the local
company's situation on the basis of MIS figures. The options 5 and 6 should give the
fiscal-profit calculation and any other calculation that might be required. Option 7
(summary of key figures) requires special attention. By means of a simple set of key
figures, an overall picture of the situation should be presented. For the screen behind
option 7, see exhibit 6-9.

summary key figures:
MIS profit/loss:
liquidity surplus/shortage:
fiscal profit/loss:
other key figures:

Exhibit 6-9. Summary key-figures screen.

The screen as presented in exhibit 6-9 looks quite simple. Calculation rules, however,
can be highly complex. One of the complexities that influences the calculation rules is
that the impact of fiscal or other regulations may depend on the destination. Price-
regulation schemes, for example, may treat home sales differently from export sales.
The same might apply for currency-exchange regulations. Special tax rules may apply

101


in case of inter-concern sales. The logical design of the module, should reflect these
issues.

Financial-impact module (component 2)
The database of this component is called "financial-impact module". It is called a
module because it offers limited "what-if' possibilities with regard to parameters like
tax and exchange rates. Its main use, however, is as a database. The aim of this
database is to provide insight in the cost on a consolidated basis for each of the
financing alternatives.

It is important that the financial impact of all financing alternatives are compared in
a consistent way. We use the method of calculation that has been worked out in detail
in chapter 3. For each financing alternative, a calculation similar to the calculation as
presented in chapter 3 is made. As we explained, in case local borrowing or local
lending is considered, interest effects should be incorporated in the calculation. The
main menu of the financial-impact module is shown in exhibit 6-10.


main menu:

1) liquidity shortage without local borrowing
2) liquidity shortage with local borrowing
3) liquidity surplus without local lending
4) liquidity surplus with local lending
5) change financial parameters
6) help message
7) exit

enter your choice (1-7) : ■

Exhibit 6-10. Main menu of the financial-impact module.

As explained in chapter 3, eight categories of financing alternatives exist. Within each
of these categories, differences in financial impact occur. A transfer-price change from
a source company which is located in a different country has a different financial
impact due to, for instance, different currency-exchange, fiscal and duty effects. The
database reflects these differences.

Preference-ordering module (component 2)
In chapter 4, one multi-criteria technique has been worked out in detail. On this
technique the design of the preference-ordering module is based. In chapter 4, we
explained our logical design by means of an example: a selection amongst financing
alternatives. For the explanation of the terms "opinion", "weight" and "normalized
weight", "score", "total score" and "preference order", we refer to chapter 4. In the
preference-ordering module this multi-criteria technique should be combined with a few
calculations which will now be explained.

102


During the feasible-solution search task, the maximal remittance per alternative and the
total remittance required for that iteration are determined. The module should calculate
the actual remittance per alternative on the basis of the preference order. We will now
illustrate this by means of an example:

Let us assume that we have two alternatives (1 and 2) and that the total remittance
required is X. The maximal amount, determined during the feasible-solution search task
for alternative 1 is Y. For alternative 2 it is Z. We presume that X > Y and that
alternative 1 is preferred. Alternative 1 should be used to the maximum: an amount
of Y. To obtain the total requirement of X, an additional amount of X -- Y is
needed. The second alternative (alternative 2) should be used for this amount. So on
the basis of the scores and weights entered by the participants, the actual remittance
per participant should directly be calculated.

The design as described above, however, presents decision support for one participant
only: each participant has his own "preference-ordering worksheet" to perform his own
multi-criteria calculations. The following design characteristics make it possible to cope
with group decisions and stimulate effective communication. The preference-ordering
module should be able to present an overall picture for all participants. Using the
simple example from above and presuming 6 participants, we could get the screen as
shown in exhibit 6-11.


actual remittances per participant
alternative 1 alternative 2
participant 1 Y X--Y
participant 4 X--Z Z

Exhibit 6-11. Actual-remittances screen.

This screen forms a good basis for discussion. Participant 4, 5 and 6 have a different
opinion than the others as they prefer alternative 2. Other overview screens should be
possible, for example a screen that shows the weights for the various criteria per
participant. Such a screen would help to focus the discussion on those aspects that
cause the differences within the group, see exhibit 6-12.

Using the same example as in exhibit 6-11, the screen with weights shows a sharp
difference in opinion within the group. With regard to the criteria C and D, the
participants 4, 5 and 6 consider criterion D of much greater importance than criterion
C. After consultation of this screen, the discussion would focus on the criteria C and
D only as with regard to the criteria A and B, no large differences exist. The interface
of the system should be such that it is easy to change the scores. Note that the group
aspects of the tool as described above should facilitate effective communication. We

103


do not consider it necessary to include a "voting tool" which calculates a kind of group
average. Expertise sharing is the aim of the group. Communication within the group
is presumed to be of such quality, that decisions can be made on the basis of consent.
In the exceptional cases that no consent can be reached, the authorizer should cut the
Gordian knot, see chapter 4. The use of a formal voting tool would disturb group
interaction.


nonnaliz;ed weight per participant
in percentages
participants:
1 2 3 4 5 6
criteria:
A 15% 19% 18% 17% 16% 18?
B 54% 50% 52% 52% 52% 51%
C 16% 16% 15% 2% 2% 1%
D 15% 15% 15% 29% 30% 30%
100% 100% 100% 100% 100% 100%

Exhibit 6-12. Screen with weights per participants.

Although a computer is used for performing fast multi-criteria calculations, participants
do not key in their own scores, see chapter 4. Each participant puts his scores down
on a multi-criteria worksheet. The analysts use these worksheets to key in the scores.

Reference database (component 3)
As explained in chapter 3, prices of reference articles play a vital role in determining
the transfer price to the object company. To have all reference information available
when necessary, the reference database provides the following information per source
article:
- transfer price to the object company,
- (adjusted) strategic minimum price and / or world market price,
- (adjusted) reference prices,
- article descriptions and article numbers of reference articles.
This reference database is designed to give information on reference prices of directly
comparable articles. The definition of "directly comparable" depends on company
strategy and the kind of articles considered. Although prices of indirectly comparable
articles are not stored in the database, in exceptional cases, these prices might be
relevant. Therefore, information on the existence of indirectly comparable articles should
be stored.

The concept of adjusted reference price needs some additional explanation. The
adjusted reference price is, for example, the reference price adjusted for pack-size
differences: in case a source article containing 20 tablets is transferred to the object
company for the price of NLG 33.—, and a reference article containing 40 tablets is
transferred to a reference company for the price of NLG 60.—, the adjusted reference
price is NLG 60.-- * 0.5 = NLG 30.--. As will be clear, the adjustment factor of 0.5
is this case, is the ratio of the pack sizes (20/40).

104


In case articles with a different dosage of active substance are considered to be
directly comparable, the adjusted reference price should take these differences in
dosage into account.

The reference database should facilitate insight in the minimum and the maximum of
the adjusted reference prices per article. By providing these two figures, participants
are able to judge reference effects of a proposed transfer price at one glance.

Disaggregation module (component 3)
The disaggregation module is a calculation tool that is designed to link aggregate-level
effects with price changes of individual articles. It can be used to determine the
sensitivity of a chosen alternative for changes in parameters like exchange rate or sales
volume. This module is closely related to the accounting module in which the
aggregated figures are used. Both modules are used in the feasible-solution search task.
The tool is designed for use in two directions: For aggregation ("up") and for
disaggregation ("down").

The main use of the disaggregation module is to calculate the exact impact of a price
change of one or more articles on the aggregate-level gross margin. Per destination,
the set up of the module should be such that the effect of a change in one of the
following external parameters or instrument variables can directly be determined:
- transfer price per article,
- sales volume per article / per destination,
- selling price per article / per destination,
- currency exchange rate,
- duty rate.

Where not only sales, but also production takes place at the object company, it might
happen that one or more source articles are transferred into one or more destination
articles. Then also the conversion factor and the local added value determine the gross
margin. The conversion factor needs some additional explanation. The conversion factor
is defined as the number of source units needed to produce one unit destination
article.

Taking a Swiss object company, for example, the input/output matrix which presents
these conversion factors, might look as presented in exhibit 6-13 ("CHF' stands for
Swiss franc).

output:
destination article: A B
input:
source article X: 1 unit 2 units
source article Y: 3 units 3 units
local added value: CHF 5 . — CHF 6.--

Exhibit 6-13. Example of an input/output matrix presenting conversion factors.

105


This input/output matrix presents the input needed per unit output. We see that the
production of one unit destination article A requires one unit source article X, three
units source article Y and a local added value of CHF 5.--. One unit destination
article B requires two units source article X, three units source article Y and a local
added value of CHF 6.--. Where such interrelationships exist, the disaggregation module
should be able to cope with the required calculations.

It should be possible to use the disaggregation module in the other direction ("down"):
to search for source articles with some desired characteristic. The disaggregation
module could be used, for example, to gain insight into the impact that transfer-price
changes have on the gross margin for the various destinations. For this purpose, the
"destination-ratio screen" is designed. Per source article, the impact that a transfer-
price change would have, is indicated, see exhibit 6-14.


destination ratios for destination Y
transfers total transfers ratio
destination Y
source article A NLG 7,500 NLG 10,000 75*
source article B NLG 100 NLG 5,000 2%
total NLG 7,600 NLG 15,000 51X

Exhibit 6-14. Screen with destination ratios.

A destination article is sold to the various destinations in a certain proportion. Let us
presume that destination article d is produced out of source article A. Article d is sold
to the destinations X and Y. (X could be the home market and Y export third
parties.) 25% of the volume of d is sold to destination X and 75% to destination Y.
A transfer-price decrease increases the gross margin. In case the payments for article
d are decreased with an amount of NLG 100, the gross margin on sales to destination
Y would increase with the equivalent of NLG 75.

In case all transfer prices are changed with the same percentage, 51% of the total
effect is caused by a change in the gross margin for destination Y. In view of
regulations, however, this might be undesirable. It might be preferable to avoid an
impact on the gross margin for destination Y. In that case, source article B would be
a good choice as the destination ratio for destination Y for this article is only 2%.

Let us presume an exchange rate of 2 NLG/CHF. If we would increase the gross
margin of the Swiss object company by CHF 100 by means of a transfer-price decrease
for source article B of 4%, (4% * NLG 5000 = NLG 200 = CHF 100) the increase
of the gross margin on destination Y sales, would only be CHF 2 (2%). The aim of
the screen above is mainly to help in identifying appropriate source articles.

106


Above, the four elements of the group decision support system have been described.
It should be realized that for every financing problem a complete group session and
the use of all modules is not always required. For every (part of) the decision process,
the authorizer and the facilitator should determine which aspects are of importance and
which tasks need to be executed. This should be done during the coordinating decision:
"Problem worth analyzing?", which is part of the scan task.

6.3 Prescriptive empirical model
The prescriptive conceptual model, as described in the previous section, will now be
applied to the Europhar situation. For a detailed explanation of the Europhar
company, see chapter 5. As in chapter 5, names of the divisions of the Eurochem
Holding NV are not always written in full. Usually only the place name is used.
Europhar Int BV in Amsterdam is referred to only as "Amsterdam". Europhar UK Ltd
in London and Eurochem Holding NV in Rotterdam are respectively referred to as
"London" and "Rotterdam". Eurobitec UK Ltd is referred to as "Oxford". As the two
destination companies of London are located in Taiwan and Australia, these country
names are used.

As in the prescriptive conceptual model, the prescriptive empirical model is described
below on the basis of the four aspects of decision support systems as explained in
chapter 4: problem specification, decision-process structure, roles of organizational
members and computerized support.


To start with, the problem at hand should be specified as either a liquidity-shortage
or as a liquidity-surplus situation. (In Year 5, we are confronted with a liquidity
shortage situation.)

Following the prescriptive conceptual model, only financing problems are addressed.
At Europhar in London, packing for Biosulfine takes place. These production activities
can be switched to Eurobitec. This would result in relocation of activities between
Eurobitec and Europhar. For the UK, a sharp distinction between financing problems
on the one hand, and relocation problems on the other hand is especially important.
As the actual decision process will show, this is not to say that no interaction between
the two areas takes place.

With regard to the time perspective, the situation is as follows: the financing situation
for Year 5 is considered at the start of the decision process in April Year 5. (The
budget for Year 6 is not made before autumn.) Decisions taken will have an impact
on the figures of Year 5, hence the short term. Year 5, therefore, should be the focus
of attention. Careful updating of all budget items should take place. This holds for the
profit and loss account as well as for the year-end balance sheet.

107


The prescriptive conceptual model requires that all eight categories of financing options
are analyzed in advance and considered simultaneously. As explained above, each
category exists of more than one option. On the basis of the description of Europhar
UK Ltd and presuming a liquidity shortage, we know that the following options exist:
- transfer-price decrease:
- source company Amsterdam,
- source company Oxford.
- internal selling-price increase:
- destination company Europhar Australia Pty Ltd,
- destination company Europhar Taiwan Ltd.
- cost transfer away from Europhar UK Ltd:
- Any cost item can in principle be subject to this cost transfer. This needs to be
worked out in the actualize and review task. The system is not a great help in
this case. The only thing that can be said is that it is very likely that cost should
be transferred to Amsterdam, as many ties exist between Amsterdam and London.
- royalty decrease:
- The only royalty paid is an annual lump sum royalty to Amsterdam. A decrease
should be considered.
- capital increase:
- A capital increase made by Amsterdam (only shareholder) should be considered.
- intracompany loan to London:
- The only loan to be considered is a loan from Amsterdam.
- current account:
- increase of the payment term from Amsterdam,
- increase of the payment term from Oxford,
- decrease of the payment term to Taiwan,
- decrease of the payment term to Australia.
- local bank loan / overdraft facility:
- The bank loan or overdraft facility should be provided under an Eurochem
guarantee.

As explained above, all kinds of criteria should be welcomed by the facilitator. A
checklist of criteria should aid the participants when considering the preference order
of the financing alternatives. This checklist is set up in advance by means of analysis
(by the participants) of all Europhar financing problems in the period Year 1-Year 4.
The 50 criteria that resulted from this analysis produced the following short checklist
for the UK case:
- financial impact,
- flexibility,
- UK fiscal effects,
- other fiscal effects,
- NHS effects,
- reference effects.


Given the priority evaluation, the following tasks should take place:

108

Sec. 6.3 PRESCRIPTIVE EMPIRICAL MODEL

1. actualization of the budget figures for Year 5 and review of all eight categories of
financing alternatives,
2. feasible-solution search,
3. preference ordering,
4. authorization,
5. implementation.

The actualize and review task should take place during April Year 5. The feasible-
solution search is planned for a two-day group session on 28 and 29 April Year 5.
After the group session, authorization and implementation need to take place.

This decision process will be described in detail in the next section. As explained in
chapter 4, we will concentrate on the decision-making tasks: actualize and review,
feasible-solution search and preference ordering.


In the case of Europhar UK Ltd the following organizational members are involved
in the decision process (for a "who is who", see chapter 5):
- authorizer / participant:
- Katwijk.
- participants:
- Berkers,
- Derecks,
- O'Brian,
- Oostbeek,
- Terwolde.
- analysts:
- Francen,
- Nijboer.
- facilitator:
- the author.
As the list shows, the organizational members that are involved in the decision process
are the same people as before the introduction of the system. As far as the support
staff is considered, however, two new people are involved: Francen and the author.

Francen is a systems analyst working in the information-management department of
Europhar Int BV. He reports to Pieterse (joint steering committee member, see
chapter 5). The facilitator role is completely new. As described in chapter 4, this role
is vital as the facilitator should guide the decision process and ensure that optimal use
is made of the computerized support. Within Europhar, however, nobody is available
in Year 5 to perform this role. This role is therefore taken by the author, who, being
an ex-Europhar employee, is considered an insider. As developer of the system, he
knows the system sufficiently well to perform this task. It should be realized that this
is a second best solution, which needed to be taken due to the experimental character
of the system.

109



For development of the software as designed above, we need a modeling tool with at
least three "dimensions". Three dimensions mean that a variable can be regarded as
a vector with three indices. Another requirement is that the tool is self documenting
in order to limit writing of system documentation to a minimum. It should be user
friendly enough for development of the system by non-programmers and run on IBM-
compatible personal computers as that is both the Delft University of Technology as
well as the Eurochem standard. The tool should offer possibilities to include system
documentation in each module, as documentation on paper is considered impractical,
especially as no standard for a word-processing package existed, either within
Eurochem, or at Delft University. The main aim is to find a practical solution. A last
selection criterion is that besides the personal-computer version, a mainframe version
should be available. This as a mainframe version is needed for large models or
databases. (For the experimental system, however, no mainframe is used). One of the
packages that meets these criteria is Personal Wizard from Comshare, see also Sprague
and Watson [1986].

We decide to use Personal Wizard as the development tool as it is already used within
Europhar. The product is known by Europhar employees, so we can use the in-house
support as provided by, for example, Francen from the information-management
department. Another reason for the choice of Personal Wizard is that Comshare
products are already used for decision support in the financing area by 6 of the top
20 European pharmaceutical companies, see Van der Ven [1989a].

For a detailed description of the modules in the Europhar case, see Van der Ven
[1988a; 1988b; 1988c]. In this subsection we concentrate on the global structure of the
various modules and on the user interface. As all six modules are set up in Personal
Wizard, we use the package as a two-dimensional spreadsheet, a three-dimensional
spreadsheet, a database and a word processor. When convenient, tailor-made menus
are set up. In other cases we use the interface facilities that are offered by the
package itself. Our approach to set up tailor-made menus is a good example of how
we interpret "simple" in the "simple and sufficient" approach.

Menu building in Personal Wizard requires the use of separate DOS files which do
not form an integral part of the module to which the menu screens belong. This is
highly impractical, especially for the non-technical user. Therefore an automatically
start-up routine is developed which creates the necessary DOS files automatically when
the module is retrieved. This approach has been published by the software house
involved, see Comshare [1989]. In this example, "simple" is interpreted from the users
and not from the builders point of view. This same approach is followed when
designing the financial-impact module or the regulation database.

For this experimental system we decide not to build interfaces with already existing
computerized systems, for example the MIS system. It is considered that design and
implementation of interfaces goes beyond the aim of the research project: developing
an initial "0 version", see chapter 1. As figures are entered in the experimental system

110


manually, they need to be checked carefully. Various checking procedures are used.
Firstly, checking variables are included in the software. Debit and credit side of the
balance sheet, for example, need to be identical. Secondly, in order to check total
figures, the original budget is used. Budget figures are entered in the software modules.
The totals included in the budget need to be identical to the totals included in the
software modules. A third and last check is at the detailed level by various people who
check whether the input is correct.

The modules in the Europhar-UK situation are now briefly described below:

Regulation database (component 1)
The categories of rules that are stored in the database are:
- rules from the UK National Health Service (NHS),
- fiscal rules from the UK Inland Revenue,
- Dutch fiscal rules,
- rules as laid down in the UK Companies Act,
- rules concerning duties and value-added tax from the UK customs and excise
authorities,
- Eurochem internal authorization rules.
These rules are considered for each of the eight categories of financing alternatives.
A distinction is made between a liquidity-shortage and a liquidity-surplus situation. For
the problem at hand, only the part concerning a liquidity-shortage situation need to
be used.

An example illustrates the approach. Below we provide the information that is
presented when the UK fiscal aspects of an intracompany loan are considered (liquidity
shortage):

"The interest rate should be at arm's length (not too high) and the interest payment
should be on an annual basis. Arm's length interest rate is related to the currency
of the loan. The amount of the loan should be acceptable in view of the arm's
length equity/debt ratio of Europhar UK Ltd ('Thin capitalization rules'). The arm's
length interest rate in case of a loan in NLG could be determined on the following
two bases:
- Dutch 'promesse disconto' + 1.5%
- LIBOR (London Inter-Bank Offered Rate) + 1%
In case of a loan in GBP, the arm's length interest rate could be determined on
the following two bases:
- British base rate + 1%
- LIBOR + 1%."

