Financial Programming and Oil Dynamics

FINANCIAL PROGRAMMING AND OIL DYNAMICS
Helios Padilla Mayer1
July 2007
ABSTRACT
Despite credit market turbulence and slowing activity in many major advanced economies, oil
prices have been reaching record highs in recent months. Besides oil-specific factors, such as
geopolitical risks and speculations, the current price boom is driven by demand and supply
forces that reinforce each other amid supportive financial conditions. This paper aims to a link
macroeconomic variables together with oil prices in order to provide complement decision
tools used by commercial and investment banks when optimizing their investment portfolios.
For that reason, we apply financial programming model with incorporated oil price variable.
We show that oil prices affect private consumption, gross domestic product, inflation, and
imports. On the other hand, we also investigate effects of macroeconomic variables on oil
market equilibrium. A decrease in oil supply as well as depreciation of the US$ lead to higher
oil prices, which in turn decrease private consumption and output, but as well stimulate
inflationary pressures. Empirical test is performed on the basis of quarterly US data from
2001 to 2007. Although financial programming models are subject to limitations and
empirical implications are difficult to apply, some general relations between selected
macroeconomic variables and oil price can be determined.
Keywords: Oil price, oil demand, oil supply, macroeconomic variables, financial
programming, error correction model, equilibrium
1
Senior Commodity Strategist, BNP-Paribas Fortis Bank.

1 Introduction
The continued volatility in the oil market is a matter of much concern. This has
witnessed large fluctuations in prices and a general underlying sense of uncertainty,
which is unsettling for normal day-to-day activities, as well as hampering sound
investment planning for the future. However, the recent high oil prices do not appear
to have harmed world economic growth, which has averaged a robust 4.9 per cent
over the past five years.
A combination of factors has contributed to the present market situation. Most
recently, these include the weakness of the US dollar, increased financial market
speculation, ongoing geopolitical developments and perceived market tightness and
bottlenecks in the refinery sector. A key concern is that crude oil prices have become
detached from the dynamics of supply and demand fundamentals.
The heightened levels of speculation have been a principal driving force behind the
volatility. Oil has become a financial asset, as other commodities. Large amounts of
money have been flowing into the commodities markets from other asset classes to
balance portfolio risks and seek higher returns. Notably, increased concern about the
falling value of the US dollar has encouraged inflows of new money into the crude oil
futures market.
Crude oil markets have been subject to shocks and consequently have been highly
volatile. Demand and supply shocks cause large movements in oil prices, which are
followed by a dynamic response in both energy demand and supply and in the energy
exploration and development activities. Modelling crude oil markets is of paramount
importance, not only because of the influence of energy price on macroeconomic
activity but also because of the role of energy in the investment plans of households
and firms. Energy cost and efficiency have become a prime concern in these plans.
There is a major difference between previous (for instance, in 1990) and current shock
prices. Due to increased energy needs from the emerging countries (mostly China and
India), demand side is mostly responsible for recent shocks. On the contrary, in the
previous years, shocks were generated by supply side. During the Gulf War, many oil
wells were destroyed which led to a shrinkage of oil supply.
Throughout this paper, a model is established to link macroeconomic variables
together with oil prices in order to complement existing policy tools used by
commercial and investment banks. As a support for prudent investment decisions,
banks follow the evolution of oil price, especially in research and trading desks. Our
model is able to simulate the effect of oil prices on macroeconomic variables and vice
versa.
We apply so called Financial Programming, which provides the diagnosis of
macroeconomic performance and analysis of the effects of macroeconomic and

structural policies on the main variables of interest for investment strategists,
including output, prices, and the balance of payments.
Theoretical and empirical literature on financial programming is scarce. Some studies
were made on certain specific aspects of the analytical framework such as the growth
aspects of financial programming by Chand (1987) or the analysis of the debt
rescheduling by Lorie(1989).
Barth and Chadha (1989) developed a small macroeconomic model that linked fiscal,
monetary and exchange rate policies of a developing country to the major
macroeconomic variables that were of interest to the policy-makers. Even if this
simulation model was more complete and comprehensive, the financial programming
framework was not in the center of the analysis, and interest rates among other
important variables were not included.
Mikkelsen (1998) constructed a model for financial programming that provided a
range of standard IMF performance criteria and the accounts for the real, monetary,
public, and foreign exchange sectors of the economy. His model, based on annual
data, was an excellent starting point for the design of a consistent framework for the
financial programming exercise. But as his predecessor, he did not include interest
rates. Moreover, the private consumption demand, which is important for aggregate
demand, was constructed on the basis of private disposable income only.
In 2004, Fair (2004) introduced a macroeconometric model in his book "Estimating
How the Macro economy Works". However, due to the size and complexity of the
model it is very difficult to get an overview of the model’s functionality.
Previous models of financial programming did not integrate oil price in order to
explain macroeconomic relationships within the economy. This was partially due to
the fact that in the past decades, only short-term oil price shocks took place whereas
the price remained relatively stable. Furthermore, the effect of oil price on
macroeconomic data could not be measured by econometric relationships due to the
lack of data. However, in recent years we observe a progressive and very likely
sustainable growth in oil price and therefore we believe that adding the oil variable in
financial programming model will contribute to an overall understanding of economy
behavior.
We extend the model developed by Mikkelsen (1989) by adding an energy market
component. Numerous studies have analyzed developments of energy market, both on
the factors influencing oil demand and supply, as well as on the impact of crude oil
price on major macroeconomic variables. Krichene (2005) described oil demand as
function of oil price, nominal effective exchange rate and U.S. interest rate while
Dées and Kaufmann (2005) found an impact of oil prices and world economic activity
on oil demand. On the other hand, LeBlanc and Chinn (2004) estimated the effects of
oil price changes on inflation for the United-States while Scoffield (2008) studied the
relationship between consumer spending and the recent rise in oil price.