For this regulation database, we used the word processing capabilities of Personal
Wizard.

Accounting module (component 1)
This "what-if' tool exists of a three dimensional spreadsheet model with figures and
calculation rules. This "cube" can be used for the manipulation of figures. On the

111


horizontal axis, the various destinations (including total company) are shown, see
exhibit 6-15. The following destinations are presented:
- NHS,
- generics,
- Europhar Taiwan Ltd,
- Europhar Australia Pty Ltd,
- export third parties Nigeria,
- export third parties China.

-71
fiscal
-* .
NHS profit
-calculation
'I I
statutory
'NHS generics etc. total
profit
and loss
account

balance-
sheet
opening

balance-
sheet close

source and
application
of funds

Exhibit 6-15. The accounting module.

On the vertical axis, all relevant sets of financial variables are indicated. Together these
variables compose the profit and loss account, opening and closing balance sheet and
the source and application of funds. Relationships between these variables are put in
the module by means of calculation rules. An example of such a calculation rule is the
relationship between inventories of finished goods on the end-year balance sheet and
value of annual goods sold on the profit and loss account:

inventories in GBP = annual sales in GBP * (inventory period in days / 365).

By changing the inventory period or the annual sales, "what-if analysis can be
performed.

On the third axis, the different internal and external reporting requirements are
indicated. Each of the four vertical "slices" of the cube is used for a different set of
figures. In the UK case, besides MIS, statutory and fiscal figures, the profit calculation
for the NHS is important. Like the other calculations, this profit calculation is
performed in one of the vertical slices. This may be illustrated by means of an
example. The NHS limits promotional costs to a certain maximum. The statutory

112


accounts, however, need to reflect all promotional costs. The calculation rule that
translates statutory promotional costs into NHS promotional costs is now as follows:
1. add all relevant statutory promotional cost items together (this total is defined as
A),
2. calculate the maximal promotional costs as accepted by the NHS (defined as B),
3. take the minimum of A and B.
The figure thus calculated represents the NHS promotional costs and is stored in the
"cell of the cube" identified by the following three indices:
- variable: total promotional cost (profit and loss account),
- destination: NHS,
- financial statement: NHS profit calculation.

Financial-impact module (component 2)
As far as possible, the financing alternatives are worked out in detail. For each source
or destination company, the financial impact of pricing and current-account alternatives
is calculated, see chapter 3. In case of a liquidity shortage with local borrowing as one
of the alternatives, the menu screen in the Europhar case looks as shown in exhibit
6-16.

submenu 2 : liquidity shortage with local borrowing

the 14 reports are (1..14) :
1) local borrowing
2) share capital increase
3) cost charge to Europhar Int BV
4) debit current account, Europhar Australia Pty Ltd
5) debit current account, Europhar Taiwan Ltd
6) credit current account, Eurobitec UK Ltd
7) credit current account, Europhar Int BV
8) intracompany loan from Europhar Int BV
9) royalty decrease
10) internal selling-price increase Europhar Australia Pty Ltd
11) internal selling-price increase Europhar Taiwan Ltd
12) transfer-price decrease Eurobitec UK Ltd
13) transfer-price decrease Europhar Int BV
14) a summary of the results of the financing alternatives (1-13)
15) help message
16) exit to main menu

enter your choice (1-16) :

Exhibit 6-16. Submenu of thefinancial-impactmodule.

The financial-impact module is a small, but complex, two-dimensional spreadsheet.

Preference-ordering module (component 2)
The preference-ordering module is a general tool which needs little adaptation to the
Europhar UK situation, see exhibit 6-17. The individual multi-criteria worksheet of each
participant is one vertical slice of the front of the cube. The summary tables that were
discussed in the previous section, are "slices" of this cube. The actual-remittances
screen, for example, is a horizontal slice of the bottom of the cube. For an explanation

113


of the multi-criteria technique, see chapter 4. For details on the background of the
calculations that are performed, see the previous section.

/ 'etc
/ i
/ / 1
opinion weight normalized transfer etc.
weight price
increase
criteria:
financial. yes 6 24Z 6 ..
.
impact
flexibility yea 10 40X 4 ..
.
etc. ..
. ..
. ..
. ..
.
total score 25 1001 5.4 3.2
preference
order 1 2
maximal
remittance 100 300

actual
remittance 100 250

(total remittance required is 350)

Exhibit 6-17. The preference-ordering module.

Reference database (component 3)
In the case of the UK, the following companies are considered to be the reference
companies:
- Europhar Belgium NV,
- Europhar France SARL,
- Europhar Germany GmbH,
- Europhar Italy SRL,
- Europhar Netherlands BV,
- Europhar Sweden AB,
- Europhar Switzerland GmbH.
For these seven companies, all the prices of the directly comparable articles are
included in the database. In the Europhar case, "directly comparable" is defined as
follows: two articles are directly comparable when they contain per tablet (or ampoule)
the identical active ingredients in an identical dosage.

Disaggregation module (component 3)
In exhibit 6-18, the design of the module is shown. The variables on the horizontal axis
are presented per article (as shown), and as total value per destination (for example:
transfer value = units * transfer price). On the vertical axis the source and destination

114


articles are shown. The third axis is used to indicate the destinations. Note that this
third axis is identical to the horizontal axis of the accounting module.

In general, this module is the largest one in size. In case of 30 variables, 100 articles
and 15 destinations, for example, total size of the module is already 45,000 cells.

total

transfer local local net gross
price added cost selling margin
value price price
Biosulfine/
Sofanaat
Briconcyl

etc.
total

Exhibit 6-18. The disaggregation module.

From a calculation point of view, the production activities in London are
uncomplicated. Packing activities take place and, as far as the Eurobitec articles are
concerned, tablets are produced out of raw material. No complex input/output
relationships exist. A destination article is always produced out of only one source
article and a source article is used in only one destination article.

In the previous section, the concept of the destination ratio has been explained. In the
case of the UK, the NHS-destination ratio is of great importance as the NHS has its
own set of rules to judge the figures of Europhar UK Ltd. By means of the use of
the destination ratio, the participants are able to determine accurately the effect of
transfer-price changes on NHS figures.

Logging facility
The logging facility is not a part of the system itself. In cooperation with the software
house involved it is designed for research purposes, see Comshare [1988]. It is a key
stroke logging facility with automatic replay possibilities. All key-strokes are interpreted
as Personal-Wizard commands and stored in a command file (the logging file). In the
replay mode, the group session or any other use can be replayed at a speed
determined by the researcher. The time and date of the key strokes are recorded in
the logging file. This facilitates analysis of the logging-file printouts.

115


6.4 Decision process after system introduction
In this section, 6 months of system use of the pilot system is described. System use
starts at the end of March Year 5, just after the pilot system is completed and
presented to the joint steering committee.

All software modules are checked carefully and contain the figures that are derived
from the original budget. In the databases, data concerning both a liquidity surplus as
well as a liquidity shortage are included. The software is ready for use.

A preparatory meeting is held on the 1st of April Year 5. The following five people
participate in this meeting: regional manager Katwijk, financial expert Berkers, price
coordinator Terwolde, the facilitator and analyst Nijboer.

These organizational members plan a group session for 28th and 29th April Year 5.
So the participants have 4 weeks to undertake the actualize and review task. Four
more people have to participate in the group session. These are: UK manager O'Brian,
UK controller Derecks, fiscal expert Oostbeek and analyst Francen.

After the group session, Eurobitec proposes changes in the relocation area due to
problems with regard to the production of Biosulfine. These changes require new
iterations of decision making in the period May-September Year 5.

In the following subsections the preparatory meeting, the preparations during the month
of April, the two days of the group session on 28/29 April and the Eurobitec
negotiations are described in detail as these are a good example of how decision
making takes places with the help of the experimental system. The author is, as
facilitator, involved in the decision process during this period. Besides logging files,
audio tapes and internal memos are used to obtain a detailed picture of the decision
process.

In the subsections below, bold headings are used to indicate the tasks. Of the path,
see chapter 4, only the task that is one level higher in the hierarchy, is indicated.
Where the outcome of a coordinating decision is obvious, this coordinating decision
is not explicitly mentioned.

THE PREPARATORY MEETING

Katwijk, Berkers, Terwolde, the facilitator and analyst Nijboer are present at the
preparatory meeting on the 1st of April Year 5.

Task: scan priority evaluation
Katwijk arranged this meeting and he explains why he did so. He expects problems
with sales of the product Retravir. Retravir, one of the current leading Europhar
products, is attacked by a new competitive product which has recently been introduced
in the UK and Australia: Denal from Smits Pharma. It has some advantages over the
Europhar product. On top of this, Smits Pharma's marketing approach seems to be

116

Sec. 6.4 DECISION PROCESS AFTER SYSTEM INTRODUCTION

more effective. Although sales of Retravir are still on target, they are expected to be
20% below budget at the end of Year 5. Budgeted sales for Retravir are GBP 20 mln.
With a gross margin of about 50%, this results in an expected decrease of gross
margin of:

20% * (50% * GBP 20 mln) = GBP 2 mln

The total budgeted sales for Europhar UK (for all products) are GBP 54 mln. As the
entire liquidity surplus of the previous year is already transferred to Amsterdam, this
shortfall on budgeted sales will result in serious financial problems for Europhar UK.
As Retravir is sold not only on the home market, but also in Australia, there will
probably be problems in Australia too.

After explaining this problem, Katwijk inquires whether there are currently other urgent
matters, which need to be worked out first. According to Berkers and Terwolde there
are no other problems which require priority. Then Katwijk authorizes the start of the
decision process.

Task: decision making actualize and review
There are eight categories of financing alternatives which should be reviewed and
discussed in order to find a suitable solution for the problem. Furthermore the year-
end balance sheet and the sales budget should be actualized, because the original
budget does no longer represent the financial situation of Europhar UK. The facilitator
puts all eight categories on the flip chart. The options are (see chapter 3):
- internal selling-price increase, to Taiwan and/or Australia,
- transfer-price decrease, from Oxford and/or Amsterdam,
- cost transfers to Amsterdam,
- decrease of royalty payments to Amsterdam,
- local borrowing (bank loan),
- intracompany loan from Amsterdam,
- capital increase,
- changing the current-account trading position. There are four possibilities: increasing
the creditors payment term from Amsterdam and/or Oxford and decreasing the
debtors payment term to Australia and/or Taiwan.
The alternatives will be worked out in April by the participants with the help of
analyst Nijboer.

Terwolde, as price coordinator, will review the alternatives dealing with prices. These
are internal selling-price increases and transfer-price decreases. Katwijk will contact his
colleague O'Brian and together they will gather data on the cost-transfer alternative
and the royalty decrease. Berkers will contact Derecks and together they will work out
the possibilities in the financing area. These are local borrowing, capital increase and
intracompany loan.

Remaining tasks are actualization of the profit and loss account and actualization of
the year-end balance sheet. The first task is performed by Katwijk and O'Brian and

117


the second by Berkers and Derecks. Oostbeek will be contacted by the other
participants to provide fiscal background information where needed.

Berkers opposes the current-account alternatives, because the Eurochem rule is one
month payment term. It is his view that the group should stick to this Eurochem rule
of 30 days settlement. So there is no need to consider these alternatives, according to
Berkers. The facilitator explains that each alternative should be worked out, even if it
seems to be against the company policy. All alternatives should be studied according
to the "rules of the game". The process structure is designed for this purpose. Berkers
accepts this point of view. He and Derecks will analyze the current-account
alternatives. As all actualize and review tasks are now assigned, Katwijk closes the
preparatory meeting.

PREPARATIONS

During April Year 5, the participants review the different financing options and
actualize the sales budget and the balance sheet. Analyst Nijboer is heavily involved
with the preparations. He performs data input of the actualized budget in the
disaggregation and accounting module. He retrieves the necessary information from the
reference database, regulation database and the financial-impact module when
requested.

Katwijk telephones UK manager O'Brian and they make an appointment for working
out the cost-transfer alternative and the royalty-decrease alternative. They actualize the
sales and cost budget. In the actualized budget all changes in the forecast year-end
balance sheet are included. Part is the information is obtained from Katwijk and
O'Brian. These have been worked out by Berkers and Derecks. By using the
disaggregation module and the accounting module, analyst Nijboer obtains the financial,
commercial and fiscal key figures based on this actualized budget.

Price coordinator Terwolde studies the transfer-price alternatives and the internal
selling-price alternatives. He phones Oostbeek, when he needs some additional fiscal
information. He contacts analyst Nijboer to obtain information from the reference
database, the regulation database and the disaggregation module. As explained in
chapter 4, the participants prefer not to use the computer personally.

Terwolde contacts the management of Eurobitec in Rotterdam and Oxford. He explains
the situation London is in at the moment and asks them to consider a transfer-price
decrease. It is not always easy to get a transfer-price decrease, because the divisions
of the Eurochem company are profit centers and each division is judged on its own
profit. This sometimes creates friction between the divisions.

Terwolde contacts analyst Nijbocr and together they use the reference database in
order to obtain information about European reference prices. It appears that prices to
the UK are relatively low compared to other destinations. However this does not apply
to all products. The transfer price of the generics might be decreased.

118


Terwolde telexes the management of Taiwan and Australia in order to make inquiries
about a possible internal selling-price increase. Apparently an internal selling-price
increase to Australia is not possible, as Australia has the same problems with Retravir
as the UK. However an internal selling-price increase to Taiwan seems a feasible
alternative as Taiwan is in a strong financial position at the moment. However, using
the financial-impact module, Terwolde notices that it is very expensive to transfer
money from Taiwan to London, because of the high import duties.

Terwolde and analyst Nijboer use the disaggregation module in order to compare
prices to Taiwan with prices to other destinations. It appears that prices to Taiwan are
relatively low. Terwolde thinks a price increase of about 20% is feasible. The regional
manager Far East and the local Taiwan management do not object to such an
increase.

UK controller Derecks visits financial expert Berkers in Amsterdam and, together with
analyst Nijboer, they actualize the local borrowing alternative, the intracompany loan
alternative and the capital-increase alternative. When fiscal problems arise during their
meeting, they telephone Oostbeek for fiscal background information. They ask analyst
Nijboer to consult the regulation database to find out who is to authorize an
intracompany loan. This appears to be the Eurochem fiscal & financial department.
Oostbeek discusses the intracompany loan with this department, but they do not
approve.

Note that until now all aspects of the problem are studied in isolation. Studying the
alternatives in isolation should be interpreted in two ways. The financing alternatives
are not considered simultaneously. Moreover the participants analyze these alternatives
individually or with a subgroup.

Before the group session, the facilitator studies the alternatives that are worked out
by the participants and analyst Nijboer. The facilitator visits all participants to explain
the new approach that will be followed. The multi-criteria technique is explained and
used in a small exercise.

FIRST DAY OF THE GROUP SESSION

The first day of the group session takes place on the 28th of April Year 5. The six
participants, including the authorizer, meet together with the two analysts and the
facilitator in one of the meeting rooms of the Europhar head office in Amsterdam.
This same meeting room has been used for the meeting in July Year 4 which has been
described in chapter 5. The layout of this meeting room is shown in chapter 4.

Before presenting a detailed account of this two day session we like to describe the
atmosphere of the meeting. Like in the meeting described in chapter 5, the atmosphere
is task oriented, informal and open. Relief of tension within the group takes place by
means of a relaxed and humorous yet businesslike approach.

119


As authorizer, Katwijk opens the meeting and asks the facilitator to structure the
decision process in the coming two days. The facilitator takes over and gives a brief
review of the decision support system that will be used. Special attention is paid to
distinction between the authorizer and facilitator role, see chapter 4. After this
introduction, the decision process starts.

Résumé of the actualize and review task
First the results of the actualization of the sales budget and the balance sheet are
discussed together with the eight financing options. The actualization of the sales
budget by Katwijk and O'Brian resulted in a slightly less unfavorable picture due to
better sales of a few small products: GBP 1.5 mm below budgeted sales (In the
preparatory meeting the shortfall on budgeted sales was estimated at GBP 2 mln).

When actualizing the balance sheet, Berkers and Derecks explain that currently an
extension to the factory is built. The costs exceed the budget by GBP 1 mln due to
additional "good manufacturing practice" regulations. The special bank loan of GBP 5
mln that was arranged originally, will not be sufficient to cover the expenditure. Total
liquidity shortage for the coming year will be GBP 1.5 mln + GBP 1.0 mln =
GBP 2.5 mln.

The key figures based on the reviewed budget represent the starting point and are
shown on one of the public screens, see exhibit 6-19.

summary key figures
MIS profit/loss: -- 771
liquidity surplus/shortage: -- 2,504
NHS excess/shortfall: -- 3,338
fiscal profit/loss: -- 2,351
(amounts in GBP * 1,000)

Exhibit 6-19. Summary key-figures screen (at start of the group session).

After the acceptation of this budget as starting point, all possible financing alternatives
are discussed. Terwolde presents the options he reviewed with analyst Nijboer and, in
a few cases, with Oostbeek.
- Internal selling-price increase to Taiwan. This seems feasible as Taiwan is doing well
as expected.
- Internal selling-price increase to Australia. This is not feasible as Australia has the
same difficulties with Retravir as the UK. Increasing the internal selling price would
lead to serious financing trouble due to a trading loss. Besides, the fiscal loss in
Australia would become unacceptably high.
- Transfer-price decrease from Oxford. This seems feasible although the idea is not
welcomed by Eurobitec. They suggest Europhar to solve its own problems, but with
the intervention of marketing director Gisbers this option seems feasible. The only
transfer price that can be discussed, is a price decrease of Biosulfine. As has been
explained earlier in this section, it is not always easy to get a transfer-price

120


decrease, because the divisions of the Eurochem company are profit centers and
each division is judged on its own profit.
- Transfer-price decrease from Amsterdam. Since the cost price for the generic articles
is relatively low compared to the transfer price, this is a feasible option. For other
articles, no possibilities exist in view of the reference prices to other European
countries.

Katwijk, O'Brian and Oostbeek have worked out two financing alternatives, together
with analyst Nijboer. Occasionally Oostbeek has been involved. These are the results:
- Cost transfers. This is not feasible since it would constitute a deviation from the
long-term fiscal policy. This would probably not be accepted by the Dutch fiscal
authorities.
- Royalty contract. As the royalty contract was changed last year in order to bring
royalty payments in line with sales of Europhar UK, a change in the opposite
direction would not be logical and probably not accepted by the Dutch fiscal
authorities.

Berkers, Derecks and analyst Nijboer have worked out the remaining financing
alternatives. Again Oostbeek has sometimes been involved. Results are as follows:
- Local borrowing. This looks feasible, although interest payments are a drawback as
they decrease liquidity. The absolute maximum is set on GBP 1 mln after discussing
the option with the Eurochem fiscal & financial department. This maximum is set
in view of the Eurochem contracts with banks in the UK. In cooperation with the
Eurochem fiscal & financial department, a bank is selected: the Northampton West
Bank. Interest rate for the loan (annual interest) would be 6.7% under an
Eurochem guarantee.
- Capital increase. This is feasible and there are no fiscal problems and no problems
with NHS regulations. The maximum as set by the Eurochem fiscal & financial
department is GBP 1 mln.
- Intracompany loan from Amsterdam. This is considered unfeasible after discussions
within the Eurochem fiscal & financial department due to arrangements made
between the Dutch fiscal authorities and Eurochem holding. These arrangements
cover financing of Eurochem's foreign subsidiaries.
- Current account Amsterdam. This is feasible, but 50 days payment term is the
maximum in view of the arm's length principle according to Oostbeek. At the
moment the payment term is 30 days, so increasing the payment term to 50 days
would result in:

20/365 * GBP 19.6 mln (being anticipated purchases during year) = GBP 1.1 mln

- Current account Oxford. Eurobitec's financial position is good. So the option seems
feasible, but Eurobitec is not very willing to help Europhar. "Let Europhar solve its
own problems". In view of NHS regulations, the maximal payment term would be
40 days (within the UK). At the moment the payment term is 30 days, so an
increase would result in:

10/365 * GBP 2 mln (being anticipated annual purchases) = GBP 55,000

121


Sbce the remittance is not that large, it is not worth trying to achieve a current-
account increase. Besides, it is certainly not worth the hassle, when London wants
a transfer-price decrease from Eurobitec.
- Current account Australia and Taiwan. Decrease of payment term is not possible as
the invoice would need to be paid before the goods would arrive, which is not
allowed by the custom authorities.

The facilitator summarizes the results and writes the remaining options on a flip chart.
These are:
- internal selling-price increase Taiwan,
- transfer-price decrease Oxford,
- transfer-price decrease Amsterdam,
- local borrowing,
- capital increase,
- current-account increase Amsterdam.

Task: decision making set financing, commercial and fiscal targets
Berkers puts forward that the financing target should be to avoid any liquidity
shortage. A surplus, however, is not necessary. Oostbeek puts forward that the fiscal
loss should preferably be less than about GBP 900,000. As far as the NHS is
concerned, O'Brian explains that the solution for the liquidity problem should not
interfere with the NHS calculations or alter the result. Negotiations with the NHS
authorities start in the autumn. O'Brian considers the current actualized figures for
Year 5 a realistic point of departure. Katwijk puts forward that a MIS profit target
does not need to be included as a separate target as decisions are made on the basis
of Europhar's consolidated performance. O'Brian agrees in principle, but if possible,
he would like to have a small MIS profit.

The facilitator summarizes the targets on the flip chart, see exhibit 6-20.

MIS profit/loss: small profit
liquidity surplus/shortage: zero
NHS excess/shortfall: shortfall of GBP 3.3 mln
fiscal profit/loss: loss of GBP 900,000

Exhibit 6-20. Summary of financing, commercial and fiscal targets.

Task: feasible-solution search select subset of alternatives and set total remittance
Katwijk proposes to split up the problem into two different iterations. First a "profit
and loss iteration" to generate extra cash flow of GBP 1.5 mln, and then a "balance-
sheet iteration" of GBP 1.0 mln. The participants and the facilitator agree with the
proposal. The profit and loss iteration considers three alternatives:
- transfer-price decrease Oxford,
- transfer-price decrease Amsterdam,
- internal selling-price increase Taiwan.

122


In this iteration the total remittance should equal GBP 1.5 mln. The balance-sheet
iteration also considers three alternatives:
- capital increase,
- current-account increase Amsterdam,
- local borrowing.
In this iteration the total remittance should equal GBP 1.0 mln.

The facilitator proposes that in the first day's meeting (April 28th) the profit and loss
iteration will be held. The balance-sheet iteration is planned for the second day (April
29th). In the remaining of the subsection, the profit and loss iteration will be
described.

Tasks: feasible-solution search select alternative and determine maximal remittance
at aggregate level & disaggregation
The three options of this iteration are discussed thoroughly. For each of these options
the tasks mentioned above take place.

- Transfer-price decrease Amsterdam. Terwolde explains the situation. From the
reference point of view none of these transfer prices can be decreased except for
generic products. The only possibility is to lower the prices of some products which
are also sold by generic houses. Taking their prices as a reference, a price decrease
of 20% would be possible.

The facilitator asks analyst Nijboer to display the landed costs per destination
(accounting module, component 1) on the public screen. On this screen it can be
seen that the landed cost for generics are GBP 1,515,000. So a decrease would
result in an extra gross margin of:

20% * GBP 1.515 mln = GBP 303,000.

- Transfer-price decrease Oxford. This alternative involves a highly complex issue, see
chapter 2 for more details on NHS regulations. The facilitator explains NHS
regulations at the request of Oostbeek as these regulations only apply to UK
pharmaceutical companies. The facilitator's explanation is supported by the NHS-
profit calculation of the accounting module, which is shown on one of the public
screens, see exhibit 6-21.

Under the current NHS regulations, royalties together with imported cost of goods
sold equal 48.3% of NHS sales. The NHS norm, however, is 47.3%. Therefore, the
NHS adds back 1% of the NHS sales to company profit. The destination-ratio
screen is shown on one of the public screens. The NHS destination ratio for
Biosulfine, the main product from Eurobitec, is 81%. If the transfer-price decrease
results in a transfer-price decrease for NHS sales of more than 1% of sales, there
is no add-back anymore as 47.3% is the NHS norm. Any additional transfer-price
decrease on NHS sales (above 1%), however, would further reduce NHS costs and
increase profitability of NHS sales. Eventually, this increased profitability could go
to a level which might result in a selling-price cut since the NHS authorities will

123


not automatically make a corresponding correction for the decreased profitability of
NHS sales. NHS sales are GBP 41 mln. 1% decrease is 1% * GBP 41 mln =
GBP 410,000. Total acceptable decrease from NHS point of view: GBP 410,000/81%
= GBP 507,000 (rounded off).

destination ratios for NHS
transfers total ratio
for NHS transfers
source article:
Biosulfine 823 1,022 811
Miconitraat 141 176 80X
Phosphoramidorine 267 311 861
(amounts in GBP * 1,000, source company Eurobitec UK Ltd)

Exhibit 6-21. Screen with destination ratios (profit and loss iteration).

Biosulfine transfers are GBP 1,022,000. So a decrease in transfer price of about
50% is required to obtain the GBP 507,000 which are needed. Berkers puts forward
that such a large decrease is not feasible, especially when the relationship with
Eurobitec is concerned. According to Katwijk, however, this 50% decrease in
transfer price is feasible and needed. The decrease is feasible because Eurobitec
recently introduced a new technology which causes a substantial decrease in the cost
price of Biosulfine. The decrease is needed because training of general practitioners
is more expensive than originally expected. Compensation for the extra marketing
cost made by Europhar should be granted by means of a lower transfer price,
according to Katwijk. As Berkers agrees, the facilitator concludes that the maximal
remittance of this option is GBP 507,000.

- Internal selling-price increase Taiwan. Terwolde explains that prices to Taiwan are
relatively low when compared with prices for other destinations since the Taiwan
prices have not been increased for a number of years. (This can be seen using the
disaggregation module.) An increase of 20% seems acceptable and Oostbeek agrees.
Oostbeek states that a price increase would still be at arm's length as the local
gross margin would remain very good. Moreover there are few other ways to
transfer money to Amsterdam. Berkers remarks that money is needed in Amsterdam
for the new research program. Therefore no money should be left idle in Taiwan.
As an overall price increase of 20% is acceptable, this will result in:

20% * GBP 3,963,000 = GBP 793,000.

Task: feasible-solution search summarize information on subset of alternatives
For the profit and loss iteration, the facilitator summarizes the situation on the flip
chart, see exhibit 6-22.

124


total remittance: GBP 1.5 mln
maximal remittances per alternative:
- transfer-price decrease Amsterdam: GBP 507,000
- transfer-price decrease Oxford: GBP 303,000
- internal selling-price increase Taiwan: GBP 793,000

Exhibit 6-22. Maximal remittances per alternative (profit and loss iteration).

Task: preference ordering select relevant criteria
This task is supported by the preference-ordering module. The facilitator asks analyst
Francen to show the checklist of criteria (preference-ordering module, component 2)
on his public screen. The following checklist is shown on one of the public screens:
- financial impact,
- flexibility,
- UK fiscal effects,
- other fiscal effects,
- NHS effects,
- reference effects.
The alternatives have different financial consequences, so the criterium "financial
impact" needs to be included in the checklist according to Katwijk. The "flexibility"
criterion can be left out of the checklist, because the three alternatives can be
considered inflexible according to Terwolde: after a transfer price or an internal selling
price is changed, it should in principle not be changed again in the near future (The
same argument applies for royalty contracts and cost transfers). Oostbeek thinks that
the criteria "UK fiscal effects" and "other fiscal effects" can be left out, because in this
case the alternatives do not differ as far as fiscal aspects are concerned. Derecks
remarks that this does not apply for the "NHS effects", which differ significantly for
the alternatives under consideration. Terwolde suggests to include "reference effects",
because the alternatives differ in view of this criterion.

"Speed & trouble" is added on request of financial expert Berkers, aiming at the
problems he expects with Eurobitec. So the following four criteria are selected:
- financial impact,
- NHS effects,
- reference effect,
- speed & trouble.

Task: preference ordering determine actual remittance per participant
The participants each receive a preference-ordering worksheet on which they fill in
their scores and weights for the three alternatives. The worksheets are then collected
by the facilitator and entered into the computer by the two analysts. Using the
preference-ordering module (component 2), a printout of the complete worksheet per
participant is made. On this printout, the normalized weight, preference order and
the actual remittance per alternative, as calculated by the computer, are presented. As
an example, the printout of the worksheet of Derecks is presented in exhibit 6-23.

125


transfer- transfer- internal
opinion weight normalized price price selling-price
weight decrease decrease increase
Amsterdam Oxford Taiwan

financial yes 9 271 3 2 4
impact
NHS yes 11 33* 3 2 2
effects
reference no - - - - -
effects
speed & yes 13 40Z 5 2 4
trouble
total 33 1002 3.8 2.0 3.3
score
preference 1 3 2
order
maximal
remittance 303 507 793
actual
remittance 303 404 793

(total remittance required 1,500, amounts in GBP * 1,000)

Exhibit 6-23. Example of a preference-ordering worksheet (printout).

The facilitator asks the participants if the outcome of the computer matches their "gut
feeling". All participants agree.

Task: preference ordering discuss differences and make group decision
The results as shown in exhibit 6-24 are now presented on the public screen.


internal
transfer transfer selling
total price price price
remittance decrease decrease increase
Amsterdam Oxford Taiwan

Derecks 1,500 303 404 793
Berkers 1,500 303 404 793
Katwijk 1,500 303 404 793
0'Brian 1,500 303 404 793
Oostbeek 1,500 303 404 793
Terwolde 1,500 200 507 793


Exhibit 6-24. Screen with actual remittances (profit and loss iteration).

126


As can be seen above, the figures of Terwolde deviate as a transfer-price decrease
from Amsterdam is his least preferred option and therefore not completely used. For
all other participants apparently a transfer-price decrease from Eurobitec is the least
preferred option. The facilitator puts forward that the difference in opinion is small
as far as the final outcome is concerned. The only difference within the group is
whether GBP 100,000 transfer-price decrease should come from Oxford or from
Amsterdam. This is only 7% of the total amount of GBP 1.5 mln that needs to be
transferred.

On proposal of the facilitator, the "normalized weight" table is shown on the public
screen, see exhibit 6-25.

normalized weight per participant
in percentages
Derecks Berkers Katwijk O'Brian Oostbeek Terwolde
financial impact 27Z 381 20* 23% 28X 211
NHS effects 33X 19% 23% 321 31% 24%
reference effect - 19% 23% - - 411
speed & trouble 40% 24% 341 45% 41% 14Z
100% 100% 1001 100% 100% 100*

Exhibit 6-25. Screen with normalized weights (profit and loss iteration).

Terwolde has a relatively high weight for "reference effect" and a relatively low weight
for "speed & trouble". On request of the facilitator, he gives an explanation on this
deviation. In 1992 he expects problems with prices in Europe. One open market makes
it impossible to have different prices to the European countries, according to Terwolde.
So his weight for the "reference effect" is higher. On the other hand he does not
expect too much problems with Eurobitec and so his weight for "speed & trouble" is
lower.

In spite of the difference in results among the participants, the "Amsterdam alternative"
is easily accepted by Terwolde on proposal of Katwijk.

Since just GBP 1.5 mln is needed, this implies that the least preferred alternative
(transfer-price decrease of Biosulfine from Eurobitec) is not fully used. The maximal
transfer-price decrease is 50% (507/1,022). The actual transfer-price decrease for
Biosulfine will now be 404/1,022 = 40%.

Task: decision making budget recalculation
During the night, analyst Nijboer enters the new prices in the disaggregation module
(component 3). Derecks and the facilitator are present. The new figures as calculated
by the disaggregation module are entered into the accounting module (component 1)
and a new set of key figures is produced.

127


SECOND DAY OF THE GROUP SESSION

29th April Year 5 is the second day of the group session. As on the first day, the six
participants, the two analysts and the facilitator are present.

Résumé of the profit and loss iteration
The facilitator gives a brief résumé of the meeting of the first day. In exhibit 6-26 the
result of the profit and loss iteration is presented.

summary key figures

MIS profit/loss: + 211
liquidity surplus/shortage:-- 1,012
fiscal profit/loss: -- 841

(amounts in GBP * 1000)

Exhibit 6-26. Summary key-figures screen (after profit and loss iteration).

The facilitator asks if the participants have questions or want to bring in additional
considerations. This is not the case. As can be seen, after this iteration the MIS and
fiscal target are reached and the NHS result is barely affected. So the only target that
still needs to be considered is the financing target. This will be done during the
balance-sheet iteration.

Task: feasible-solution search select subset of alternatives and set total remittance
The three alternatives which will be considered in the balance-sheet iteration are:
- capital increase,
- current-account increase Amsterdam,
- local borrowing.
The total remittance should equal GBP 1.0 mln.

Task: feasible-solution search select alternative and determine maximal remittance
at aggregate level
The three options of this iteration are discussed. For each option the task mentioned
above takes place.
- Capital increase. There are no problems to be expected with the capital increase
alternative. Oostbeek discussed this option with other members of the Eurochem
fiscal & financial department. It is possible to increase the capital by GBP 1.0
mln.
- Current-account increase Amsterdam. Using the accounting module (component 1) it
can be seen that the transfers, (against new transfer prices!) equal GBP 19.3 mln
annually. Increasing the current account from Amsterdam from 30 days to 50 days
would result in:

20/365 * GBP 19.3 mln = GBP 1.06 mln.

128


So the maximal remittance of this alternative equals GBP 1,060,000. At the
beginning of the meeting of the first day, this has been discussed.
- Local borrowing. On the first day of the meeting Berkers, Derecks and Oostbeek
already explained that the maximum for the local borrowing (the Northampton West
Bank loan) alternative is set on GBP 1.0 mln, because of the Eurochem contracts
with a few banks in the UK. The maximal remittance of this alternative equals
GBP 1.0 mln. In case the loan would be borrowed from the first of July onwards,
interest (before tax) on this loan (for Year 5) would be:

6.7% * 6/12 * GBP 1 mln = GBP 33,500

The amount of the loan should be decreased with this amount. So the effective
maximal remittance of the loan becomes:

GBP 1,000,000 -- GBP 33,500 = GBP 966,500

Task: feasible-solution search summarize information on subset of alternatives
For the balance-sheet iteration, the situation is summarized on the flip chart, see
exhibit 6-27.

total remittance required: GBP 1 mln
maximal remittances per alternative:
- capital increase: GBP 1,000,000
- current account increase Amsterdam: GBP 1,060,000
- bank loan: GBP 966,500

Exhibit 6-27. Maximal remittances per alternative (balance-sheet iteration).

Task: preference ordering select relevant criteria
Analyst Nijboer presents the checklist with criteria (preference-ordering module,
component 2) on one of the public screens. "NHS effects" is the only item of the
checklist which is taken into account. It is mainly taken because the current-account
alternative influences the capital employed as determined by the NHS, see chapter 2.
Other criteria of the checklist are not included, because the differences between the
alternatives are small.

The criterion "authorization" is added as the three alternatives require different
authorization levels. The criterion "long-term solution" is added, because a capital
increase has a different time perspective than the other two alternatives.

Task: preference ordering determine actual remittance per participant
The participants fill in their preference-ordering worksheets, which are then collected
by the facilitator and entered into the preference-ordering module (component 2) by
the analysts. On request of the facilitator the participants confirm that the outcome
matches their "gut feeling".

129


Task: preference ordering discuss differences and make group decision
The results of this preference ordering, as presented on the public screen, can be seen
in exhibit 6-28.

total capital current local
remittance increase account bank loan
Derecks 1,000 1,000 0 0
Berkers 1,000 0 1,000 0
Katwijk 1,000 0 1,000 0
0*Brian 1,000 1,000 0 0
Oostbeek 1,000 0 1,000 0
Terwolde 1,000 0 1,000 0


Exhibit 6-28. Screen with actual remittances (balance-sheet iteration).
The screen shows that the current account alternative is preferred by most participants.
However the outcome for O'Brian and Derecks deviates. On request of the facilitator,
the normalized weights are shown on the public screen, see exhibit 6-29.

normalized weight per participant
in percentages
Derecks Berkers Katwijk O'Brian Oostbeek Terwolde
NHS effects 33X 24% 33X 30Z 24% 281
authorization - 35X 25Z - 40X 291
long-term solution 67% 41Z 42* 701 36Z 43Z
100Z 100X 100% 100% 100% 100%

Exhibit 6-29. Screen with normalized weights (balance-sheet iteration).

O'Brian explains the deviation. He says they have no opinion concerning "authorization",
as this is a divisional head office activity. Furthermore they have a higher weight for
"long-term solution", as in the UK situation, too often problems have been solved in
a temporary way.

Katwijk points out that he understands the local desire for a long-term solution, but
in view of the uncertainty with regard to developments in the market he prefers to
stick to the current-account alternative because it is more flexible. O'Brian and Derecks
agree to this.

Task: decision making budget recalculation
Total remittance needed is GBP 1 mln. An extension of the payment term by 19 days
is sufficient to obtain this as:

19/365 * GBP 19.3 mln = GBP 1 mln.

130


The facUitator asks analyst Francen to calculate the key figures after an increase of the
Amsterdam payment term from 30 to 49 days. For the last time the summary key
figures are presented, see exhibit 6-30.

summary key figures
MIS profit/loss: + 208
liquidity surplus/shortage: + 63
fiscal profit/loss: -- 843


Exhibit 6-30. Summary key-figures screen (after balance-sheet iteration).

Task: problem solving formal agreement by authorizer
The facUitator writes the action steps which need to be authorized on the flip chart:
- Transfer-price decrease for genetics from Amsterdam. The first action step will be
authorized by Katwijk. This concerns the transfer-price decrease for generics from
Amsterdam by about 20%. Exhibit 6-31 shows the old and the new prices for the
generics in NLG as calculated on the night of the first day.

articles: old new
Nimatapine 21.— 17.--
Nalvodexine 82.— 66.—
Cotergenine 990.— 790.—
Olupentine 27.— 21.—

Exhibit 6-31. Old and new transfer prices for generic products.

- Internal selling-price increase to Taiwan. All the destination articles, which are sold
to Taiwan, will be increased by 20%. Exhibit 6-32 shows the old and the new
prices, for these articles in GBP (as calculated on the night of day 1).

articles: old new
Briconcyl 29.25 35.10
Ogmentin 74.75 89.70
Sastril 10 18.31 21.97
Sastril 50 91.40 109.68
Darmicum 9.54 11.45
Opanol 3.53 4.24
Morvelan 14.34 17.21

Exhibit 6-32. Old and new internal selling prices to Taiwan.