The International Energy Agency (2004) brought to light the vulnerability of oil-
importing countries to higher oil prices. Assuming a sustained US$ 10 per barrel
increase in oil price, they forecasted a downfall of the US gross domestic product by
0.3 percent. As for the OECD, they would loose 0.4 per cent of GDP in the first and
second year of higher prices. The IMF (2003) forecasted that such a sustained oil price
increase would lead to a transfer of about 0.25 per cent of GDP from oil importers to
oil exporters and a similar transfer of income from oil consumers to oil producers. In
Elekdag Lalonde, Laxton, Muir and Pesenti (2008) studied the international
transmission mechanism of shocks behind the evolution of oil prices through an
adjustment of the Global Economy Model (GEM) and found out significant wealth
transfers between the regions.
Moreover, in 2002, as Balke, Brown, Stephen, Yucel and Mine pointed out, through a
vector autoregressive model ,that U.S. gross domestic product seems to respond
asymmetrically to oil prices, that is rising oil prices appear to slow down U.S.
economic activity by more than decreasing oil prices stimulate it.
Finally, Schindler and Zittel (2007) show that the world is at the beginning of a
structural change of its economic system and that the way of dealing with energy
issues will have be modified fundamentally. As a result, the rules governing the energy
market might be unique during this transition period and we cannot just rely on past
experiences to analyze it. (46)
Our analysis will focus on demand side shocks for the period 2001 to 2007 as data
availability allow setting up econometric relationships. The model developed in the
paper is applicable to the United States market; this is one of the bigger oil consumers
as well as (still) one of the major world economies, despite the fact that the economy
has been recently put under pressure due to the sub-prime crisis and the weakness of
the dollar. The model could be applied to other markets by changing coefficients in
econometrics relationships. Indeed, we expect these coefficients to be different for
each economically homogenous area.
The rest of the paper is organized as follows. Chapter 2 describes the structural model
of the paper: it explains the purpose of financial programming and sets the theoretical
framework through the goods, money and foreign exchange markets. Furthermore, we
set the simultaneous equations model, which explains behaviour of energy markets. In
chapter 3, we apply US empirical data to study behavioural relations and the main
accounting identities for the markets mentioned above. Simulation experiments
(forecasts and shocks) are conducted in chapter 4, where interrelations between the
energy market and all four other markets are discussed. Concluding remarks and the
limitations of our model are presented chapter 5.

2 Methodology
Financial programming is a comprehensive set of policy measures designed to achieve
a given set of macro economic goals. The policies are designed to eliminate
disequilibrium between aggregate demand and supply, which typically manifests itself
in balance of payments problems, high inflation rates, and low or falling output.
Financial programs emphasize the importance of monetary, fiscal, and exchange rate
policies in controlling domestic demand and correcting balance of payments
disequilibrium. In other words, economic policies enable to reach macroeconomic
objectives such as satisfying the requirements for external and internal balance-
equilibrium in balance payments, maintaining a sustainable growth, and getting the
inflation rate under control among other things. However, financial programs also
incorporate the effects of other policy measures, most prominently those aimed at
increasing aggregate supply. Such measures should help minimize the losses in output
and employment during the adjustment period, while eventually leading to a balance of
payments position that is sustainable.
An integrated system of macro economic accounts covering national accounts, the
balance of payments, and the fiscal and monetary accounts provides the information
needed to assess the performance of the economy and the need for policy adjustment:
(1) national accounts, (2) fiscal accounts,2
(3) monetary accounts and (4) balance of
payments accounts.
The integrated system of macroeconomic accounts provides the information needed to
assess the performance of the economy. But now a question is rising: How to calibrate
the policy in order to apply it to a specific country in order to be applicable to analyst.
The accounting framework develop by (FP) must be complemented by the
specification of a set of behavioral relationships. These relations indicate the typical
response of some of the variables included in the accounting framework to changes in
other variables. The behavioral relationships together with the accounting identities
form a schematic quantitative representation, or "model," of the relevant economic
processes. This framework can be used to assess the changes in policy variables, that
is, variables that are under the authorities’ control, needed to achieve given policy
objectives (for variables such as inflation and the balance of payments, which are
endogenously determined. Consideration of all sectors of the economy is necessary to
allow the repercussions of any policy changes to be recorded consistently.
The structure of the model will be formulated in terms of the demand and supply goods
(goods market), money (money market) and foreign exchange (foreign exchange
market). Moreover, as we are also considering the influence of oil price, we add a
supplementary market: the energy market. The model has been designed to provide
explicit links between monetary and exchange rate policies and energy price
development strategies.
2
In our paper we assume that public sector expenditures and revenues are exogenously determined and
therefore public sector discussion is excluded from the paper.

2.1 The Goods Market
One of the main challenges for policy-makers is to create an environment for a
sustainable long-run economic growth. In the long run, the economy tends to operate
in full capacity, so the economic growth is determined by the supply side. Several
macroeconomic variables are related with a slow growth: high fiscal deficit, high and
volatile in inflation, high debt burdens volatile exchange rates, etc. An accurate
analysis in these aggregates, and consequently the design of appropriate policies,
requires accurate economic information and statistics to guide policymakers in their
pursuit of economic objectives and e orts to respond to changes in the economic
environment. Below we present the main macro economic aggregates and the basic
macroeconomic relationships in the real sector.
Q Gross Output Value of all goods and services produced in the economy
VA Value added Gross output – intermediate consumption
C Consumption Intermediate + final consumption
IC Intermediate Consumption Inputs into production
FC Final Consumption Goods and services imported and domestically produced
I Investment Gross investment - depreciation
D Depreciation Difference between net and gross investment
A Absorption Sum of consumption and gross investment
X Exports Exports of goods and non-factor services
M Imports Imports of goods and non-factor services
GDP Gross Domestic Products Sum of value added across all sectors of the economy
S Savings GNDI - C
CAB Current Account Balance X – M + Yf + TRf
GNDI Gross National Disposable Income Total income available to residents
)()( MXAMXICGDP
VAGDP
For the modeling purposes, we consider GDPQ . The output of our model is gross
domestic product. Inflation is a sustained rise in the general level of prices and can be
measured by the GDP deflator or the CPI. The GDP deflator (PQ) is computed as:
100*
GDPreal
GDPnominal
PQ
The national income entity is derived in the following way:
CABIS
A negative current account balance (current account deficit) occurs whenever a