Katwijk wUl approach the Far East regional manager to obtain formal authorization
for the internal selling-price increase to Taiwan. This has been discussed with the
people involved before the group session.
- Transfer-price decrease for Biosulfine: The third action step is the transfer-price
decrease of 40% for the Eurobitec article Biosulfine. The current price is GBP 6.50

131


per unit. So the new price will be GBP 3.90. In cooperation with marketing director
Gisbers, Katwijk will try to get this authorized by the financial director of Eurobitec.
Related to this issue, Katwijk proposes another action. He thinks relationships
between Oxford and Europhar UK in general should be improved. He will bring
this up at the meeting with the financial director of Eurobitec. He will also discuss
this with marketing director Gisbers.
- Increase of the payment term from Amsterdam. The following action step is increasing
the current account from Amsterdam from 30 to 49 days. On request of Katwijk
Berkers will submit this action for authorization by Marcus, the financial director
of Europhar. Berkers says he does not expect any trouble.

Katwijk proposes a last action step: a check for the currency-exchange rate and interest
stability. The two financial specialists (Berkers and Derecks) do not consider this
relevant as the interest rate and die exchange rate are very stable. They consider this
a waste of time.

The facilitator asks if there are any questions or comments amongst the participants.
Since there are not, he asks Katwijk to close the meeting. Katwijk thanks the
participants, analysts and the facilitator for their contribution. He stresses that he
enjoyed the open, informal and constructive attitude of all involved and the meeting
is closed.

THE EUROBITEC NEGOTIATIONS

Above, the decision process and system use are described in some detail for the month
of April Year 5. Use of system elements during the period May-September Year 5 is
now globally described below. Please note that the use of the system in this period
differs from its original purpose: to determine a financing policy. In the part described
below, the system is used to support negotiations with Eurobitec. This subsection shows
the close link between decision support and negotiation support.

All action steps decided upon during the group session are authorized and
implemented in the beginning of May Year 5. At the end of May Year 5, Eurobitec
in Oxford is faced with an overcapacity problem in the production area. This is due
to the recent start of the business in Oxford. In this initial phase the plant is not yet
working at full capacity. In order to solve this problem the management of Eurobitec
proposes to relocate tabletting and packaging activities of the product Sofanaat to
Oxford.

At this moment the raw material Biosulfine is produced in Oxford. Biosulfine is then
transported to Europhar in London, where it is used to produce Sofanaat tablets. Then
these tablets Sofanaat are packed by Europhar. The original situation is shown in
exhibit 6-33.

132


raw
materials
X.Y.Z

Eurobitec
f
raw
material
Biosulfine

Europhar
f
semi-finished
Sofanaat
(tablets)

Europhar
T
finished
Sofanaat
(packed)

Exhibit 6-33. The production process for Sofanaat.

Management of Eurobitec wishes to relocate tabletting and packaging activities from
London to Oxford. However, Europhar in London would then get overcapacity
problems. A shift of these activities would seriously affect Europhar's business.
Europhar people have the impression that Oxford is trying to roll off its problems on
London.

As explained, the Europhar divisions are independent profit centers, see chapter 3.
Bargaining is the predominant character of the meetings between the two companies.
In the negotiations that take place with regard to this relocation of tabletting and
packaging, a transfer price for the packed Sofanaat needs to be determined. Until now,
Biosulfine and not Sofanaat has been sold to Europhar.

On behalf of Europhar, Katwijk negotiates with Eurobitec. As Europhar and Eurobitec
belong to the same fiscal unit, fiscal effects of the relocation for the unit as a whole
are nil. For the problems caused in the financing and commercial area, Europhar
requires adequate compensation. From the point of view of Europhar, the aims of
these negotiations are:
- to minimize the degree of tabletting and production activities that are relocated to
Eurobitec,
- to obtain a low transfer price for the packed Sofanaat that replaces the Biosulfine.

Katwijk considers arranging a second group session in order to determine a negotiation
policy towards Eurobitec. After discussing this with Marcus, the Europhar financial
director, they decide not to do so. Since the overall policy, which is already
determined at the end of April is clear, a new group session is not necessary.

133


To determine the negotiation policy, various meetings are held. In evaluating the
various negotiation strategies, Katwijk, Derecks, Terwolde and Berkers are involved.
Support is provided by analyst Nijboer. Analyst Francen, the facilitator, O'Brian and
Oostbeek are not involved. In various stages of the negotiation process, Terwolde
works out the consequences of the relocation and transfer-price alternatives and
together with analyst Nijboer the financial, commercial and fiscal key figures for the
various alternatives are calculated. This is done by using the accounting module
(component 1) and the disaggregation module (component 3).

As Katwijk is able to present hard figures on the commercial and financing
consequences of this unexpected move from Eurobitec, he is able to make a strong
case. After a series of negotiation talks a compromise with Eurobitec is reached in
September of Year 5. The main point of this compromise is that only the tabletting
activities are relocated to Eurobitec. The packaging activities remain with Europhar.
Another point of the compromise is that the transfer price for Sofanaat tablets is
relatively low. This low transfer price is an adequate compensation for the negative
commercial and financing effects.

task. / activity component / module
3
a
1 2
f r
1
g P
c u r i e
c e n f
o 1 f i e
n a o m r
t
preparations
actualize and review X X X X X

first day of the group session
résumé of the actualize and review task X
determine maximal remit, at aggregate level & disaggregation X X
select relevant criteria X
determine actual remittance per participant X
discuss differences and make group decision X
budget recalculation X X

second day of the group session
résumé of the profit and loss iteration X
determine maximal remittance at aggregate level X
select relevant criteria X
determine actual remittance per participant X
discuss differences and make group decision X
budget recalculation X
the Eurobitec negotiations
negotiation support X

legend:
1,2,3: component 1, 2 or 3
accon: accounting module
gulat: regulation database
prefo: preference-ordering module
finim: financial-impact module
disag: disaggregation module
refer: reference database

Exhibit 6-34. Use of the various modules during the experimentation period.

134


SUMMARY OF COMPUTERIZED-SUPPORT USE

In exhibit 6-34 the use of the various modules during the experimentation period is
summarized.

6.5 Conclusion
The previous section gives us a description of six months experimental use of the
system.

The system is presented to Europhar at the end of March in Year 5. When
performing "real-life" tests, the researcher has no full control over the experiment.
From the point of view of the experiment, we are pleased to observe that a problem
came up in the beginning of April Year 5. This problem gave us the possibility to test
the system in a "real-life" situation.

We would like to put forward that system use during this period was still experimental
as, for example, the facilitator is still somebody external to Europhar and the decision
support system was not yet integrated in the MIS.

A last point to note is that before and during the group session, the participants used
the system for decision support and stuck to the decision-process structure. After the
group session, mainly the computerized support was used for negotiation support. As
explained in chapter 4, negotiation support is not the aim of the system.

With the prescriptive empirical model in experimental use for six months, we have a
basis to evaluate the system.

135


136

7

EVALUATION
7.1 Introduction
The second research question, as formulated in chapter 1 (B'), is the following:

What are the effects of the use of the prescriptive empirical model on organizational
decision making?

As explained in chapter 1, the prescriptive empirical model should be regarded as an
experimental system and not as a full system. The aim of the evaluation of the
empirical model is to establish the value of the proposed concept: the prescriptive
conceptual model. In this chapter we evaluate the prescriptive empirical model.

We realize, however, that "evaluation is perhaps the most difficult aspect of decision
support systems development especially since such systems are evolutionary in nature",
see Keen and Scott Morton [1978, p. 213]. Choosing when to evaluate them is as
difficult as choosing how.

The decision to evaluate the system in September of Year 5 is mainly based upon the
consideration that in September the attention of the participants shifted towards
preparing the budgets for Year 6. Although various plans existed to add new features
to the system, these would not be carried out until after the budget-preparation period
due to time constraints of organizational members involved.

Turning to the question how to evaluate the system, Keen and Scott Morton [1978]
present a multitude of evaluation approaches to be matched to a specific situation.
Each of these approaches can contribute to insight in the performance of the
prescriptive empirical model and hence in the feasibility of the prescriptive conceptual
model.

As Keen and Scott Morton only consider approaches to evaluate systems already in
use, however, one indicator of system performance is omitted: system use itself. We
consider system use during the experimentation period as described in the previous
chapter our "hardest" and best indicator of system quality: during six months, the
system has been used to solve "real-life" problems, see chapter 6. In this period, system
use has been supported by the participants and top management. As rather large sums
of money were involved in decision making, this expresses confidence in the system.

Following Keen and Scott Morton [1978] the following approaches have been
considered for evaluation of the experimental system:

137

EVALUATION Ch. 7

Output of the system
The ideal performance measure would be to determine the impact of the system on
financing, commercial and fiscal company goals. The exact determination of system
impact, however, is impossible, see exhibit 7-1. In the first place, the effect is disturbed
as most changes in the environment are reflected in the company goals. Another aspect
is that the figures that result from financing decisions are subject to negotiations with
the various authorities. Thirdly, it should be realized that the time lag between
financing decisions and company goals is about three or four years. As a surrogate
measure, we will use the impact on company goals as expected by the participants.
financing
decisions

financing, fiscal
- ► and commercial
effects

time l a g

interpretation
disturbing factors -► and
negotiation

company
goals

Exhibit 7-1. The indirect effect between financing decisions and company goals.

Changes in the decision process
As decision-process structure is one of the main elements of the system, it will be
clear that we compare the decision process before and after introduction of the system.
To gain insight in the process structure before and after introduction of the system,
meetings have been recorded on video and audio tape and computer use has been
logged by means of the specially designed logging facility. The previous chapter shows
that the decision process after system implementation neatly matches the process
structure as designed in the prescriptive conceptual model. Time spent on decision
making and number of alternatives are also analyzed. For a comparison of the decision
situations as described in the chapters 5 and 6, see exhibit 7-2.

138


relevant actions with frequency
information available but not shared with other participants
considerable loss of time due to search for information
number of times that participants are missing
need for properly actualized budget
information on alternatives lacking
number of operational or relocation alternatives evaluated
number of financing alternatives evaluated (at aggregate level)
general complaints about the complexity of the issue
disaggregation without systematic selection of articles
quantitative insight in consequences of actions impossible when needed
sensitivity analyses with regard to external parameters impossible when needed
no proper introduction or résumé when relevant
change of subject of the meeting without proper conclusion
manual calculation error
choice task mixed with integrative-design task

legend:
ai frequency of actions during the meeting of July Year 4
b: frequency of actions during the group session

Exhibit 7-2. Comparison of frequency distributions of actions taken during two
meetings.

Manager's assessment of the system's value
This method is the one on which our evaluation is mainly based, see the next section.
Keen and Scott Morton [1978, p. 223] describe this evaluation method as "effective" as
decision makers "frequently have a very shrewd sense of what a resource or service is
'worth'". This is not to say that we would not have preferred a harder, more objective
way of performance measurement. According to Ives and Olson [1983], however, "user
information satisfaction provides a meaningful 'surrogate' for the critical but
unmeasurable result of an information system, namely, changes in organizational
effectiveness".

Anecdotal evidence
Keen and Scott Morton consider it useful that a more formal evaluation is
supplemented by "anecdotal" evidence such as insights, examples, opinions and events
collected by a skilled observer. In our case this anecdotal evidence, like letters and
internal memos, supplements the evaluation interviews. Quotes taken from the anecdotal
evidence are included in the overall evaluation as presented in section 7.3. An example
is presented in exhibit 7-3. Note that this letter is adapted to the Europhar situation,
see chapter 5.

139

EVALUATION Ch. 7

Mr N. Katwijk
Manager European Region
Europhar International BV
Amsterdam
The Netherlands

London, June 2nd, Year S.

Dear Mr Katwijk,

As decided earlier this week, the evaluation of the TP-DSS system is
scheduled for our Europhar UK board meeting on June 14th. On your request,
X send you our view on its value in advance.
Derecks and I have discussed both the group session on the 28th/29th April
in Amsterdam and the work and effort that went into the preparation of that
meeting.
Preparation
As is well appreciated, the complexity of the difficulties faced by Europhar
UK Ltd with the interrelationship of the NHS and the fiscal authorities has
been a problem looking for a solution for a number of years. In addition,
our financing problems, one way or other, are chronic. Therefore, it was
with enthusiasm that we viewed the system devised by Delft University.

We found the whole system well researched and considered that all the
relevant factors with their interrelationships and effects had been
carefully assessed and included. Personal Wizard proved to be an adequate
piece of software. Pre-reading and instruction were also thought to be of
good quality.

The group session
Due to the high quality of the preparation for this meeting and the positive
attitudes that were taken by the participants (team setting), the meeting
turned out to be extremely valuable. It was well controlled by the
facilitator without his deliberately trying to influence the participants'
views. The group contained sufficient expertise for the task and, as has
already been stated, was well motivated. There were good practical
facilities and the software modules worked.
By the end of the second day it was clear that the value of the system was
well appreciated, particularly as it was possible in two days to reach clear
decisions with a positive effect on all three aspects (financing, commercial
and fiscal) of our complex problem. This is particularly rewarding as these
difficulties have been constant for a number of years and it appears that
the reason that they have not been solved previously was our inability to
relate adequately the cross-effects of a number of interrelated decisions
reached in isolation.

In summary, both Derecks and 1 found the meeting stimulating and the system
practical and valuable.
Yours Sincerely,

Europhar UK Ltd

J. O'Brian,
General Manager

Copy : Mr R. Marcus, Financial Director
Mr V. Gisbers, Marketing Director

Exhibit 7-3. A letter from the UK general manager.

140

Sec. 7.2 QUESTIONNAIRE AND ANSWERS

7.2 Questionnaire and answers
Following Keen and Scott Morton's approach of the manager's assessment, see above,
evaluation interviews played a major role in determining system performance. In
September Year 5, questionnaires were sent to all participants with the request to
think about the questions in advance. The interviews took place during October
Year 5. It is important to realize that, as described in the previous chapter, the
participants used the system together before the interviews took place. Exchange of
information and opinion with regard of the system took place on a regular basis. It
is likely that parts of the questionnaire were discussed amongst participants in
advance. This implies that the answers given are not independent of each other. We
used individual interviews to elicit as much comment as possible. Another reason for
individual interviews was that it was much more practical than a group interview,
especially as two participants live in the UK.

We did choose for evaluation by means of structured interviews for the following two
reasons:
- As all participants had to fill in the same questionnaire during the interview, it
would be possible to compare and aggregate the answers. The questionnaire had
the advantage that no important aspects would be forgotten.
- In case not all issues were exactly covered by the questionnaire, the interviewer
would have the possibility to elicit comments.

The interviews took on average one hour and a half. After the questionnaires were
filled in, they were signed for agreement. Besides the interviews with participants, two
short interviews took place with top managers (Gisbers and Marcus) involved. During
these interviews a few questions from the participants questionnaire were presented.
We would like to stress, however, that perception by top management of the quality
of the system was completely indirect: none of the two top managers had any direct
experience with the system. Their point of view was fully based upon the signals that
they received from the organizational members that were involved. These two
interviews took 15 minutes each. For a detailed description of the set up and outcome
of the interviews, see Van der Ven [1989a].

Answers to the questions are relevant. Perhaps even of greater relevance, however, are
the comments provided. In combination with the answers given, it is our impression
that the comments give a profound insight in the points of view of the organizational
members involved. In exhibit 7-4 a scheme with the answers of the participants is
provided. Answers concern the impact on three categories of variables: company goals,
confidence and insight, and the decision process.

141

EVALUATION Ch. 7

question * answer frequency
A: impact on company goals ni mi si hsi -
1 impact on financing company goals - 2 4 -
2 impact on commercial company goals - 1 5 -
3 impact on fiscal company goals - 3 2 1
4 impact on financing, commercial and fiscal goals - 1 2 3
A * n It n n n it n
* - - 2 -
B: impact on confidence and insight ni mi si hsi -
1 general problem understanding of transfer pricing 1 5 - -
I It it it it it n tt

2 general problem understanding of the UK situation
- 1 1
* 2 2 2 -
2 * tt n n tt it it it
-
3 insight in the reasons behind a specific decision
* - - 2 - -
1 4 1
4 insight in effects of financing alternatives
-
1 5 - 2
5 insight in effects of external parameters - - 2 2
6 insight in regulations - 4 1 1
7 insight of one's own point of view 1 2 3 -
8 insight of point of view of other participants - 1 5 -
9 insight in financial impact (cost) - - 4 2
10 insight in disaggregation - - 5 1
11 insight in reference prices 1 3 - 2
12 confidence in the decisions made - - 5 1
1 2 * n u n it it
* - - 2 -
13 acceptance of the decision within the organization
i » n it it it it it n
- 2 1 2 1
* - - 1 1
C: impact on the decision process ni mi si hsi -
1 number of financing alternatives evaluated 6 -
- - main text
2 time spent on decision-making (notes deterioration) see
3 effectiveness of time spent in decision making - 1 4 1
4 effectiveness of communication with participants - 1 3 2 -
5 effectiveness of communication with non participants 1 2 2 1 -

legend:
* : answers from the two top managers
ni : no impact
mi : modest improvement
si : significant improvement
hsi : highly significant improvement
: no opinion

Exhibit 7-4. Frequency distribution of answers on impact of system use.

The seven-point scale that was used in the questionnaire ranged from "highly
significant deterioration" to "highly significant improvement". The answers on question
C-2 (time spent on decision-making) are the only ones that indicate a deterioration.
The frequencies of the answers obtained for question C-2 are:

-: highly significant deterioration,
1: significant deterioration,
4: modest deterioration,
1: no impact,
-: modest improvement,
-: significant improvement,
-: highly significant improvement,
-: no opinion.

Exhibit 7-4 shows that the impact is positive in most cases. In a few cases there is no
impact but, except for time spent (question C-2), the impact is never negative.

142


Exhibit 7-5 presents the contribution of the various system elements to system
performance.

question answer frequency
D: contribution of the system elements ni spc pc lpc _
1 contribution of the authorizer role 2 2 1 1 -
2 contribution of the analyst role 1 - 2 3 -
3 contribution of the facilitator role - - 1 5 -
A contribution of the three roles - 1 3 2 -
5 contribution of system use outside the group session - - A - 2
6 contribution of system use during the group session - - 2 A -
7 contribution of actualize and review task - - 3 2 1
8 contribution of feasible-solution search task - - 5 1 -
9 contribution of preference-ordering task - 2 3 1 -
10 contribution of dlsaggregatlon task - - A 2 -
11 contribution of the entire process structure - - 2 A -
12 contribution of the accounting module - - 3 1 2
13 contribution of the regulation database 2 - 1 - 3
1A contribution of the preference-ordering module - 2 1 1 2
15 contribution of Che financial-impact module _ - A - 2
16 contribution of Che dlsaggregatlon module - - 2 2 2
17 contribution of the reference database - 1 3 A - 2
18 contribution of Che six software modules _ . 2 -
19 contribution of the problem specification - - 2 A -
legend:
ni : no impact
spc : small positive contribucion
pc : positive contribution
lpc : large positive contribution
: no opinion

Exhibit 7-5. Frequency distribution of answers on contribution of system elements.

Exhibit 7-5 shows that the contribution is positive in most cases. In some cases there
is no impact but the contribution is never negative. We considered it important to
search for a clear interpretation of the results. The null hypotheses that we consider
are:
section A: the system does not cause any improvement with regard to the
company goals,
section B: the system. does not have any positive impact on problem
understanding, insight or confidence,
question C-2: the system does not have any negative impact on time spent,
section C (other): the system does not have any positive impact on the decision
process,
section D: system elements do not contribute positively to system
performance.

We use the following decision rule for rejection of the null hypotheses:
- all six participants answered the question, so there are no "no opinions", and
- all six answers are positive (except question C-2). This means that all participants
marked an "improvement" in the sections A, B or C (except question C-2) or a
"positive contribution" in section D. For question C-2 (time spent) all answers
should be negative.

143

EVALUATION Ch. 7

The choice for this decision rule needs some explanation. From a statistical point of
view, the respondents are a sample. The central question is, what is the population (or
universe) to which this sample relates.

With regard to Europhar, all organizational members involved were interviewed. In
other words, we interviewed the complete population, see Newbold [1988, p. 8]. So
sample and population are identical. By presenting all answers, like we did above, all
information that exists is already conveyed to the reader. We only need to summarize
the results. This is the domain of descriptive statistics, see for instance Harnett [1982,
p. 2].

It is also possible to consider the respondents a sample of the population of all
decision-makers who might participate in a similar experiment. "Similar" means using
a similar system in similar circumstances. Both the system and the circumstances must
be similar, to provide that the answers have the same probability distribution. We
include "similar systems" rather than identical systems, because different companies
constitute different environments to the experiment and need different - adapted -
versions of the prescriptive conceptual model. The answers of our sample would then
be used to estimate the opinion of the entire population with regard to the effect of
the use of a similar system in a similar company. This is the domain of inductive
statistics, see Harnett [1982, p. 2 and 250] and Newbold [1988, p. 231].

AJI obstacle of using statistical inference is that the sample is not taken at random as
all respondents are drawn from the same experiment. By consequence, the answers of
different respondents are not independent. Organizational members are already working
close together for a number of years. System development and system use are
explicitly a group effort and the impact of the system is extensively discussed amongst
participants and with top-management. The evaluation questionnaires were sent in
advance and it is likely that these are also discussed amongst organizational members
before the time of the interviews. So in fact we have drawn one respondent only: the
group of organizational members involved. That the group opinion is determined by
means of individual interviews, does not change this fundamental issue.

Given this point, however, we still want to aggregate the answers given by the six
participants to a rejection or non rejection of the null hypotheses. Presuming mutual
independence, we might use a statistical test to select a decision rule. Note that the
selection of this decision rule is not based upon formal statistical inference. Statistics
is only used to provide guidance.

As the results are measured on an ordinal scale, a non-parametric test is used, see
Harnett [1982, pp. 695-696]. We have a sample of matched pairs (the before and after
introduction situation), so the matched-pairs sign test is appropriate.

Below we present the way of calculation for all questions except question C-2 (time
spent). The approach to question C-2 is similar. (We only have to interchange the
roles of positive and negative.) We compare the "after introduction" situation with the
"before introduction" situation. The before introduction situation is by definition "no

144


impact". We apply a one-sided test. If we require 2.5% significance, we get the table
that is presented in exhibit 7-6, see Conover [1980, pp. 434-435].

n - Is P(X-l) - 50.00X (> 2.5X)
n - 2: P(X-2> - 25.00* (> 2.51)
n - 3: P(X-3) - 12.501 (> 2.5Ï)
n - A: P(X-4) - 6.251 (> 2.51)
n - 5: P(X-5) - 3.12Z (> 2.5X)
n - 6: P(X-6> - 1.56Ï (< 2.51)
n - 6: P(Xa5) - 10.94* (> 2.51)
legend:
n: number of respondents with a positive or negative opinion.
X: number of respondents with a positive opinion.

Exhibit 7-6. Table with probabilities.

For example, "n = 6: P(X = 6) = 1.56%" means: Given the null hypothesis and six
respondents with a positive or negative opinion (n = 6), the probability that six
respondents give a positive answer (X = 6) is 1.56%. The figures of this table can be
translated into the decision rule that we presented above.

With the exception of question C-2 on time spent, there are no negative answers.
Thus we can make a distinction between non-rejection of the null hypothesis caused
by lack of observations, because of "no-opinion" answers and non rejection because of
"no-impact" or "no-contribution" answers. We would like to stress, that in a statistical
sense, this does not make any difference. That we decided to make a distinction in
the presentation of the results, is because we consider it relevant for the
interpretation. The following two examples might illustrate this issue:

- According to four respondents, the disaggregation module (question D-16) had a
"large" (answered twice) or a "very large" (also answered twice) contribution to
system performance. As two respondents were not directly involved in software use,
they indicated "no opinion".

- With regard to the impact of system use on problem understanding (question B-
2), the respondents answered as follows: Two times "no impact", two times "modest
improvement" and two times "significant improvement".

From a statistical point of view, both answers are identical in the sense that four
positive answers are given, which means that the null hypothesis cannot be rejected.
In the case of the disaggregation module (question D-16), however, all participants
involved are positive. The only problem is that we lack observations. The case of
problem understanding (question B-2) is different: two respondents deliberately answer
that their problem understanding did not increase. Therefore, non rejection of the null
hypothesis caused by one or more "no-impact" or "no-contribution" answers (in some
cases combined with "no opinion" answers) is interpreted as no impact or no
contribution. Rejections caused by one or more "no-opinion" answers only, are
interpreted as lack of observations. The classification shows that we require a lot

145

EVALUATION Ch. 7

before we grant the classification positive impact or positive contribution as for this
classification we need a positive answer from all six participants. Note that we use
the term significant in another sense than in the questionnaire.

question ov jd
A: impact on company goals
1 impact on financing company goals
2 impact on commercial company goals
3 impact on fiscal company goals
4 impact on financing, commercial and fiscal goals
B: impact on confidence and insight
1 general problem understanding of transfer pricing
2 general problem understanding of the UK situation
3 insight in the reasons behind a specific decision
4 insight in effects of financing alternatives
5 insight in effects of external parameters
6 insight in regulations
7 insight of one's own point of view
8 insight of point of view of other participants
9 insight in the financial impact (cost)
10 insight in disaggregation
11 insight in reference prices
12 confidence in the decisions made
13 acceptance of the decision within the organization
C: impact on the decision process
1 number of financing alternatives evaluated
2 time spent on decision making (note: deterioration)
3 effectiveness of time spent in decision making
A effectiveness of communication with participants
5 effectiveness of communication with non participants

legend:
ov jd! overall judgement
* : a non-significant deterioration for time spent
p : positive impact
n : no impact
: lack of observations

Exhibit 7-7. Overall judgement on impact of system use.

On the basis of the information provided in exhibit 7-5, the five system elements with
the highest scores are:
- problem specification,
- decision-process structure,
- facilitator role,
- group session,
- six software modules.

The answers on these five questions dominate the answers on the other questions. For
an explanation of the meaning of dominance, see Albers [1987, pp. 93-99]. For a
detailed account of all results that were obtained, we refer to Van der Ven [1989a].

146


question ov jd
Ds contribution of the system elements
1 contribution of the authorizer role n
2 contribution of the analyst role n
3 contribution of the facilitator role * P
4 contribution of the three roles P
5 contribution of system use outside the group session
6 contribution of system use during the group session * P
7 contribution of the actualize and review task
8 contribution of feasible-solution search task p
9 contribution of preference ordering task P
10 contribution of disaggregation task P
11 contribution of the entire process structure * P
12 contribution of the accounting module
13 contribution of the regulation database n
14 contribution of the preference-ordering module -
15 contribution of the financial-impact module -
16 contribution of the disaggregation module -
17 contribution of the reference database -
18 contribution of the six software modules * P
19 contribution of the problem specification * P

legend:
ov jd: overall judgement
* I top-five score
p : positive contribution
n : no impact
: lack of observations

Exhibit 7-8. Overall judgement on contribution of system elements.

7.3 Overall evaluation
In this section the results as shown above, the comments given during the evaluation
interviews, the analysis of the decision processes as described in chapter 5 and 6 and
the anecdotal evidence are integrated into an overall evaluation.

Company goals
Most participants considered the impact on the integration of all three aspects to be
"significant" or "highly significant". A comment made was that "the sum of total
integration was more important than the component elements". Top management
indicated that they expected that the system caused a "significant improvement" of the
attainment of company goals. The comments stressed the capacity of the system to
integrate the three main aspects of financing decisions: financing, commercial and
fiscal.

General problem understanding
It is interesting to note that general problem understanding is not significantly
increased, according to the participants. Background of these answers could be that all
knowledge concerning the various aspects was already available within the company.
At the time of the evaluation in Year 5, the participants and top management were
already working closely together for nearly five years in order to solve transfer-pricing
and financing problems. All aspects related to the problem were already extensively
studied. This is perhaps best expressed by one of the comments made: "I considered

147

EVALUATION Ch. 7

myself already an expert" and: "ample expertise on the financing, commercial and fiscal
aspects ... was [already] available". The two top managers, however, both had the
impression that problem understanding within the organization had increased. This in
contrast to the participants.

Insight in reasons behind a specific decision
The "why" behind a decision is better understood. When the decision processes in
chapters 5 and 6 are compared, this will be clear. In the comments, different system
aspects are indicated as the cause for this increased insight. The group session, the
software modules, the process structure and the integration of financing, commercial
and fiscal aspects are named as causes for the increased insight.

Insight in the effects of financing decisions and external parameters on company
goals / accounting module
This module is seen as an "excellent 'what-if tool" which helps to communicate the
impact of the various alternatives to the other participants. The insight in the effects
of financing decisions is increased significantly. As far as the external parameters are
concerned, the Eurobitec relocation operation is named a few times. As not all
participants were involved in the Eurobitec negotiations the improvement of the insight
was not significant. The frequent use of the accounting module both inside as well as
outside the group session, however, can be regarded as an important indicator of its
performance. Comments made stress both the usefulness of this module inside as well
as outside the group session.

Insight in regulations / regulation database
Regulations are of overriding importance. The task descriptions in the prescriptive
conceptual model, for example, indicates regulations as input for nearly every task.
According to some participants, the system contributed to better insight in regulations
due to a more effective communication with the experts within the group. Although
the "structured approach [of the regulation database] provided a fresh view", the
regulation database was seldom used and is the only module which got a "no impact"
evaluation. We consider the cause of this to be twofold. Firstly, the set up of the
regulation database already provided this "fresh view". The participants already knew
most of its contents (the regulations), because they were involved in setting up the
module. Therefore, no further need to consult the database existed. Secondly, it is our
impression that the on-line storage of information in text mode, at least for this
purpose, is not suitable. The approach lacked flexibility.

Insight in one's own and other's point of view / preference-ordering module
As the outcome indicates, the strength of the preference-ordering module was
especially that it provided insight in the point of view of others. It was considered a
"very good aid for mtra-group communication" as "others were 'forced' to express and
explain their point of view". Own preferences ("gut feeling") of participants were
already well understood. With regard to this aspect, some additional points were made.
Firstly, that "informality of the meeting was of great importance" as "together with the
structured approach it stimulated explicit disclosure". Secondly, the multi-criteria
technique was considered to be "complex". Additional training was requested as

148

Sec. 7.3 OVERALL EVALUATION

participants wanted to "play around with this module" for better understanding. A last
point that was made was that the tool should not be used as a "voting instrument",
but only to build "consensus".

Insight in the financial impact / financial-impact module
Although not all participants were involved in the use of this module, all answers
given were positive. Especially "the reasoning behind it", as discussed in chapter 3, was
considered useful. The module "revealed previously unknown consolidated impacts" and
the outcome was sometimes considered "surprising".

Insight in disaggregation / disaggregation module
Of all modules, the disaggregation module obtained the highest scores. These scores
dominate the scores for the other modules. Insight in the disaggregation effect scored
also very high. Background of these scores is the introduction of the NHS-destination
ratio. This new concept is a "very effective tool" as it "reduces a complicated activity
to a single sheet of paper". The destination-ratio "was frequently used" during the
experimentation period. It is realized that "without computer support the calculation
of these effects would have been impossible". The need for the combined use of the
disaggregation and accounting module was put forward.

Insight in reference prices / reference database
The contribution of this module was not highly valued. This is not because the
information provided was not considered useful. The reason is that "reference prices
were already systematically calculated" and therefore the database "provided no new
information", although the reference database provided some extra facilities which were
appreciated. In other words, a reference database already existed within Europhar.

Confidence in the decisions made
Both participants and top management indicated a significant impact of system use on
their confidence. The two elements that caused this higher confidence were: integration
of all aspects (financing, commercial and fiscal) and the structured approach to
decision making.

The degree of acceptance within the organization
Although not all respondents answered this question, especially the comments indicate
that the confidence within the organization increased. In relation to the degree of
acceptance two aspects were relevant: the negotiations with Eurobitec and the
authorization procedures. With regard to Eurobitec "the negotiation position improved"
as Europhar was "able to convince by means of hard figures". With regard to
authorization procedures, the system and expertise "gave top management greater
confidence in decisions made". As Gisbers and Marcus both indicated a "significant
increase" for "confidence in the decisions made", this point of view is endorsed by
top management itself.

The number of financing alternatives evaluated
As the comparison of the decision process before and after implementation of the
system already shows (see chapter 5 and 6), the number of financing alternatives (at

149

EVALUATION Ch. 7

aggregate level) evaluated during a meeting increased from 4 to 13. The answers from
the participants indicate a significant improvement, mainly due to the analysis and
review task in April of Year 5. "For the first time the whole range of alternative
financing actions were considered at one session. Previously a single alternative would
have been selected on expediency grounds".

Time spent and effectiveness of time spent
Time spent is the only aspect of the system that obtained negative scores. According
to five participants, time spent increased, "mainly in April Year 5". The impression of
the participants that time spent in preparation increased, is supported by the
comparison of the decision processes as described in chapter 5 and 6. The effectiveness
of time spent, however increased substantially, according to the participants. "More
time was spent in developing background knowledge and in developing alternatives -
this must be a good thing". "The quality of time spent was significantly higher". "Due
to interest and the need to understand, more time was involved than originally
anticipated".

Effectiveness of communication
The effectiveness of the communication between participants increased significantly.
According to the comments, the following elements are believed to have contributed
to better communication between the participants:
- all participants were present at one meeting,
- informality of the meeting was combined with a structured approach,
- more background knowledge was available.
It was put forward that "some previous sessions were not an example of effective
communication". This last comment is supported by the descriptive empirical model in
chapter 5. According to five participants, the effectiveness of communication with other
organizational members improved. Communication with the support staff and Eurobitec
were named in this respect.

Structure of the decision process
One of the five highest scores obtained is for the structure of the decision process.
Here is put forward that "previous attempts on group-decision making have resulted
in few decisions of any value being made" and that "before introduction of the system,
the decision process was very unstructured". Use of the software modules and the
facilitator role are considered "essential". The three tasks feasible-solution search,
preference ordering and disaggregation are considered to have contributed significantly
to system performance. The highest scores were for the disaggregation task. Lowest
scores, although still significant, were for the preference-ordering task.

Roles of organizational members
The use of explicitly defined roles and the role of the facihtator did contribute
significantly to system performance. As the authorizer and analyst role already existed,
no significant difference was noticed in comparison with the old situation. With regard
to the facilitator role the comments can be summarized as follows: "This central role
was extremely important to keep the discussion on the correct track. The facihtator
had the role of watchdog. He maintained the decision-process structure. The

150

Sec. 7.3 OVERALL EVALUATION

combination process structure - facilitator can be regarded as the strongest point of
the system. It was important that the facilitator did not try to influence the
participants views." The facilitator role obtained one of the five highest scores. With
regard to the three roles in general, the comment was made that the explicit
introduction of team setting was realistic and necessary. With regard to the analysts
the comment was made that their extra contribution was very much linked to the use
of the software modules. Nijboer's positive contribution during the actualize and review
task was explicitly mentioned.

System use inside and outside the group session
The group session had one of the five highest scores and was described as "the best
meeting on the UK I ever participated in". It was considered that during the group
session, the structure, roles and software were integrated. Due to this integration, the
system "achieved it's proper value" during the group session. Outside the group session,
both the preparation and the Eurobitec negotiations after the session are considered
to have benefitted from the system. Although not all participants were sufficiently
involved in the actualize and review task to give an opinion, the answers that were
obtained were positive. The task was considered to have been "the basis of the group
session and the follow up". In itself, however, this task made "relatively little impact".
The Eurobitec negotiations after the group session benefitted from the system due to
better understanding of the "integral relationship" between the various aspects of the
problem. As not all participants were involved, however, scores are not significant.

All six software modules
The contribution of the six software modules to system performance obtained a "top-
five" score. The "integration of all modules was effective". "Computerized support
played an extremely important part in speeding up the process. Previously such
decisions would have been reached only after a series of meetings which would have
required the attendance of the whole group at each meeting. This is not practical in
view of the time constraints of each participant." "Personal Wizard proved to be an
adequate piece of software".

Problem specification
Problem specification also obtained a "top-five" score. Exclusion of operational and/or
relocation alternatives is considered "unavoidable" because "inclusion would have led to
a vague discussion as the expertise with regard to these matters was missing amongst
participants". Limitation of the problem space, however, "is only possible when the
budget (especially sales) has been actualized in advance". The inclusion of non-pricing
financing alternatives was merited as "it enabled participants to concentrate on all
financing problems simultaneously".

7.4 Conclusion
As explained in chapter 4, our design approach was explicitly to keep the system
"simple and sufficient". In an internal Europhar publication, see Van der Ven [1989a],
the system is described by Europhar as "advanced and sophisticated". The comments

151

EVALUATION Ch. 7

on the multi-criteria technique also indicate that the participants have another view on
the simplicity of the system.

A second conclusion is that none of the system aspects in itself is sufficient to provide
effective decision support. It is the integration of problem specification, decision-
process structure, roles of organizational members and computerized support that aid
the participants in their decision-making task. It is the entire concept, and not merely
the software modules or one of the other aspects that leaded to improved decision
making. To stress this point, we would like to finish this chapter with a piece of
"anecdotal evidence". In the beginning of May Year 5, just after the group session, the
author, who also performed the role of facilitator, had an informal chat with Katwijk.
Katwijk said the following:

"Over the weekend, I have been thinking. We have been struggling with problems in
the UK for several years on end. We did not manage to master the problem. Now
you come around with your system, sway your arms in front of us during two days,
and see, you manage to 'solve' our problem as far as the UK problem can be solved.
How can you succeed where we could not? We are not that stupid and you are not
that smart.

I have been thinking. It was not you, Mike, neither the software, nor us. What you
did was, you blended all know-how available in the group together by means of the
structure and you combined it with computers which checked our expectancies by
means of 'what-if. It was the combination of all things which made it a success".

152

8

SUMMARY AND CONCLUSIONS
8.1 Summary of findings
International transfer pricing is a good example of an area where multinational
corporations are confronted with a "hostile, complex and turbulent" environment.
Consequently, it is a good test case for the approach to aid organizational decision
makers as described by Huber [1984b]. In chapter 1, we formulated our hypothesis
that group decision support systems are helpful for improving decision making with
regard to transfer prices in the pharmaceutical industry. The first research question
(A') which we tried to answer in this thesis is:

What are the characteristics of the prescriptive conceptual model?

The three main conclusions with regard to this first research question are:

- The provision of computerized decision support is in itself not sufficient for improving
group decision making. In addition to computerized tools, clear roles of the support
staff and participants should be defined. Problem specification, use of support tools
and roles of organizational members should be integrated by means of a
prescriptive decision-process structure.

- Within organizations, usually sufficient hardware and software is available to provide
group decision support. In our case, we used two personal computers and the
software tool Personal Wizard. High technology is not a prerequisite for effective
decision support. It might even be argued that sophisticated technology distracts the
attention from more important issues like problem specification, decision-process
structure and roles. From a cost perspective, our approach to group-decision
support is a viable one. No expensive hardware is needed to set up a decision
room. The decision room is created on an ad-hoc basis by using the personal
computers of the participants involved. Besides the cost aspect, this approach offers
the possibility to experiment freely and to introduce group decision support
gradually into the organization. We consider our approach especially valuable for
providing group decision support at the administrative level, see also chapter 4.