country spends beyond its means or absorbs more than it produces. The gross national
disposable income is:
ACABGNDI
2.2 The Money Market
Monetary policy is closely related to fiscal and exchange rate policies. The central
bank, commercial banks and other financial institutions are considered to be monetary
institutions. Most of central banks act independently of the government because
governments may be tempted to instruct central banks to print money rather than raise
revenues. Thereby, several central banks have been made independent of politicians in
order to immunize monetary policy from political pressures.
The money demand equation can be written as follows:
),( YiL
P
M
d
The theory of liquidity preference postulates that the quantity of real money balances
demanded depends on the interest rate (i) and on the income (Y). The interest rate is the
opportunity cost of holding money. When the interest rate rises, people want to hold
less of their wealth in the form of money. In conclusion, the quantity of real money
balances demanded should be negatively related to the interest rate and positively
related to income.
As money supply is a policy variable decided by the FED, we assume that there is a
fixed supply of real money with respect to the interest rate:
P
M
P
M
s
The money market equilibrium is found at the intersection of money demand and
money supply, which also determines equilibrium interest rate. For a fixed money
demand curve, the money supply determines the interest rate, or a targeted interest rate
determines the required size of the money supply.
2.3 The Foreign Exchange Market
We examine an accounting framework and key concepts used in analyzing the balance
of payments and external debt in an open economy. Measuring and assessing the
external position of a country are essential steps in the economic policy making
process. When capital mobility is imperfect, shocks to the investment/saving balance
affect both current account and domestic interest rates. We will address statistical
statements on economic transactions and financial flows between a country and the rest
of the world to determine a country’s external position and need for adjustment.

Current account balance (CAB) is identical to the economy’s resource gap (identified
as the difference between economy’s saving (S) and investment (I). On the other hand,
it is defined as a sum of trade balance (net exports of goods) (TB), net factor income
(Yf),3
and net transfers from abroad (TRf). Furthermore, current account also reflects
the gap between income (GNDI) and absorption (A) in the economy.
AGNDITRYTBCABIS ff
The current account balance is always matched by net claims on the rest of the world:
the change in net foreign assets of non-bank entities or non-monetary financial flows
( FI) and the change in net foreign assets of the banking system or monetary financial
flows ( RES).
RESFICAB
The capital and financial account records an economy’s net foreign borrowing and
capital transfers. The non-monetary financial flows ( FI) are the sum of foreign
direct investment (FDI) and net foreign borrowing (NFB) and together with monetary
financing ( RES) equal the current account balance.
0RESNFBFDICAB
The financial flows associated with the CAB deficit involve reduction in the net
foreign position of the economy. Thus, it is important to see whether the government
can service the associated change in net foreign investment position (due to a CAB
deficit) without modifying its current economic policies – for example, by reducing
absorption – or by bringing about changes in interest or exchange rates. If there is no
spontaneous financial flow, the government must adjust its policies to seek private
investment, seek borrowing abroad, or draw down international reserves.
2.4 The Energy Market
There exist several methods to model oil price. We are going to concentrate on two
most common approaches: financial models and structural models. Financial models
are based on financial theory and concentrate on the relationship between spot and
futures prices while structural models assign a key role to variables explaining the
characteristics of the physical oil market.
2.4.1 Financial Models
3
Net factor income is composed of (1) net factor income from labor = Remittances from domestic
workers abroad minus those of foreign workers at home , and (2) net factor income from capital =
Interest receipts from domestic assets held abroad minus interest payments on foreign loans

Financial models forecasts can be used to determine the relationship between the spot
futures prices. There are two kinds of transactions in the commodities market:
• spot trading where delivery of the commodity takes place immediately or
with a minimum lag due to technical delays, and
• forward trading where an agreement is made between two parties to
exchange a given quantity of product at some fixed future date for a previously
fixed amount of money
It is the commodity price risk and its higher volatility that drove the development of
liquid derivative markets through the years as hedging activities became more and
more indispensable for many sectors of the economy. Thus, the main reason to engage
into a forward contract is to become independent of the unknown future price of a
volatile product such as a commodity. For the purpose of our analysis, we use 1-
month futures contracts as a proxy for crude oil price.
2.4.2 Structural Models
We apply a simultaneous equation model to forecast oil price. Simultaneous equation
models are a set of linear equations that involve same endogenous variables. The
simultaneity arises when one or more of the explanatory variables are jointly
determined with the dependent variable typically through an equilibrium mechanism
(here the oil production and oil price).
Let Qd denote the oil demand of the entire world and WTI denote the crude oil
nominal price (in US$ per barrel) on the NYMEX.
The oil demand function is
11*11 uzWTIQd (2.1)
where 1z is some observed variable affecting oil demand. This is typically a structural
equation. The coefficient 1 measures the change of oil demand with respect to
change in oil prices. Thus, 1 is interpreted as the oil demand elasticity. We expect
1 to be negative as higher prices cause a decrease in demand.
As oil price is not exogenously determined, we add a complementary structural
equation: the oil supply.
22*22 uzWTIQs (2.2)
Similarly to the oil demand equation, the coefficient 2 represents the elasticity of oil
supply. Economic theory predicts 2 to be positive: the goods supply is increasing
with the price.

As we apply error correction model, residuals 1u and 2u are assumed to be
uncorrelated, independently and identically distributed with a mean of zero and
standard error sigma and uncorrelated with the exogenous variables.
The equilibrium oil price (WTI) is determined by the intersection of supply and
demand, therefore
imposedQQQ sd (2.3)
We choose oil supply to represent an imposed quantity on the oil market.
3 Results
3.1 Data
We use data provided by the International Monetary Fund (IMF) for the period of 1992
to 2007 on a quarterly basis.4
All variables are seasonally adjusted by a moving
average filter to avoid problems that may arise when time series data is quarterly and
subject to a seasonal pattern. Some of the variables we use are taken in nominal values
and others in real values. The values in constant prices are taken with respect to a base
year, which is the first quarter of 2000.
Econometric relationships are given in the logarithmic form of real values. If identity
implies nominal values, all variables in the same equation must be taken nominally to
get a consistent model. Thus, real values are multiplied by the corresponding price
level.
X Real value of variable X
XV Nominal value of variable X
X Logarithmic form of variable X
To avoid spurious regression problem we use stationary time series. If at least one of
the explanatory variables, Xk ,in the regression equation displays a trend over time, it is
very likely that the dependent variable, Y, will also display a similar trend. In this case,
we are likely to obtain highly significant coefficients and high values for R2
even if
such results are completely spurious.
3.2 Error Correction Model Approach
For our analysis, we use error correction model (ECM) because (1) its disequilibrium
error term can be regarded as a stationary variable, (2) it is easy to fit the ECM into a
general to specific approach, and (3) it overcomes the problem of non-stationarity of
variables. ECM involves the first-differences of our variables and the short-run (or
4
For a detailed specification of data, refer to the author.

disequilibrium) relationship between these variables. Suppose that the equilibrium
relationship between two variables X and Y takes the form:
1
* tt XKY (3.1)
where K and 1 are constants.
Using lower-case letters to denote the logarithmic form of variables, the preceding
equation may be rewritten as:
tt xy *10 (3.2)
where )ln(0 K . Equation 3.2 holds only if Y and X are in equilibrium. But
unfortunately this will rarely be the case; and the extent of disequilibrium between the
two variables is given by tt xy *10 . What one usually observes is a
disequilibrium relationship involving lagged values of X and Y:
ttttt yxaxaay 11210 *** (3.3)
where t is the disturbance. For simplicity, we included first-order lags only. A
reparameterization of equation 3.3 yields:
ttttt xyxay )*(* 11011 (3.4)
where 1 , /)( 211 aa and /00 a .
The term in parentheses in equation 3.4 can be regarded as the disequilibrium error
from the period t − 1. Thus, the current change in Y depends on the change in X in
period t and on the extent of the disequilibrium error in period t−1. Equation 3.4 is a
first-order error correction model (ECM).
Throughout this paper, we use quarterly data; so we take into account up to four-order
lags, which is the most we can take considering the time period, under which we
estimate our equations.
3.3 The Goods Market
In this part we consider long-run equilibrium and we assume that aggregate supply
curve is vertical as economy operates in full capacity. This means that fiscal and
monetary policies can influence output in the short-term but not in the long-term.
3.3.1 Supply Side
Let us first consider the Cobb-Douglas form of the production function:
21
* EKzQL (3.5)
where z is total productivity of factors, QL is long-term output or GDP, K is the stock
of capital, and E is the labour input.

The labour input E is defined by:
POPUNE *)1( (3.6)
where UN is the level of unemployment and POP is the number of inhabitants in the
United States.
The stock of capital in period t is the sum of total investment during period t plus the
stock of capital in period t − 1 minus depreciation of this stock.
111 )1( ttt KITK (3.7)
Total investment can be divided into public investment (IG) and private investment
(IP).
IGIPIT (3.8)
Our modelling assumption requires that the number of equations equals to the number
of unknown (endogenous) variables. Therefore, we add some equations depending on
other exogenous variables and we define them as added equations:
PQQIGPIIG Q *** (3.9)
where IGQ is public investment as a share of GDP, PQ is the implicit GDP deflator
and PI is the price level of investment goods. As variables are expressed in constant
(real) prices, we add the implicit GDP deflator and the price level of investment
goods.
PI
PQQIG
IG
Q **
(3.10)
Rewriting the production function in logarithmic form and add a disturbance term
leads to:
1210 ekql (3.11)
where )ln(0 C , 1 and 2 are long-run output elasticities with respect to capital and
labour, respectively.
We cannot estimate this equation cannot be directly as considered time series exhibit
constant stochastic trend over time, which could result in spurious correlation
problems. To overcome this problem we use the error correction model (ECM) and
the equilibrium equation yields:
tekql *001.0*698.1*154.0133.2 (3.12)
where t is the deterministic trend. Cochrane (1988) provided evidence that real GDP
in the US follows a deterministic trend. As Figure 3.1 shows, our model is coherent
with the actual GDP values:

Figure 3.1: Actual (LQA) and forecasted (LQ1) logarithmic forms of GDP, 1996-
2008
3.3.2 Demand Side
Gross domestic product (GDP) is defined as
MXITCTQ (3.13)
Dividing consumption into private (CP) and public (CG) and adding investment in
inventories (IS) to total investments leads to the equation we use in our model:
0MXISITCGCPQ (3.14)
Public consumption and investment together with real exports are assumed to be
exogenously determined.
The Private Consumption Model
Private consumption depends on a variety of factors:

(1) Private sector disposable income YDP (private sector income minus taxes). As we
consider normal goods only, demand for consuming goods should increase with a rise
in income;
(2) Interest rate I. If interest rate increases, consumers will save more money in order
to purchase more consuming goods later and increase their utility. We use 3-month
Treasury bill interest rate and covert it in index in our model;
(3) Financial liabilities of households (hereafter referred to as debt) FL;
(4) Financial assets FA;
(5) Unemployment UN which is also converted in index;
The function of long-term private consumption is then defined by:5
4321
*** FLUNIYDPACP (3.15)
and the logarithmic form is written as
fluniydpcp **** 43210 (3.16)
In order to estimate this equation we use an error correction model (ECM) and results
are the following:
tfluniydpcp *009.0*621.0*015.0*003.0*189.0926.111 (3.17)
where t is deterministic trend.
Inflation
Long-term inflation is determined by the intersection of aggregate demand and long-
term aggregate supply curve. In order to model inflation, we use a modified version of
the Friedman-Lucas money surprise model:
)(* T
QQ (3.18)
where is the actual rate of inflation, *
is the expected inflation, is an increasing
function and T
Q is the output trend of the output in the long-term. This relationship
arises because real output only deviates from trend in the model if the central bank
increases money supply growth rate in a surprise way, causing a surprise increase in
inflation rate. Estimating equation 3.18 results in:
QL
Q
pqpq tt ln*424.01 (3.19)
Surprisingly the coefficient is negative and this theoretical relationship does not match
reality.
5 Financial assets (FA) variable was not included in the model as coefficients on the current and lagged variables
were not significant.