- The multi-criteria technique used in the preference-ordering module was clearly new
to the participants. They considered the approach complex and requested additional
training. On the basis of literature study, however, we considered our approach
extremely simple and exactly fitting into our "simple and sufficient" approach. In
chapter 4, we discussed the practicality gap. The evaluation of the preference-
ordering module indicates that this practicality gap is even larger than we expected.

The second research question (B') that we tried to answer in this thesis is:

153

SUMMARY AND CONCLUSIONS Ch. 8

What are the effects of the use of the prescriptive empirical model on organizational
decision making?

In complex cases like ours it is very difficult to establish a direct cause - effect
relationship between company goals and the instruments used. The aim of the system
is to aid decision makers by means of implementation of the prescriptive conceptual
model. Improvement of decision making is the focus of attention. As it is in our case
impossible to measure the direct impact of the system on company goals, evaluation
of the experimental system took place by means of a scala of secondary criteria for
performance measurement. These criteria relate to:
- the quality of the chosen solution as perceived by participants and top management,
- the actual decision process and actual use of the system compared to the
prescriptive conceptual model and the descriptive empirical model,
- the improvement of the decision process as perceived by top management and the
participants.
Applying these criteria shows that the quality of the decision process has improved.

Although the system is developed specifically for the Europhar company, the points of
departure of the prescriptive conceptual model have been checked by means of:
- a survey among transfer-price experts, see Van der Ven [1989a],
- study of transfer-pricing literature,
- study of the environment and structure of the pharmaceutical industry.
Hence we made it plausible that within most European pharmaceutical companies:
- international transfer pricing, and not management-control transfer pricing is the
focus of attention,
- transfer-price determination is a part of the problem of local company financing
and transfer prices are one financing instrument amongst various others,
- transfer prices are decided upon by a group of experts and managers.
As system use and system evaluation at Europhar both indicate an improved decision
process, we consider this experiment a first test of a more generalizable prescriptive
conceptual model. On the basis of various contacts within the European
pharmaceutical industry, we expect that our approach can contribute to more effective
decision making for other pharmaceutical companies as well.

Our experience might be helpful for others who are active in the area of designing
group decision support systems. It is our impression that in design, careful attention
should be paid to the following issues:

- Proper analysis of the task structure is of overriding importance and should receive
sufficient attention. In this respect, we consider the technique as developed by Bots
and Sol [1987] useful. This technique is an excellent aid to "structure an ill-
structured problem".

- In group decision support literature, the problem of "minimum frequency of use" of
a system is usually approached by means of widening the scope of the group
decision support system. The design should be such that group decision making
related to various organizational areas can be supported, see Huber [1984b].

154

Sec. 8.1 SUMMARY OF FINDINGS

Consequently, the support system should be of a general nature. In our
experimental system, however, another approach is taken. We use a relatively easy
to use software tool (Personal Wizard). Frequency of use should be guaranteed by
use of the developed application-domain support by decision makers. We focus on
integrating the decision support into the organization. So support should be
available for all those who are concerned with determining transfer prices and other
financing alternatives, whether working individually or with others. We have the
impression that the importance of application-domain knowledge is not sufficiently
stressed in current (group) decision support literature.

- When developing group decision support systems for use within organizations, it
should be accepted that some decision makers are computer illiterate and wish to
remain that way. These decision makers want to concentrate on their own area of
expertise and prefer not to be "bothered" by computers. Therefore, not every
participant should be forced to provide (some) of the input for the group session
via a terminal. An alternative way of using a group decision support system is to
have a staff member available who operates the computer in his role of "analyst".
System use via an analyst can be seen as a gradual way of integrating group
decision support in the organization.

- When implementing a (group) decision support system, careful attention should be
paid to the existing infrastructure with regard to software tools available and formal
information flows as provided by existing information systems. Whenever possible,
the existing tools, which are already known to technical support staff and
participants, should be used. Decision support should build on the infrastructure as
far as information systems are concerned. Decision support should focus on making
existing information more effective. It should be avoided to create new formalized
information flows.

8.2 Future research
In essence, see Keen and Scott Morton [1978], development of decision support
systems is an interactive ongoing process. Developments of decision support systems
for transfer pricing should continue. Our research indicates the following new
directions:

- The disaggregation module and the accounting module received the highest scores
during the evaluation interviews. Technically speaking, integration of the
disaggregation module and the accounting module is possible. What would be the
implications for use and maintainability if this would be done? What are the
advantages and disadvantages? How can the accounting- and disaggregation module
be improved?

- The preference-ordering module is considered to be complex. What training is needed
to stimulate effective communication with the help of this tool? How can the
preference-ordering module be improved?.

155

SUMMARY AND CONCLUSIONS Ch. 8

- Participants started to use the system for internal negotiation support. What
enhancements to the system need to be made to support negotiations more
effectively? How can the system be used during negotiations with, for example,
fiscal authorities and health-care services? What facilities should be added to make
it an effective internal and external negotiation support system?

- In an early stage of the development process, it was tried to bring regulations
concerning transfer prices and other financing alternatives together in a knowledge-
based system. After a few discussion drafts, it was realized that the development of
such a system could not be handled within the framework of the research project.
As explained in chapter 6, we decided to store fiscal information in text mode. As
explained in chapter 3, regulations are the main concern in international transfer
pricing. It is our impression that the development of a knowledge-based system for
this area would be a very worthwhile exercise.

- The project limited itself to the decision room and other "close-proximity" decision
support without using electronic mail. Communication between the divisional head
office and the subsidiary, however, is of significant importance. During the six
months of the experiment, frequent communication took place by means of telex,
telefax, letters and visits. What would be the advantages and disadvantages of
including video conferencing, electronic mail and other ways of communication? What
would be the contribution of a local decision network at the head office?

- The project limited itself to the development of an initial "version 0" system, see
chapter 1. The project was an explicit research project. After this "research" effort,
it would be interesting to continue with "development". In this respect, the following
aspects may be considered:

- The set up of a system in other software tools. What are advantages and
drawbacks of using another tool? This question is especially relevant in case we
want to implement the system at a company that uses other software tools for its
own budgeting and reporting purposes.

- The integration of the software modules with already existing systems that provide
financial management information. New budgets and transfer-price information,
for example, could then automatically be incorporated in the software modules.

- According to Brussaard [1982], users often exaggerate the differences that exist
between organizations. On the basis of this exaggeration, they claim that systems
developed by others are not appropriate for their organization. System designers,
on the other hand, also exaggerate, according to Brussaard. They usually claim that
general use of their system is possible. As explained above, we claim that our
prescriptive conceptual model can be used to develop decision support for other
pharmaceutical companies. It would be interesting to test our claim by means of
future experiments.

156

REFERENCES
ABPI, The Pharmaceutical Industry and the Nation's Health, The Association of the
British Pharmaceutical Industry, London, England, 1985.
AKZO, Annual Report 1985, AKZO N.V., Arnhem, The Netherlands, 1985.
ALBERS, L.H., Het Gewichtloze Gewogen, Ph.D. Thesis, Delft University Press, Delft,
1987.
AMMER, C. AND D.S. AMMER, Dictionary of Business and Economics, Free Press, New
York, New York, 1977.
BENBASAT, I., D. GOLDSTEIN AND M. MEAD, "The Case Research Strategy in Studies
of Information Systems," Management Information Systems Quarterly, Vol. 11, No. 3
(1987), pp. 369-386.
BEUC, The Consumer and the Pharmaceutical Products in the European Economic
Community, Bureau Europeen des Unions de Consommateurs, Brussels, Belgium,
1984.
BOOT, M. AND H. KOPPELAAR, "The Natural Computer or Paradigm Lost," in J.
Goetschalckx and L. Rolling (Eds.), Lexicography in the Electronic Age, Elsevier
Science Publishers, Amsterdam, The Netherlands, 1982, pp. 7-18.
BOSMAN, A., Een Metatheorie over het Gedrag van Organisaties, Stenfert Kroese,
Leiden, t h e Netherlands, 1977.
BOSMAN, A., "Decision Support Systems, Problem Processing and Co-ordination," in
H.G. Sol (Ed.), Processes and Tools for Decision Support, Elsevier Science
Publishers, Amsterdam, The Netherlands, 1983, pp. 79-92.
BOSMAN, A., "Ontwerpen, Procesmodellen en Beslissen," Informatie, Vol. 28, No. 4
(1986), pp. 310-321.
BOSMAN, A., "Relations Between Specific Decision Support Systems," Decision Support
Systems, Vol. 3, No. 3 (1987a), pp. 213-224.
BOSMAN, A., "Decision Support Systems: A Discipline or a Vision?," in H.G. Sol, C.A.
Th. Takkenberg and P.F. de Vries Robbé (Eds.), Expert Systems and Artificial
Intelligence in Decision Support Systems, Kluwer, Dordrecht, The Netherlands, 1987b,
pp. 23-39.
BOSMAN, A. AND H.G. SOL, "Knowledge Representation and Information Systems
Design," in L.B. Methlie and R.H. Sprague Jr. (Eds.), Knowledge Representation for
Decision Support Systems, Elsevier Science Publishers, Amsterdam, The Netherlands,
1985, pp. 81-91.
BOTS, P.W.G., An Environment to Support Problem Solving, Ph.D. Thesis, Delft
University Press, Delft, The Netherlands, 1989 (forthcoming).
BOTS, P.W.G. AND H.G. SOL, "An Environment to Support Problem Solving," Decision
Support Systems, Vol. 3, No. 3 (1987), pp. 225-231.
BOTS, P.W.G. AND H.G. SOL, "Shaping Organizational Information Systems through
Co-ordination Support," in' R.M. Lee, A.M. McCosh and P. Migliarese (Eds.),
Organizational Decision Support Systems, Elsevier Science Publishers, Amsterdam,
The Netherlands, 1988, pp. 139-154.
BOUEE, J.B., "Potential Evolutions of Medical, Industrial, and Distribution Practices in
the Japanese Pharmaceutical Business Strategy of the Japanese Government," in
Pricing and Forecasting in the Pharmaceutical Industry: A Matter of Growing Concern,

157

References

EPHMRA/ESOMAR Seminar Proceedings, Copenhagen, Denmark, 1987, pp. 39-
41.
BRATTHWAITE, J., Corporate Crime in the Pharmaceutical Industry, Routledge & Kegan
Paul, London, England, 1984.
BRANDT, W. ET AL, North-South: A Program for Survival, Pan Books, London,
England, 1980.
BREUK, R., "Promotie van Farmaca," Wemos Bulletin, No. 6 (1985), pp. 4-6.
BRUSSAARD, B.K., "Datapraet, Wat Bestuurders van Automatisering Moeten Weten,"
Informatie, Vol. 24, No. 9 (1982), p. 480.
BUI, X.T, AND M. JARKE, "A DSS for Cooperative Multiple Criteria Group Decision
Making," in Proceedings of the 5th International Conference on Information Systems,
1984, pp. 101-113.
BUI, X.T., M. JARKE, Communications Design for Co-op: A Group Decision Support
System, ACM Transactions on Office Information Systems, Vol. 4, No. 2 (1986), pp.
81-103.
BURSTALL, M.L., The Community's Pharmaceutical Industry, Office for Official
Publications of the European Communities, Luxemburg, Luxemburg, 1985.
CHANDLER, A.D., The Visible Hand: the Managerial Revolution in American Business,
Harvard University Press, Cambridge, Massachusetts, 1977.
CLAESSENS, M.N.A.J. AND F.J. VOGT, Optimaal Beslissen bij Meerdere Criteria, Internal
Report, Delft University of Technology, Delft, The Netherlands, 1988.
COMPTAX, An International Tax Management System, Comptax, Helsingborg, Sweden,
1987.
COMSHARE, "PWlearn," Comshare Nieuws, Vol. 6, (1988), p. 7.
COMSHARE, "Tips & Hints," Comshare Nieuws, Vol. 7, Special Edition (1989), p. 3.
CONKLIN, H., "Ethnography," in D.L. Sills (Ed.), International Encyclopedia of the
Social Sciences, No. 5, Free Press, New York, New York, 1968, pp. 115-208.
CONOVER, W.J., Practical Non-parametric Statistics, John Wiley and Sons, New York,
New York, 1980.
COOK, R.L. AND K.R. HAMMOND, "Interpersonal Learning and Interpersonal Conflict
Reduction in Decision-Making Groups," in R.A. Guzzo (Ed.), Improving Group
Decision Making in Organizations, Academic Press, New York, New York, 1982, pp.
13-40.
DAU, MAUD User Manual, Decision Analysis Unit, University of London, London,
England, 1986a.
DAU, A Brief Description of MAUD, Decision Analysis Unit, University of London,
London, England, 1986b.
DAVIS, J.H., Group performance, Addison-Wesley, Reading, Massachusetts, 1969
DAVIS, G.B. AND M.H. OLSON, Management Information Systems: Conceptual
Foundations, Structure, and Development, McGraw-Hill, New York, New York, 1985.
D E JONG, H.W., "Competition and Economic Power in the Pharmaceutical Industry,"
in H.W de Jong (Ed.), The Structure of European Industry, Martinus Nijhoff, Leiden,
DESANCns, G. AND B. GALLUPE, "Group Decision Support Systems, A New Frontier,"
Database, Vol. 16, No. 2 (1985), pp. 3-10.
DESANCns, G. AND B. GALLUPE, "A Foundation for the Study of Group Decision
Support Systems," Management Science, Vol. 33, No. 5 (1987), pp. 589-609.

158

References

DHSS, The Pharmaceutical Price Regulation Scheme 1986, Department of Health and
Social Security, London, England, 1986.
DOWNEY, H.K. AND R.D. IRELAND, "Quantitative versus Qualitative: Environmental
Assessment in Organizational Studies," Administrative Science Quarterly, Vol. 24, No.
1 (1979), pp. 630-637.
ECCLES, R.G., The Transfer Pricing Problem, A Theory for Practice, Lexington Books,
Lexington, Massachusetts, 1985.
EDEN, L., "The Microeconomics of Transfer Pricing," in A.M. Rugman and L. Eden
(Eds.), Multinationals and Transfer Pricing, Croom Helm, London, England, 1985,
pp. 13-46.
EFPIA, EFPIA in Figures, The Pharmaceutical Industry in Europe 1985-1986, European
Federation of Pharmaceutical Industries' Associations, Brussels, Belgium, 1987.
EG, "Richtlijn van de Raad van 21 December 1988," Publicatieblad van de Europese
Gemeenschappen, Vol. 32, No. L 40 (1989), pp. 8-11.
FITZGERALD ET AL., "Introduction," in E. Mumford et al. (Eds.), Research Methods
in Information Systems, Elsevier Science Publishers, Amsterdam, The Netherlands,
1985, pp. 3-9.
FORTUIN, L. AND F.A. LOOTSMA, "Future Directions in Operations Research," in
A.H.G. Rinnooy Kan (Ed.), New Challenges for Management Research, Elsevier
GALBRATTH, J.R., Organization Design, Addison-Wesley, Reading, Massachusetts, 1977.
GiNZBERG, M.J. AND E.A. STOHR, "DSS: Issues and Perspectives," in M.J. Ginzberg,
W. Reitman and E.A. Stohr (Eds.), Decision Support Systems, Elsevier Science
GRAY, P., "Group Decision Support Systems," in E.R. McLean and H.G. Sol (Eds.),
Decision Support Systems: A Decade in Perspective, Elsevier Science Publishers,
Amsterdam, The Netherlands, 1986, pp. 157-171.
GUZZO, R.A., "The Study of Group Decision Making: Approaches and Applications,"
in R A Guzzo (Ed.), Improving Group Decision Making in Organizations, Academic
Press, New York, New York, 1982, pp. 1-11.
HAAYER-RUSKAMP, F.M. AND M.N.G. DUKES, Drugs and Money, University of
Groningen, Groningen, The Netherlands, 1986.
HANKIN, R., "The Transparency Directive - Progress and Timetable," in Pharmaceutical
Product Price Control, and the EEC Transparency Directive, Conference Proceedings,
Legal Studies & Services, London, England, 1988.
HANSEN, R., "Pharmaceutical Regulation: The Past and Future Twenty-five Years," in
G.T. Smith (Ed.), Health Economics: Prospects for the Future, Croom Helm, London,
England, 1987, pp. 259-268.
HARNETT, D.L., Statistical Methods, Addison-Wesley, Reading, Massachusetts, 1982.
HlRSCHLEiFER, J., "On the Economics of Transfer Pricing," The Journal of Business,
Vol. 29, No. 3 (1956), pp. 172-184.
HUBER, G.P., "Issues in the Design of Group Decision Support Systems," Management
Information Systems Quarterly, Vol. 8, No. 3 (1984a), pp. 195-204.
HUBER, G.P., "The Nature and Design of Post-Industrial Organizations," Management
Science, Vol. 30, No. 8 (1984b), pp. 928-951.
HUBER, G.P. AND R.R. MCDANIEL, "The Decision-Making Paradigm of Organizational
Design," Management Science, Vol. 32, No. 5 (1986), pp. 572-589.

159

References

IBFD, The Taxation of Companies in Europe, International Bureau of Fiscal
Documentation, Amsterdam, The Netherlands, 1984.
IBFD, Annual Report 1986, International Bureau of Fiscal Documentation, Amsterdam,
The Netherlands, 1986.
IBFD, The Tax Treatment of Transfer Pricing, International Bureau of Fiscal
Documentation, Amsterdam, The Netherlands, 1987.
IMS, The Pharmaceutical Market Review 1987, IMS World Publications, London,
England, 1988.
IVES, B., M.H. OLSON AND J.J. BAROUDI, "The Measurement of User Information
Satisfaction," Communications of the ACM, Vol. 26, No. 10 (1983), pp. 785-793.
JAQUES, E., A General Theory of Bureaucracy, Heinemann, London, England, 1976.
JARKE, M., "Group Decision Support through Office Systems: Developments in
Distributed DSS Technology," in E.R. McLean and H.G. Sol (Eds.), Decision
Support Systems: A Decade in Perspective, Elsevier Science Publishers, Amsterdam,
JARKE, M. AND M.T. JELASSI, "View Integration in Negotiation Support Systems,"
DSS-86 Transactions, (1986), pp. 180-188.
JARKE, M., M.T. JELASSI AND M.F. SHAKUN, "Mediator: Towards a Negotiation
Support System," European Journal of Operational Research, Vol. 31, No. 3 (1987),
pp. 314-334.
KASSIC1EH, S.K., "A Decision Support System for Transfer Pricing Environments,"
Proceedings of the 18'h Annual Hawaii International Conference on System Sciences,
University of Hawaii, Honolulu, Hawaii, 1985, pp. 439-445.
KEEN, P.W.G., "Value Analysis: Justifying Decision Support Systems," in R.H. Sprague
Jr. and H.J. Watson (Eds.), Decision Support Systems, Putting Theory into Practice,
Prentice-Hall, Englewood Cliffs, New Jersey, 1986, pp. 48-64.
KEEN, P.G.W. AND M.S. SCOTT MORTON, Decision Support Systems, An Organizational
Perspective, Addison-Wesley, Reading, Massachusetts, 1978.
KOPPELAAR, H., "Intelligente Kantoornetwerken," Kennissystemen, Vol. 2, No. 5 (1988),
pp. 27-29.
KRENS, F., "Overdrachtsprijzen," in P.J.W. Duffhues, J.G. Groeneveld and J.A.M.
Oonincx, Financiële Leiding en Organisatie, Samson, Alphen aan den Rijn, The
Netherlands, 1979, pp. 2650.1-2650.27.
LANDRETH, H., History of Economie Theory: Scope, Method and Content, Houghton
Mifflin, Atlanta, Dallas, 1976.
LARICHEV, O.I., "Systems Analysis and Decision Making," in P. Humphreys, O.
Svenson and A. Vari (Eds.), Analyzing and Aiding Decision Processes, Elsevier
LEE, S. AND J. SHIM, Micro Management Science, Microcomputer Applications of
Management Science, Wm. C. Brown Publishers, Dubuque, Iowa, 1986.
LITTLE, J.D.C., "Models and Managers: The Concept of a Decision Calculus,"
Management Science, Vol. 16, No. 8 (1971), pp. 466-485.
LOOTSMA, F.A., "Wikken en Wegen van Wezen en Schijn," Informatie, Vol. 30, No.
7/8 (1988), pp. 461-540.
LOOTSMA, F.A., J. MEISNER AND F. SCHELLEMANS, "Multi-criteria Decision Analysis
as an Aid to the Strategic Planning of Energy R & D," European Journal of
Operations Research, Vol. 25, No. 2 (1986), pp. 216-234.