We also estimate price level of investment goods (PI) as a weighted average of price
level of domestic goods (P) and price level of the imported goods (PM) because
investments in specific countries use domestic as well as foreign goods.
0*)1(* 22 PMPPI (3.20)
A numerical evaluation leads to:
0*377.0*623.0 PMPPI (3.21)
3.3.3 Integration of Crude Oil Price Influence
Oil prices will have a direct impact on the GDP model and the private consumption
model only. Since other variables are linked together, oil prices will most probably
have an influence on the US economy as a whole.
The GDP Model
It has now become a usual conclusion of economic literature: increasing oil price and
volatility dampen economic growth. However we cannot insert a price in the Cobb
Douglas function, as it is not a production factor such as K or E. The oil price will
have an influence on the US output through private consumption and imports as we
prove in chapter 3.5.
The Private Consumption Model
Rising oil and gasoline prices are putting a dent on income which should in turn
influence private consumption; the WTI price must be taken into account in private
consumption model and it should be preceded by a negative coefficient. We can
divide private consumption into oil consumption and consumption of other goods.
Consider a two-period model and assume that households have the same disposable
income for both periods ; if oil price increases from period 1 to period 2, a bigger part
of their income will likely be spent in period 2 for the same quantity of oil they used
to purchase in period 1. Consumers have less money to spend on other goods. After
running an appropriate ECM, the simplified equation is now given by:
twti
fluniydpcp
*013.0*048.0
*318.0*034.0*002.0*280.0938.82
(3.22)
where t is deterministic trend. The upgrade of the original model is striking as shown
in Figure 3.2.
Figure 3.2: Actual log form of private consumption (LCPA), forecasted log forms of
private consumption without WTI (LCP1) and with WTI (LCP2)

Inflation
We extend the estimated inflation equation 3.19 by adding WTI prices as well as
interest rates, which measure the effect of monetary policy on inflation. The
equilibrium relationship yields:
iwti
QL
Q
pq *014.0*204.0ln*763.33187.4 (3.23)
Unlike in the evaluation of the Friedman-Lucas Money Surprise model, the
coefficient on the output gap (deviation of actual output from the output trend) is now
positive, as well as coefficient on the WTI price. As oil is inelastic product,
consumers do not seem to be very sensitive to changes in oil price. When firms
increase the price of their final products, wages follow this upward movement, which
leads to an increase in inflation rate. Nevertheless, repercussions of the rise in oil
price on inflation might be limited because of governments try to meet an inflation
target through monetary policies.
3.4 The Money Market
3.4.1 Supply Side
Supply of M1 is determined by:
HMM *1 (3.24)

where HM is reserve money of the central bank. The parameter μ is chosen by the
FED according to the economy situation and their price control policy. We can
observe the tendency of money multiplier μ to converge to the value 1.77.
HMM *77.11 (3.25)
3.4.2 Demand Side
Money demand is determined by the consumer’s behaviour. In real terms, M1
depends on:
(1) Real national disposable income YD. With an increasing income, households will
want to consume more goods and will therefore demand more money;
• Bonds nominal interest rate I. If interest rate is high, money is deposited on a bank
account or bonds are purchased;
• Inflation rate that approximates the cost of holding money. We use the implicit GDP
deflator PQ.
The form of the M1 function is:
321
)(***
1
PQIYDA
P
M V
(3.26)
We rewrite this equation in logarithmic form:
pqiydm ***1 3210 (3.27)
The simplified equilibrium ECM estimated equation yields:
tpqiydm *058.0*469.6*138.0*663.2411.491 (3.28)
where t is deterministic trend. As money market is always assumed to be in
equilibrium, the supply and demand for M1, as defined in equations, must be equal.
Higher oil prices lead to inflationary expectations and thus higher money demand. In
the observed period, WTI and GDP deflator (PQ) are highly collinear (the correlation
coefficient equals to 0.93) as well as the effect of WTI on M1 demand is absorbed by
inflation rate. Therefore, we do not have to (and we cannot) insert oil price variable
directly into the money demand model.

3.5 The Foreign Exchange Market
3.5.1 Import Demand Model
We use a standard import demand model with total income, price of domestic goods
and price of imported goods as explanatory variables. Goldstein and Khan (1985)
presented two trade models: the perfect substitutes model and the imperfect
substitutes model. While the first one is mainly used for the trade of homogenous
goods, the latter is applied to study trade of manufactured and aggregate goods. For
the purpose of our analysis, we will consider the second one. The main assumption
underlying the imperfect substitutes model is that imports and exports are not perfect
substitutes for domestically produced goods. Considering this framework, the basic
demand for imports model is as follows:
tttt PMPYDM ln*ln*ln*ln 3210 (3.29)
where Mt, demand for real imports, is a function of national income (YDt), prices of
domestic goods and non-factor services or cross prices (Pt) and prices of imports or
own prices (PMt). In this section, we follow the recent formulations by Tang (2003),
Ho (2004) and Narayan (2005). We divide national income (YDt) into its final demand
expenditure components (CONSt + ITt + Xt) and specify a computable disaggregate
import demand model as follows:
tttttt pmpxitconsm ***** 321312110 (3.30)
Estimating the above equation through an ECM results in:
tpmp
xitconsm
*015.0*783.0*975.1
*315.0*439.0*005.2100.211
(3.31)
where t is deterministic trend.
As expected, imports increase with higher level of domestic consumption, investment
and exports (most of time imported goods are used as inputs for production of final
export goods), as well as higher level of price of domestic goods. However, imports
decrease with an increase own prices.
Increasing oil price will have an impact on the import model and therefore we add the
WTI variable to the econometric relationship. The estimated equation is as follows:
twtipmp
xitconsm
*005.0060.0*690.0*910.0
*290.0*520.0*305.1440.112
(3.32)