160

References

LORSCH, J.W. AND P.R. LAWRENCE (Eds.), Studies in Organization Design, R.D. Irwin
and The Dorsey Press, Homewood, Illinois, 1970.
MARCH, J.G. AND H.A. SIMON, Organizations, John Wiley & Sons, New York, New
York, 1958.
MCLEAN, E.R. AND H.G. SOL (Eds.), Decision Support Systems: A Decade in
Perspective, Elsevier Science Publishers, Amsterdam, The Netherlands, 1986.
MEIJBOOM, B.R., Planning in Decentralized Firms, A Contribution to the Theory on
Multilevel Decisions, Ph.D. Thesis, Tilburg University, Tilburg, The Netherlands,
1986.
MINTZBERG, H., "An Emerging Strategy of 'Direct' Research," Administrative Science
Quarterly, Vol. 24, No. 1 (1979), pp. 582-589.
MOSTERMANS, Th.G.A., Het Planningsproces in een Gedecentraliseerde Onderneming,
Master's Degree Thesis, Tilburg University, Tilburg, The Netherlands, 1982.
MURRAY, R., "Introduction," in R. Murray (Ed.), Multinationals Beyond the Market,
Intra-firm Trade and the Control of Transfer Pricing, The Harvester Press, Brighton,
England, 1981, pp. 1-14.
NATH, R., Comparative Management: A Regional View, Ballinger, Cambridge, England,
1988.
NEFARMA, Jaarverslag Nederlandse Asociatie van de Farmaceutische Industrie, Nefarma,
Utrecht, The Netherlands, 1985.
NEWBOLD, P., Statistics for Business and Economics, Prentice-Hall, Englewood Cliffs,
New Jersey, 1988.
NOWOTNY, O.H., "Verrechnungspreise Zwischen Verbundenen Unternehmen," Pharma
Industrie, Vol. 49, No. 4 (1987), pp. 339-346.
OECD, Transfer Pricing and Multinational Enterprises, Report of the OECD Committee
on Fiscal Affairs, OECD, Paris, France, 1979.
OECD, Transfer Pricing and Multinational Enterprises, Three Taxation Issues, Report of
the OECD Committee on Fiscal Affairs, OECD, Paris, France, 1984.
OWEN, M., "Difficulties for Companies Imposed by Variability," in Pharmaceutical
Product Price Control, and the EEC Transparency Directive, Conference Proceedings,
Legal Studies & Services, London, England, 1988.
PBPA, Notes on 'factor (F)' Seminars, Pharmaceutical Benefits Pricing Authority,
Sidney, Australia, 1988.
PHILLIPS, L.D., "Decision Support for Managers," in H.J. Otway and M. Peltu (Eds.),
77ie Managerial Challenge of New Office Technology, Butterworths, London, England,
1984, pp. 80-98.
PHILLIPS, L.D., "Systems for Solutions," Datamation Business, Vol. 31, Special Report
(1985), pp. 26-29.
PHILLIPS, L.D., "Computing to Consensus," Datamation, Vol. 32, No. 19 (1986), pp.
68.2-68.6.
PLASSCHAERT, S.R.F., Transfer Pricing and Multinational Corporations: An Overview of
Concepts, Mechanisms and Regulations, Saxon House, Westmead, England, 1979.
PLASSCHAERT, S.R.F., "Transfer Pricing Problems in Developing Countries," in A.M.
Rugman and L. Eden (Eds.), Multinationals and Transfer Pricing, Croom Helm,
London, England, 1985, pp. 247-266.

161

References

QUIRIN, G.D., "Fiscal Transfer Pricing: Accounting for Reality," in A.M. Rugman and
L. Eden (Eds.), Multinationals and Transfer Pricing, Croom Helm, London, England,
1985, pp. 132-148.
RASCH, R.H., An Analysis of Information Structure and Optimal Transfer Pricing
Decision Rules for Decentralized Firms, Ph.D. Thesis, University of Texas, Austin,
Texas, 1980.
REINSHAGEN, P., "Twee Miljard Exportguldens, De Farmacie is een Bedrijfstak die We
Onvoldoende Koesteren," Intermediair, Vol. 24, No. 48 (1988), pp. 47-51.
ROY, B., Methodologie Multicritè d'Aide a la Decision, Economica, Paris, France, 1985.
SCHNELLE, E., The Metaplan-method, Metaplan, Quickborn, West-Germany, 1977.
SCRIP, "Satisfaction for UK Industry," Scrip, No. 1232 (1987a), p. 5.
SCRIP, "UK Price Increases Coming Through," Scrip, No. 1268 (1987b), p. 2.
SCRIP, "Registration and Patent Concerns Dominate," Scrip, No. 1270/1271 (1988), pp.
1-5.
SIMON, H.A., "Applying Information Technology to Organization Design," Public
Administration Review, Vol. 33, No. 3 (1973), pp. 268-278.
SIMON, H.A., Administrative Behavior, MacMillan, New York, New York, 1976.
SIMON, H.A., The New Science of Management Decision, Prentice-Hall, Englewood
Cliffs, New Jersey, 1977.
SIMON, H.A. ET AL, "Decision Making and Problem Solving," Interfaces, Vol. 17, No.
5 (1987), pp. 11-31.
SOL, H.G., Simulation in Information Systems Development, Ph.D. Thesis, University of
Groningen, Groningen, The Netherlands, 1982.
SOL, H.G., "Aggregating Data for Decision Support," Decision Support Systems, Vol. 1,
No. 2 (1985a), pp. 111-121.
SOL, H.G., "DSS: Buzzword or OR Challenge?," European Journal of Operational
Research, Vol. 22, No. 1 (1985b), pp. 1-8.
SOL, H.G., "Paradoxes around DSS," in C.W. Holsapple and A.B. Whinston (Eds.),
Decision Support Systems: Theory and Application, Springer-Verlag, Berlin, West-
Germany, 1987a, pp. 3-18.
SOL, H.G., "Conflicting experiences with DSS," Decision Support Systems, Vol. 3, No.
3 (1987b), pp. 203-211.
SOL, H.G. AND M.B.M. VAN DER VEN, "A Group Decision Support System for
International Transfer Pricing Decisions in the Pharmaceutical Industry," in M.G.
Singh, K.S. Hindi and D. Salassa (Eds.), Managerial Decision Support Systems,
Elsevier Science Publishers, Amsterdam, The Netherlands, 1988a, pp. 81-94.
SOL, H.G. AND M.B.M. VAN DER VEN, "Integrating GDSS in the Organization: The
Case of a GDSS for International Transfer Pricing," in R.M. Lee, A.M. McCosh
and P. Migliarese (Eds.), Organizational Decision Support Systems, Elsevier Science
Publishers, Amsterdam, The Netherlands, 1988b, pp. 15-30.
SOL, H.G. AND M.B.M. VAN DER VEN, "Designing and Evaluating Decision Support
Systems: The Case of a Group DSS for International Transfer Pricing," in J.
Klabbers, D. Crookall, H. de Jong and W. Schepers (Eds.), Simulation-Gaming:
On the Improvement of Competence in Dealing with Complexity, Uncertainty and
Value Conflicts, Pergamon Press, Oxford, England, 1989 (forthcoming).
SPRAGUE, R.H. JR. AND E.D. CARLSON, Building Effective Decision Support Systems,
Prentice-Hall, Englewood Cliffs, New Jersey, 1982.

162

References

SPRAGUE, R.H. JR. AND B.C. MCNURLIN, Information Systems Management in Practice,
Prentice-Hall, Englewood Cliffs, New Jersey, 1986.
SPRAGUE, R.H. JR., AND H.J. WATSON, University Support Programs Offered by
Vendors of DSS Generators, in R.H. Sprague Jr. and H.J. Watson (Eds.), Decision
Support Systems, Putting Theory into Practice, Prentice-Hall, Englewood Cliffs, New
Jersey, 1986, pp. 266-274.
SUCHAN, J., T. BUI AND D. DOLK, "GDSS Effectiveness: Identifying Organizational
Opportunities," in Proceedings of the 20th Hawaii International Conference on
Information System Sciences, Western Periodicals Company, North Hollywood,
California, 1987, pp. 441-449.
TAKKENBERG, CA.Th., Planning en Methode van Onderzoek, Ph.D. Thesis, University
of Groningen, Groningen, The Netherlands, 1983a.
TAKKENBERG, CA.Th., "CAP: A Decision Support System for the Planning of
Production Levels," in H.G. Sol (Ed.), Processes and Tools for Decision Support,
Elsevier Science Publishers, Amsterdam, The Netherlands, 1983b, pp. 249-259.
UHR, L., Pattern Recognition, Learning and Thought, Prentice-Hall, Englewood Cliffs,
New Jersey, 1973.
VAITOS, C.V., Intercountry Income Distribution and Transnational Enterprises, Clarendon
Press, Oxford, England, 1974.
VAN CAYSEELE, P., Regulation and International Innovative Activities in the
Pharmaceutical Industry, Ph.D. Thesis, Leuven University, Leuven, Belgium, 1987.
VAN DER PLAS, A.H.W.M., Een Planningsmethode voor een Onderneming, Master's
Degree Thesis, Tilburg University, Tilburg, The Netherlands, 1983.
VAN DER SCHROEFF, H.J. AND J.G. GROENEVELD, Kosten en Kostprijs, Kosmos,
Amsterdam, The Netherlands, 1980.
VAN DER VEN, M.B.M., System Documentation of the TP-DSS System: Component I,
Section Report, Section of Information Systems, Delft University of Technology,
Delft, The Netherlands, 1988a.
VAN DER VEN, M.B.M., System Documentation of the TP-DSS System: Component II,
Delft, The Netherlands, 1988b.
VAN DER VEN, M.B.M., System Documentation of the TP-DSS System: Component III
and Logging Facility, Section Report, Section of Information Systems, Delft
University of Technology, Delft, The Netherlands, 1988c.
VAN DER VEN, M.B.M., A Group Decision Support System for Transfer Pricing in the
Pharmaceutical Industry: The Evaluation Interviews, Section Report, Section of
Information Systems, Delft University of Technology, Delft, The Netherlands, 1989a.
VAN DER VEN, M.B.M., A Group Decision Support System for Transfer Pricing in the
Pharmaceutical Industry: Decision Making before Introduction of the TP-DSS System,
Delft, The Netherlands, 1989b.
VAN SCHAIK, F.D.J., "Effectiveness of Decision Support Systems," Data Base, Vol. 18,
No. 3 (1987), pp. 58-59.
VAN SCHAIK, F.D.J., Effectiveness of Decision Support Systems, Ph.D. Thesis, Delft
University Press, Delft, The Netherlands, 1988.
VERDONER, L.A., Fiscale Aspecten van Intercompany Pricing in Internationaal Verband,
Ph.D. Thesis, Kluwer, Deventer, The Netherlands, 1988.

163

References

VERENIGING VAN BELASTING WETENSCHAPPEN, Internationale Winstallocatie, Kluwer,
Deventer, The Netherlands, 1987.
VERLAGE, H.C., Transfer Pricing for Multinational Enterprises, Ph.D. Thesis, Rotterdam
University Press, Rotterdam, The Netherlands, 1975.
VERLAGE, H.C, "De Rol van Overheidsprijzen bij de Financiering van Multinationale
Ondernemingen," Bedrijfskunde, Vol. 52, Nr. 3 (1980), pp. 269-277.
VOGEL, D.R., J.F. NUNAMAKER JR., J.F. GEORGE AND A.R. DENNIS, "Group Decision
Support Systems: Evolution and Status at the University of Arizona," in R.M. Lee,
A.M. McCosh and P. Migliarese (Eds.), Organizational Decision Support Systems,
Elsevier Science Publishers, 1988, pp. 287-304.
VON GREBMER, K., "International Transfer Pricing in the Pharmaceutical Industry,"
Intertax, Vol. 4, No. 5 (1987a), pp. 92-99.
VON GREBMER, K., "Pricing Medicines," in G.T. Smith (Ed.), Health Economics:
Prospects for the Future, Croom Helm, London, England, 1987b, pp. 229-249.
VON WlNTERFELDT, D., "Pitfalls of Decision Analysis," in P. Humphreys, O. Svenson
and A. Vari (Eds.), Analyzing and Aiding Decision Processes, Elsevier Science
WnTE, E., "Field Research on Complex Decision-Making Processes - The Phase
Theorem," in International Studies of Management and Organization, (1972), pp. 156-
182.
WOLCOTT, H., "Criteria for an Ethnographic Approach to Research in Schools,"
Human Organization, No. 34 (1975), pp. 111-128.
YIN, R.K., Case Study Research, Sage Publications, Beverly Hills, California, 1984.
ZELENY, M., Multiple Criteria Decision Mating, McGraw-Hill, New York, New York,
1982.

164

List of exhibits

Exhibit 1-1. Categories of models 5
Exhibit 1-2. Overview of the research approach 8
Exhibit 2-1. Organizational structure of a typical pharmaceutical company. . 14
Exhibit 2-2. First part of the NHS-profit calculation 17
Exhibit 2-3. Second part of the NHS-profit calculation 18
Exhibit 3-1. Eccles' contingency approach to management control 26
Exhibit 3-2. Financial impact of a transfer-price decrease 35
Exhibit 3-3. Financial impact of an increase in current account 35
Exhibit 3-4. Financial impact of an intracompany loan 36
Exhibit 3-5. Calculation of the interest payments for an intracompany loan. 36
Exhibit 3-6. Summary of the financial-impact results 37
Exhibit 3-7. The place of pharmaceutical companies in Eccles' contingency
approach to management control 39
Exhibit 3-8. The importance of various kinds of transfer pricing for the
pharmaceutical industry. 40
Exhibit 3-9. The basic pattern between source company, object company and
destination 41
Exhibit 3-10. Relationships within a multinational 42
Exhibit 4-1. Proposed structure of the problem-solving task 52
Exhibit 4-2. Proposed structure of the decision-making task 53
Exhibit 4-3. Formula for normalized weight 55
Exhibit 4-4. Formula for total score 55
Exhibit 4-5. Table with total scores 55
Exhibit 4-6. A basic taxonomy of group decision support systems 57
Exhibit 4-7. Possible set up of an ad-hoc decision room 61
Exhibit 5-1. Organizational scheme of the Eurochem group 71
Exhibit 5-2. Relationships of object company Europhar UK Ltd 73
Exhibit 5-3. Summary of financing alternatives decided upon at the end of the
meeting in July Year 4 77
Exhibit 5-4. Summary of financing alternatives decided upon in August
Year 4 78
Exhibit 5-5. Frequency distribution of actions taking place durmg the meeting
of July Year 4 80
Exhibit 5-6. Structure of the problem-solving task in the descriptive conceptual
model 82
Exhibit 5-7. Structure of die scan task 83
Exhibit 5-8. Structure of the decision-making task in the descriptive conceptual
model 84
Exhibit 6-1. The prescriptive conceptual model compared to the descriptive
conceptual model 88
Exhibit 6-2. Structure of the problem-solving task in the prescriptive conceptual
model 90
Exhibit 6-3. Structure of the decision-making task in the prescriptive conceptual
model 92
Exhibit 6-4. Structure of the feasible-solution search task 95
Exhibit 6-5. Structure of the preference-ordering task 97
Exhibit 6-6. Global outline of the computerized support 99
Exhibit 6-7. Four types of variables. . . . ' 100

165

List of exhibits

Exhibit 6-8. Main menu of the accounting module 101
Exhibit 6-9. Summary key-figures screen 101
Exhibit 6-10. Main menu of the financial-impact module 102
Exhibit 6-11. Actual-remittances screen 103
Exhibit 6-12. Screen with weights per participants 104
Exhibit 6-13. Example of an input/output matrix presenting conversion factors. 105
Exhibit 6-14. Screen with destination ratios 106
Exhibit 6-15. The accounting module 112
Exhibit 6-16. Submenu of the financial-impact module 113
Exhibit 6-17. The preference-ordering module 114
Exhibit 6-18. The disaggregation module 115
Exhibit 6-19. Summary key-figures screen (at start of the group session). . . 120
Exhibit 6-20. Summary of financing, commercial and fiscal targets 122
Exhibit 6-21. Screen with destination ratios (profit and loss iteration) 124
Exhibit 6-22. Maximal remittances per alternative (profit and loss iteration). 125
Exhibit 6-23. Example of a preference-ordering worksheet (printout) 126
Exhibit 6-24. Screen with actual remittances (profit and loss iteration) 126
Exhibit 6-25. Screen with normalized weights (profit and loss iteration). . . . 127
Exhibit 6-26. Summary key-figures screen (after profit and loss iteration). . . 128
Exhibit 6-27. Maximal remittances per alternative (balance-sheet iteration). . 129
Exhibit 6-28. Screen with actual remittances (balance-sheet iteration) 130
Exhibit 6-29. Screen with normalized weights (balance-sheet iteration) 130
Exhibit 6-30. Summary key-figures screen (after balance-sheet iteration). . . . 131
Exhibit 6-31. Old and new transfer prices for generic products 131
Exhibit 6-32. Old and new internal selling prices to Taiwan 131
Exhibit 6-33. The production process for Sofanaat 133
Exhibit 6-34. Use of the various modules during the experimentation period. 134
Exhibit 7-1. The indirect effect between financing decisions and company
goals 138
Exhibit 7-2. Comparison of frequency distributions of actions taken during two
meetings 139
Exhibit 7-3. A letter from the UK general manager 140
Exhibit 7-4. Frequency distribution of answers on impact of system use. . . 142
Exhibit 7-5. Frequency distribution of answers on contribution of system
elements 143
Exhibit 7-6. Table with probabilities 145
Exhibit 7-7. Overall judgement on impact of system use 146
Exhibit 7-8. Overall judgement on contribution of system elements 147

166

Samenvatting in het Nederlands

EEN GROEPS-BESLISSINGSONDERSTEUNEND SYS
TEEM VOOR TRANSFERPRIJS BESLISSINGEN IN DE
PHARMACEUTISCHE INDUSTRIE
Dit proefschrift brengt verslag uit van een onderzoek naar de mogelijkheden van
groeps-beslissingsondersteunende systemen voor het oplossen van transferprijs problemen
in de pharmaceutische industrie. Onderzocht is welke de eigenschappen van een
dergelijk systeem zouden moeten zijn (onderzoeksvraag 1). Vervolgens is een
experimenteel systeem met deze eigenschappen gebouwd. Deze experimentele versie is
gebruikt voor het oplossen van transferprijs problemen in de praktijk. Op basis hiervan
is een indruk verkregen van de geschiktheid van het ontworpen systeem om
besluitvorming te ondersteunen (onderzoeksvraag 2).

Vier modellen en hun onderlinge relatie staan centraal in dit onderzoek. Deze
modellen zijn:
A. een beschrijvend empirisch model,
B. een beschrijvend conceptueel model,
C. een voorschrijvend conceptueel model,
D. een voorschrijvend empirisch model.
Een beschrijvend empirisch model (A) beschrijft de besluitvorming zoals die in een
specifieke pharmaceutische onderneming is aangetroffen. Met name op basis van dit
model is de achterliggende besluitvormingsstruktuur blootgelegd: een beschrijvend
conceptueel model (B). Vervolgens is een voorschrijvend conceptueel model (C)
ontworpen. Dit model geeft weer hoe de besluitvorming op effektievere wijze zou
kunnen plaatsvinden. Het voorschrijvend conceptueel model is gebaseerd op analyse
van de bestaande situatie (het beschrijvend conceptueel model (B)) en studie van
literatuur op het gebied van beslissingsondersteuning, transferpnjzen en de
pharmaceutische industrie. Het voorschrijvend conceptueel model geeft een antwoord
op onderzoeksvraag 1. De implementatie van dit voorschrijvend conceptueel model
leidt tot wat is gedefinieerd als het voorschrijvend empirisch model (D). De evaluatie
van dit experimenteel systeem op basis van gebruik in de praktijk verschaft inzicht in
de sterke en zwakke kanten van het voorschrijvend conceptueel model (C). Op deze
wijze wordt een antwoord verkregen op onderzoeksvraag 2.

Op basis van literatuurstudie wordt geconcludeerd dat de pharmaceutische industrie
zich bevindt in een complexe, vijandige en turbulente omgeving. Een belangrijk aspekt
hierbij is de verhouding van de industrie tot de gezondheidsautoriteiten die
verantwoordelijk zijn voor de kosten van het medicijngebruik. Daarnaast blijkt dat deze
industrietak wordt gekenmerkt door een hoge mate van centralisatie, internationalisatie
en vertikale integratie. De diversificatiegraad binnen de industrietak is echter laag.

De kenmerken van pharmaceutische bedrijven zoals hierboven genoemd leiden tot de
konklusie dat het transferprijs beleid in deze bedrijfstak kan worden gekwalificeerd als
"internationaal transferprijs beleid". Dit internationaal transferprijs beleid wordt immers
bepaald op basis van regelgeving in diverse landen bij grensoverschrijdende transfers

167

Samenvatting

van goederen tussen verschillende delen van eenzelfde onderneming. Deze regelgeving
is met name afkomstig van fiskale en gezondheidsautoriteiten. Daarnaast spelen
financieringsaspekten een belangrijke rol. Transferprijs mutaties moeten dan ook worden
beschouwd als één - zij het slechts één - van de financieringsinstrumenten die de
onderneming ten dienste staan om geld binnen de onderneming te transfereren. Andere
instrumenten zijn dividenden betaald aan een houdstermaatschappij, interne royalty
betalingen, doorbelastingen van kosten en dergelijke. Tevens wordt gekonkludeerd dat
het reëel is om te veronderstellen dat de betrokkenen bij het vaststellen van
transferprijzen samenwerken in een "team" waarbinnen kennisuitwisseling centraal staat.
De betrokkenen zijn fiskale specialisten, financieringsspecialisten, de prijscoördinator en
tenslotte lokaal en centraal management. Door hen wordt naar consensus gezocht.

De analyse van de in de pharmaceutische industrie bestaande transferprijs situatie
zoals hierboven omschreven wordt bevestigd door een enquête die onder 12
transferprijs experts is gehouden. Deze twaalf experts zijn werkzaam in de Europese
pharmaceutische industrie.

Op basis van literatuurstudie op het gebied van beslissingsondersteuning worden bij het
ontwerp van een groeps-beslissingsondersteunend systeem de volgende vier aspekten
van belang geacht:
- een probleem-specificatie,
- een besluitvormingsstruktuur,
- een rolverdeling voor de betrokken personen,
- computerondersteuning.
Naast deze vier aspekten is het van belang dat het gebruik van het systeem niet
beperkt blijft tot computerondersteuning voor vergaderingen (de zogenaamde
groepssessies). De betrokkenen moeten het systeem zowel binnen als buiten deze
groepssessies kunnen gebruiken. Het systeem is dan ook niet ontworpen voor algemene
groeps-beslissingsondersteuning maar specifiek voor het toepassingsgebied transferprijs
bepaling. Hiermee hangt samen dat de groepssessies niet plaatsvinden in speciaal
hiervoor gebouwde ruimtes (zogenaamde "beslissingskamers") maar in één van de
vergaderzalen van de betrokken onderneming zelf. Zodra dit wenselijk wordt geacht
moet een groepssessie kunnen worden gehouden. De computerondersteuning tijdens
de groepssessie wordt geleverd met behulp van personal computers die ook buiten de
groepssessie voor beslissingsondersteuning worden gebruikt.

De vermelde analyse van de bestaande situatie, die leidt tot het beschrijvend empirisch
model (A), vindt plaats bij een grote Europese pharmaceutische onderneming (in deze
dissertatie als "Europhar" aangeduid). Uit het op deze analyse gebaseerde beschrijvend
conceptueel model (B) blijkt dat het groeps-besluitvormingsproces het best kan worden
omschreven door het begrip "negatieve selektie". "Negatieve selektie" houdt in dat in
principe de eerste oplossing waar de betrokkenen mee accoord gaan wordt gekozen.
Systematische generatie en evaluatie van alternatieven vindt niet plaats.

Het ontworpen systeem (het voorschrijvend conceptueel model (C)) wordt beschreven
aan de hand van de vier hierboven genoemde aspekten. De probleem-specificatie bakent
het terrein af tot een financieringsprobleem. Tevens wordt er een onderscheid gemaakt

168

Samenvatting

tussen de financieringsproblematiek op korte en lange termijn. De korte termijn situatie
wordt geanalyseerd op basis van een tot in detail gespecificeerde begroting. In de
besluitvormingsstruktuur wordt een expliciet onderscheid gemaakt tussen de taken
"voorbereidend ontwerp", "integrerend ontwerp" en "keuze". Er wordt dus een fasering
aangebracht. In de rolverdeling worden de volgende rollen onderscheiden: participant
in het besluitvormingsproces, autorisator van het besluitvormingsproces, analist en
"facilitator". De analist is degene die zowel in als buiten de groepssessies de computer
bedient. Hij is ook degene die de applikatie mee ontwikkelt. De facilitator struktureert
het besluitvormingsproces, zowel in als buiten de groepssessies. De
computerondersteuning bestaat uit drie databanken en drie rekenmodules. De databanken
bevatten gegevens over de regels omtrent financiering zoals die door diverse
autoriteiten zijn gesteld, kosteneffekten van de diverse financieringsalternatieven, en
referentieprijzen. Twee rekenmodellen zijn specifiek voor transferprijs- en
financieringsproblemen, en zijn bestemd voor de kwantitatieve "als-dan" analyse. Het
derde rekenmodel is bestemd voor het toepassen van multi-criteria analyse. Doelstelling
van dit rekenmodel is om effektieve communicatie tussen de leden van de groep te
bevorderen. Het wordt niet gebruikt om een "groepsmening" of "groepsgemiddelde" uit
te rekenen.

Op basis van het hierboven omschreven voorschrijvend conceptueel model (C) wordt
een experimenteel systeem (het voorschrijvend empirisch model (D)) gebouwd en
geïmplementeerd. Hiervoor wordt het softwarepakket "Personal Wizard" van Comshare
gebruikt. Het experimenteel systeem is bestemd om besluitvorming met betrekking tot
de Engelse dochteronderneming van Europhar te ondersteunen.

Na zes maanden experimenteel gebruik vinden de evaluatiegesprekken plaats. Hierbij
wordt het gebruik op basis van het beschrijvend empirisch model (A) met het gebruik
op basis van het voorschrijvend empirisch model (D) vergeleken. De zes participanten
en twee topmanagers nemen deel aan deze interviews. Het blijkt dat het voorschrijvend
empirisch model (D) niet alleen wordt gebruikt voor beslissingsondersteuning, doch
ook voor ondersteuning van onderhandelingen met een aan Europhar gelieerde
onderneming. Uit de evaluatie interviews komt naar voren dat zowel op financierings-
commercieel als fiskaal gebied de invloed van het experimenteel systeem positief is. Bij
de vraag naar de belangrijkste elementen van het systeem worden met name die
elementen genoemd die bijdragen tot een integratie van alle aspekten, zoals de
beslissingsstruktuur, de groepssessie en de rol van de facilitator. Daarnaast bieden
volgens de participanten de rekenmodellen tot dan toe ongekende mogelijkheden voor
de "als-dan" analyse.

Drie punten komen uit dit onderzoek naar voren:
- Computerondersteuning op zichzelf is geen garantie voor verbetering van de kwaliteit
van de besluitvorming. Andere aspekten zoals hierboven geformuleerd dienen in het
ontwerp te worden meegenomen.
- Beslissingsondersteuning voor groepen is in principe mogelijk met reeds in veel
organisaties aanwezige hard- en software. De relatief dure "beslissingskamers" zijn
dan niet noodzakelijk. Naast het kostenvoordeel biedt dit ook de mogelijkheid om
groeps-beslissingsondersteuning geleidelijk in organisaties te introduceren.

169

Samenvatting

- De multi-criteria techniek die onderdeel uitmaakt van de aangeboden
beslissingsondersteuning kan, op basis van literatuurstudie, als eenvoudig worden
bestempeld. De gebruikers vonden de techniek echter ingewikkeld. Dit duidt op
een grote discrepantie tussen theorie en praktijk.

170

Acknowledgements
First of all I would like to thank my promotor, Henk Sol. He is a stimulating
personality and knows the delicate balance that is required in academic research: on
the one hand he stimulates his Ph.D. students to independent thinking and developing
their own approach, on the other hand he encourages cooperation and requires that
the various research projects fit into larger scientific programs.

From the start of the project onwards, I have had frequent contact with most members
of the Ph.D. committee. I am grateful that they were willing to share their profound
expertise with me. This expertise ranged from knowledge-based systems, statistics and
methodology to multi-criteria techniques and transfer pricing. I enjoyed the discussions
and the provoking comments. These comments contributed significantly to the quality
of my research.

It will be clear that our research would have been impossible without the cooperation
of "Europhar". I would like to thank the top-managers, the participants, the analysts
and the joint steering committee members. In particular thanks for the hospitality and
friendship that I encountered while working in the United Kingdom.

During the years, technical assistance of Comshare was a great help. I experienced our
cooperation as an example of how universities and software houses can work together
effectively in the area of software development. Special thanks go to Bert Jaspers,
Henk Teunissen and Nely van Westerop.

I would like to thank all my colleagues at the Department of Information Systems in
Delft for the pleasant cooperation. The comments on earlier drafts of my thesis and
our discussions have been extremely valuable. I particularly like to mention Frans van
Schaik, Pieter Bots and the other members of the Kongsi.

Many students were involved in development of the system. I enjoyed their informal
and result-oriented approach. Especially thanks to Eric Dielessen. The technical
support that he provided at "Europhar" and during the various presentations was of
great value.

As the list on the next page shows, many institutions helped me to progress in my
research. Some institutions gave me valuable information, others provided financial and
practical support. For their help I would like to express my sincere thanks.

Last but not least I would like to thank Koos Schonewille and Jan van Boekel for
their friendship and practical assistance. They showed a commitment to quality and
service that is rare to find.

171

List of contributing institutions
- AKZO, pharmaceutical division, Oss, The Netherlands
- Association of the British Pharmaceutical industry, London, United Kingdom
- Bouman & Company, Den Bosch, The Netherlands
- Catacombestichting, Valkenburg, The Netherlands
- Ciba Geigy, pharmaceutical division, Basel, Switzerland
- Comshare, Delft, The Netherlands
- Decision Analysis Unit, London, United Kingdom
- Delft University of Technology, Delft, The Netherlands
- Duphar, Weesp, The Netherlands
- Euroforum, Eindhoven, The Netherlands
- Foundation for Medical and Pharmological Research, Utrecht, The Netherlands
- Gist Brocades, pharmaceutical division, Delft, The Netherlands
- Grasso's Koninklijke Machinefabrieken NV, Den Bosch, The Netherlands
- Hank Schut Management Consultants, Laren, The Netherlands
- Hoffman la Roche, pharmaceutical division, Basel, Switzerland
- International Medical Statistics, The Hague, The Netherlands
- International Bureau of Fiscal Documentation, Amsterdam, The Netherlands
- Management Forum, Guildford, United Kingdom
- National Facility of Informatics, The Hague, The Netherlands
- Nefarma, Utrecht, The Netherlands
- Office of Health Economics, London, United Kingdom
- Omnia Tangit Oeconomia, Tilburg, The Netherlands
- Stibbe, Blaisse & De Jong, Amsterdam, The Netherlands
- Tilburg University, Tilburg, The Netherlands
- Pharmaceutical Computer Club, London, United Kingdom
- Vidia, Zeist, The Netherlands

About the author
Michael van der Ven (1956) obtained his master's degree in business from Tilburg
University in the Netherlands in 1983 (cum laude). He specialized in information
systems and marketing. He developed an information system at Scott Bader Ltd, a
British chemical company. After six month's study at the Sorbonne in Paris he joined
one of the large European pharmaceutical companies in 1984. As a financial staff
member, he was responsible for the financial side of transfer pricing.

On the basis of a research agreement between Delft University of Technology and this
company, he started his Ph.D. research in the beginning of 1986. He presented his
research on scientific and management conferences in various European countries.
Michael van der Ven acted as a management consultant in the area of (transfer)
pricing and group decision support.

172

Index
accounting exchange rate 37
accounting module 100, 112, 113, 148
aggregation 15
analyst role 64, 150
Annual Financial Return 17
application domain 60, 64
arm's length 31
authorizer role 63, 150
average capital employed 20
bounded rationality 4
case study 6, 9
centralization 1, 13, 14, 21
choice phase 50
choice task 51, 54, 62, 93
communication technologies 45
comparable uncontrolled price method 31, 32
competitive organization 27
computer illiterate 59, 155
computerized support 61, 64, 81, 85, 98, 110
computing technologies 45-47
Comshare 69
conceptual model 5
consolidated profit 43
cooperative organization 27
coordinating decision 48, 51
cost structure 12, 14
cost-plus method 32
current account 33, 35
decentralization 1, 12, 24
decision network 57-59
decision room 57-61
decision rule 143-145
decision support generator 60
decision support system 2-4, 6
decision-group effectiveness 46
decision-group technologies 45-47
decision-making paradigm 3, 45, 47
decision-making task 49, 50, 53, 84
Department of Health and Social Security 16
descriptive conceptual model 5, 7, 81-85
descriptive empirical model 5, 7, 74, 78-81
design phase 50, 51
destination company 42
destination ratio 107
destination third 42
disaggregation 15
disaggregation module 106, 115, 116
diversification 12, 26

173

Index

dividend 33
divisional profit 25, 26
divisionalization 24
EFPIA 11
empirical model 5
Eurobitec negotiations 132
Eurobitec UK Ltd 70, 71
Eurochem 69, 71
European commission 15, 16
European pharmaceutical industry 11, 12
Europhar 7, 71
Europhar Australia Pty Ltd 71
Europhar International BV 34, 67, 71
Europhar Taiwan Ltd 71
Europhar UK Ltd 7, 34, 67, 71
ex ante 25, 39
ex post 25, 39
excess/shortfall calculation 20
expertise sharing 54, 56, 63
export third parties 72
external sale 42, 73
external selling price 42, 73
facilitator role 63, 150
financial-impact module 102, 113, 149
financing alternative 38
fiscal unit 69, 93, 133
generic sales 18, 72, 73
generics 16, 113
governmental interference 13
group decision support system 4, 57
group session 58, 119-132, 151
gut feeling 126, 129, 148
hardware 61, 69
hidden agendas 56, 63
human ethical drugs 11
human health care 13
hyper games 56, 63
ill-structured 2, 48, 57
implementation task 50
infrastructure 89, 155
innovation 47
integral decision making 51, 85, 100
integrative-design task 51
intelligence phase 49
internal reporting system 100
internal sale 42, 73
internal selling price 42, 73, 121, 125

174

international transfer pricing 29-38
internationalization 2, 13
intracompany loan 34, 112, 122
inventories 113
joint steering committee 68
key figures 100
local borrowing 37, 102, 114, 122, 129
local decision network 58, 59, 156
local lending 79, 102, 103
logging facility 102, 116, 138
lotus 1-2-3 69, 74, 75
management information system 28, 69, 100
market structure 12
Maud 54
memo 58, 74
minimum frequency of use 59, 154
multi-criteria decision making 54
multi-criteria technique 55, 62, 94, 99, 103, 115, 148, 152, 153
multinational 1
myriad of problems 81, 91
negotiation support 56, 132, 156
NHS-profit calculation 17, 18
normalized weight 55, 127, 130
null hypothesis 145
object company 40, 73, 105-107
organizational decision making 4, 6, 43, 49
overall evaluation 147
parent company 14, 41, 71
partial decision 79, 85
participant role 63, 88
path 51
Personal Wizard 111, 140, 151, 153, 155
Pharmaceutical Price Regulation Scheme 17
pilot system 6, 9, 117
post-industrial 3, 4, 45-47
practicality gap 54, 153
preference ordering 88, 93-101, 125, 129, 150
preference-ordering module 102, 113, 114, 148
preparatory-design task 50, 51, 84
preparatory meeting 117, 119, 121
prescriptive conceptual model 5, 61, 87
prescriptive empirical model 6, 87, 108, 137
problem solving 48-50, 52, 53, 82-84, 89, 90, 92, 97, 99, B l
problem specification 61, 78, 81, 88, 108, 151
profit and loss iteration 123-128
profit center 27, 28
project management 46

Index

project-management technologies 47
public screen 46, 57, 61
questionnaire 141
reference database 105, 115, 149
reference price 42, 73, 105
regulation database 102, 112, 147
relocation alternative 38, 78
remote decision network 58
reporting system 100
resale-price method 32
research approach 4, 87
research questions 1, 4
return on capital 17, 20
scan task 49, 83, 90, 108
scenario analysis 100
score 54-56, 99, 103, 115, 126
sensitivity analysis 100
share capital 33
simple and sufficient 64, 65, 111, 151
software house 68, 111, 116
software modules 112, 140
source company 41, 42
subgroup meeting 79
subsidiary company 14, 15, 33, 67
support staff 59, 63, 85, 100, 110, 150
system impact 138
System W 69
task structure 51, 62, 91, 100, 154
team setting 11, 25, 39, 6, 63, 85, 100, 140, 151
teleconferencing facilities 58
therapeutic area 13
three-tier demand system 15
total adjusted profit 17, 18, 20
transfer price 1, 23
turbulent environment 21, 45
user interface 60, 101, 111
version 7, 111, 156
vertical integration 13, 26, 41
weight 55, 99, 103, 115, 125-127, 130
well-structured 2, 79
what-if 60, 81, 85, 100, 148, 152
worksheet 103, 105, 126

176

Dissertation transfer pricing

More Related Content

Similar to Dissertation transfer pricing (20)

Recently uploaded (20)

Dissertation transfer pricing