As in the case of private consumption, Figure 3.3 shows that the WTI-adjusted model
fits better actual import values
Figure 3.3: Actual log form of imports (LMA), forecasted log forms of imports without
WTI (LM1) and with WTI (LM2)
3.6 The Energy Market
3.6.1 Oil Demand Equation
The total demand of oil is an aggregate value of the OECD and non-OECD countries’
demand for oil. In our analysis we assume that oil demand equals oil supply as we use
crude oil production as endogenous variable.
We try to identify other variables than oil prices that affect oil demand. Krichene
(2005) described the oil demand as function of oil price, nominal effective exchange
rate and U.S interest rate while Dées and Kaufmann (2005) found an impact of oil
prices and world economic activity on oil demand.
The world GDP is computed as an index and is expected to affect oil demand
positively. Indeed, better economic situation will stimulate higher demand for all
kinds of goods and services including crude oil. As Cooke (2006) observed, a rise in
oil consumption is stimulated by an increase in the demand for goods and services
that results in greater production (and hence GDP). On the other hand, when

consumption declines, so does GDP. Thereby the increase or decrease in GDP (which
measures the production of goods and services), tends to drive the demand and
consumption of oil.
Secondly, we are interested to observe impact of monetary policies on oil prices. In
particular, we want to test how shocks on interest and exchange rates affect oil price
developments. NEER, the nominal effective exchange rate, is computed as a weighted
average value of a country’s currency relative to all major currencies being traded
within an index of currencies. The weights are determined by the importance a home
country places on all other currencies traded within the index, as measured by the
balance of trade. A higher NEER coefficient (above 1) means that the home country’s
currency will usually be worth more than an imported currency (an appreciation of
domestic currency), and a lower coefficient (below 1) means that the home currency
will usually be worth less than the imported currency (a depreciation of domestic
currency). To model the oil demand, U.S NEER is used since the dollar is still used as
the medium of exchange for most of oil transactions. As WTI crude oil prices are
denominated in US$, depreciation of US$ makes oil imports cheaper in non-US$
currencies. Therefore, we expect a negative sign of NEER coefficient on oil demand.
Alhajji (2004) analyzed the impact of a US$ devaluation on the world oil industry.
US$ devaluation increases demand for oil in countries with non-US$ appreciating
currencies. It also increases demand for gasoline in the US as thousands of Americans
spend their vacations at home instead of travelling to Europe where the cost of the
vacation is at least 40 percent higher than three years ago. Regardless the OPEC
decisions, US$ devaluation on its own may tighten supply, increase demand and keep
oil prices high for an extended period of time.
Thirdly, we want to see whether nominal interest rates could serve as important
explanatory variable in the oil demand equation. We examine 3-month T-bill and 6-
month LIBOR rates as a trade-off between short- and long-term interest rates. We
expect that the interest rate coefficient will have a negative sign in oil demand
equation as worldwide low interest rates stimulate aggregate demand and therefore
energy demand.
The simplified ECM estimated equation for oil demand is given by
tneerGDPwtiq worldoil *030.0*073.0*837.1*066.0232.14 (3.33)
There was no correlation between the oil demand and the interest rates. In order not to
reduce degrees of freedom in the oil demand equation, we decide to omit interest rates
Variable as coefficient proved to be non-significant. One possible explanation could
be that oil demand is relatively inelastic so interest rates do not have such significant
effect as for other consumption goods consumption.
As expected, oil demand reacts inversely to oil price, positively to economic activity
and negatively to the exchange rate. All coefficients can also be interpreted as
elasticities. Low price elasticity implies that a big shock in oil price will have a small
impact on the quantity demanded. For most of goods, when the demand is inelastic,
the consumer buys substitute products but in the case of oil there are no substitute

products (yet) and therefore it is unlikely that in the current environment of increasing
oil prices oil demand will be significantly reduced.
The relationship between oil demand and GDP is usually scrutinized within the
context of income elasticity of demand. High-income elasticity of demand means that
if country is progressing, investment and consumption activities are booming and
therefore, there is a greater demand for oil. On the other hand, if an economy enters in
a recession, a slow-down in economic activities will cause a decrease in demand for
oil.
In order to check for biasness of coefficients, we analyze the correlation matrix and
find out that explanatory variables are slightly correlated.
WTI GDPworld NEER
WTI
GDPworld
NEER
1
0.17
-0.21
0.17
1
-0.51
-0.21
-0.51
1
The biasness can be reduced by either (1) extending a data coverage period as this will
lead to a lower correlation between explanatory variables or by (2) retrieving one of
the explanatory variables. The first proposal is not a feasible option as demand side
shock started in 2001 and we want to analyze the behaviour of oil markets after in this
period. We try to retrieve one of the explanatory variables but it leads to a mis-
specified regression. We therefore decide to leave demand equation unchanged keep
in mind that our estimators are probably biased.
3.6.2 Oil Supply Equation
The total oil production is the aggregate value of the OPEC and non-OPEC oil
producing countries. Supply is harder to model than demand because it is difficult to
understand the OPEC production quota decisions and there are many exogenous
factors that influence oil supply.
As higher oil prices stimulate production in theory, we expect to observe a positive
coefficient on oil prices as explanatory variable in the oil supply equation. The hardest
task now is to identify other variables that affect the oil production.
One possible explanatory variable is oil reserves (OR). Higher level of reserves
indicates possible higher production capacity, however, one should be aware that
existing reserves are placed in more difficult accessible areas so higher extraction and
refinery costs are related to greater oil supply. Nevertheless, positive correlation is
expected between the oil reserves and oil production.
Oil inventories are simply the stocks of oil already extracted and will probably have a
negative impact on oil production. The refining capacity is related to the changes in
the downstream sector. Refinery capacity utilization is measured as the ratio of the
total amount of crude oil through distillation units to the operable capacity of these

units. Dées and Kaufmann (2005) suggest that a lack of spare refining capacity is one
cause of the increase in oil price. Likewise the lack of sufficient spare production
capacity could explain the sharp rise in oil price.
As for the demand, the simplified econometric equation is given by:
orwtiqoil *433.0*031.0180.1 (3.34)
Only oil price and the oil reserves seem to impact oil production. Their coefficients
have expected sign but are not really significant. Increase in the oil price will
stimulate a rise in the quantity supplied. As oil resources are limited, one cannot
expect an indefinite increase in supply along with higher oil prices. Secondly,
constant increase in production will lead to excess supply, increase in inventories and
that would lead to a decrease in oil prices. Positive coefficient of oil reserves confirms
the ability to produce more if more reserves are discovered.
As in case of oil demand, oil supply is fairly inelastic with respect to price changes.
Higher prices do not bring forth new capacity because suppliers are unwilling or
unable to increase production.
In order to check for the biasness of coefficients, we compute the correlation matrix:
WTI OR
WTI
OR
1
-016
-0.16
1
The correlation between oil price and oil reserves is low enough to believe that
explanatory variables are not correlated. Therefore, we expect that our estimators are
unbiased.
The supply equation could be improved if we considered the production of OPEC and
non-OPEC economies as two separate variables. The non-OPEC nations act as price
takers whereas OPEC producers use a myriad of factors to determine levels of
production and installed capacity. Thus, OPEC production would be determined by
the total demand of oil, the production of non-OPEC countries, the oil stocks, and
geopolitical factors. However, modelling of OPEC production exceeds the format of
the paper, so we do not modify our oil supply equation.
So far, our focus has been made on fundamentals analysis. Extraordinary
circumstances in form of supply disruptions may at any time result in a shock to
economies and oil price. Such circumstances may include: attacks on Saudi oil
installations and worldwide supply lines, disruption of the already reduced Iraqi
production; Iran stopping its deliveries due to an attack related to its nuclear
development programs; Venezuela deciding not to deliver oil any longer to the US
and divert it to China instead, fighting in Sudan and other African producer countries;
and OPEC to admitting that their reserves have been overstated. Furthermore,
speculations in trading activities are becoming more important determinant of oil
price developments.
Most of these events could be incorporated in the equation as a dummy variable. The
equation could be extended by five dummy variables: geopolitical factors, supply

disruptions, environmental disasters, technological breakthroughs, and speculations.
However, as we are not sure whether simple dummy variables and estimated
coefficient would really reflect the market situation, we do not include them in the
analysis.
4 Simulation Results
In order to run the model, we use the numerical computing environment Matlab.
However, prior to that, we must make certain assumptions about oil related variables.
In the Energy Market section, we use the oil demand and the oil supply equations to
explain the movements of oil prices. As expected, these two econometric relationships
do not yield the same quantity of oil (demanded and supplied). As a matter of fact, we
did not account for the stocks of oil barrels for which the data are not easily given out.
We can solve this inconsistency by (1) setting a third equation describing the oil price
as a function of the gap between oil supply and oil demand; or (2) considering the oil
supply as exogenous and to keep only one equation: this assumption will allow us to
make shocks on oil supply. We decide for the second option as too many random
factors affect oil supply and our econometric relationship for oil supply does not
really fit the actual data.
The final objective of this paper is to provide a tool that would allow detect how
changes in selected macroeconomic variables and conditions on oil markets affect
each other. Through the financial programming model, we manage to equilibrate
conditions on goods, money and external sector markets while taking into account
movements in the oil market. This framework can be then used for (1) forecasting
purposes of selected variables, and (2) detecting the size of the effect of change of
exogenous variable on selected endogenous variables in the model. If the model is fit
properly to indicate correct movements of selected variables, it can serve as additional
tool for bankers while undertaking investment decisions. In order to meet this
objective we decide to focus on six major endogenous variables:
(1) Oil price (WTI). Given the constant increase in recent years, oil prices are
currently by far the most analyzed variable.
(2) Current account balance (CAB). Through their fiscal and monetary policy tools,
governments try to reach a balanced CAB.
(3) Private Consumption (CP). As CP is a measure of purchasing power, an important
indicator of the country’s wealth.
(4) Gross domestic product (GDP). This is the major indicator of the level of
economic activity in the country.
(5) GDP deflator (PQ). Most governments concentrate on inflation targeting as part of
their stabilization programs.
(6) Money demand (M1). Along with inflation targeting, monetary policy decision-
makers aim to control money supply in the countries.

4.1 Forecasting
Our forecast starts in the fourth quarter of 2006 because some time series are no
longer available for next quarters. Below we present different assumptions made on
the missing exogenous variables.
(1) Investment in inventories that represents goods produced during a given period but
not sold during this period (IS). Due to its nature this variable is not easily predictable.
It can be negative as well as positive. Indeed, goods might have been produced in an
earlier period and sold in the current period. So we decided to take the same values
than for the three previous quarters considering the economical conditions were
similar.
(2) Nominal Effective Exchange Rate (NEER) is computed as a benchmark of U.S
dollar with respect to other currencies. Since 2005, the dollar is constantly
depreciating with respect to the Euro. We predict than a decrease in NEER in 2007
with the same rate than it was in 2006.
(3) Unemployment (UN): the employment is rather stable so we keep the same value
than in 2006:4.
(4) Government Revenues (REV): Time series data indicate that government revenues
jump by 30% in the second quarter of the year. We assume the same increase for our
forecasted value. After this jump, the government revenues are decreasing but stay
around 10% above the value before the jump.
Variable Units 2006:4 2007:1 2007:2 2007:3
IS US$ billion 0.08 0 -0.08 0.08
NEER Index 107.47 107.16 104.61 102.68
UN % 4.00 4.00 4.00 4.00
REV US$ billion 488.11 461.13 600.38 510
The degree of accuracy of the model reduces as we go forward in time due. The 2008
forecasts for the year would be quite inaccurate due to three types of errors:
(1) Errors due to the forecast of exogenous variables not available in 2008;
(2) Errors due to the predefined econometric relationship transmit to the period from
2006:4 to 2008:4;
(3) Errors due to the resolution of the system by Matlab
We could diminish these errors by wedging our model but this exceeds the context of
the paper. Our forecast runs until the third quarter of 2007. The results are given in the
table below:

Variable Unit 2006:4 2007:1 2007:2 2007:3 Variation(2006:
4 to 2007:3)
WTI US$/barrel 60.41 57.41 64.24 71.00 17%
CAB US$ billion
% GDP
-116.33
-0.012
-89.49
-0.008
-127.27
-0.011
-161.53
-0.015
-45.2
39%
CP US$ billion 10504 10535 10321 10558 0. 5%
GDP US$ billion 10992 10996 10921 11049 0.5%
PQ Index 117.17 117.32 117.55 117.87 0.6%
M1 US$ billion
% GDP
1187.80
0.108
1147.80
0.104
1136.60
0.104
1119.90
0.101
-67.9
-6%
4.2 Shocks
We study the impact of shocks imposed on exogenous variables on the key major
endogenous variables described above. We choose three exogenous variables: (1) Oil
supply because it has a major influence on oil price, (2) the interest rate and (3) the
nominal effective exchange rate because they are two key variables for fiscal and
monetary policies.
In reality, two kinds of shocks can occur:
(1) A sustained shock, growing by a certain percentage at a base day and keeping the
same growth rate for the next days;
(2) A one shot shock that is just a shock happening at a base day and in the next days
it comes back to its actual value.
Due to dynamics of our model, analyzing the one-shot or the sustained shock will not
make a difference because the variables reach their equilibrium values at the first
quarter. As the forecasts above start in the fourth quarter of 2006 the shock will be
done at this precise point.
Since we are interested to analyze the impact of these sustained shocks on endogenous
variables, we choose values that could represent real situation:
(1) Oil supply decreased by 1% in 2006: 4;
(2) Real interest rate decreased by 1% in 2006: 4 - shock of 22.6% on the interest rate
index;
(3) Nominal effective exchange rate decreased (depreciated) by 5% in 2006: 4.

4.2.1 Decrease in Oil Supply
A 1% downward6
shock on oil supply is assumed.
Figure 4.1: Effect on the level of oil price
Figure 4.2: Effect on the percentage change in oil price
6
We do not graph the upward shock because it is perfectly symmetric with the downward shock.

Since we assume same shocks for all variables - a jump in 2006: 4 and then growing
at the same rate -, we create a table where we report percentage changes of forecasted
values with respect to a decrease in oil supply:
Variable 2006:3 2006:4 2007:1 2007:2 2007:3
WTI 0 15 15 15 15
CP 0 -0.91 -0.91 -0.91 -0.91
GDP 0 -0.63 -0.63 -0.63 -0.63
PQ 0 0.75 0.73 0.71 0.72
M1 0 3.18 3.04 2.91 3.03
CAB 0 0 0 0 0
The only variable directly affected by the oil supply shock is oil price. The percentage
variation of oil price due to a 1% downward shock push up the price by 16%. Figure
4.1 shows that a 1% decrease in the supply increases oil price to US$ 69.6 per barrel
from the predetermined level of US$ 60 per barrel.
An impact of oil production on prices explains pressures exerted by governments on
OPEC to increase oil production. Instability in the Middle East such as the disastrous
effects of a war in Iran are concerning. Our conclusions are supported by the
following fact: while the value of barrel increased by 161% during the war between
Iran and Iraq in 1979-1980, the supply had decreased by only 5%.
Other variables affected by oil supply shock are actually affected by oil price. Private
consumption, which can be divided into the oil consumption and consumption of
other goods, is expected to be negatively correlated with oil price. Considering the
same disposable income for this period, a bigger part of this income will be likely
spent for the same quantity of oil and less will be available for spending on other
goods. Scherer (2008) showed that in 2008, 6.5 percent of household spending were
dedicated to energy compared to 5.8 percent in 2007 and less than 4.1 percent in
2002. The table above shows that a 1 % point decrease in oil supply will decrease
private consumption by 91 basis points.
The gross domestic product decreases by 63 basis points with a 1 % point decrease in
oil supply. The effect of imports partially counterbalances a decrease in private
consumption.
Inflation is fundamental in financial programming. A decrease in oil production as
well as an increase in oil price is going to raise the goods prices. Brown and Cronin
(2007) as well as Roubini (2004) analyzed the effect of commodities prices on
inflation. With an increase of oil price, firms will increase their final products prices,
as oil is a major input product in most of industries. Wages will adjust to this increase,
which will lead to a general surge in price level. The table implies that a drop in
supply will increase inflation by 75 basis points. Thus, we can conclude the oil price
has a greater impact on inflation than GDP.

Money demand is directly linked to the inflation. Higher inflationary pressures will
create greater demand for money as consumer need to spend more money for the
same quantity of goods. A 1% downward shock will lead to a rise in money
demanded by 3%.
4.2.2 Decrease in Interest Rate
We assume a 1% downward shock on the interest rate that represents a 22.6% shock
on the interest rate index.
Variable 2006:3 2006:4 2007:1 2007:2 2007:3
WTI 0 0 0 0 0
CP 0 0.06 0.06 0.06 0.06
GDP 0 0.03 0.03 0.03 0.03
PQ 0 0.48 0.46 0.53 0.52
M1 0 6.91 6.63 7.63 7.41
CAB 0 0.01 0.01 0.01 0.01
In our model, interest rate has no effect oil price as oil is explained by three
exogenous variables and the interest rate is not a part of it. However, in reality, we
could expect the interest rate affecting the oil price through the exchange rate. Indeed,
MacDonald and Nagayam (2000) noticed a long-run relationship between the real
interest rate and the real exchange rate.
One of the determinants of the propensity to consume current income is the real
interest rate. Sorensen and Whitta-Jacobsen (2005) analyze different channels through
which interest rate could affect the private consumption. If interest rate is growing,

consumers will substitute their present consumption to the future and increase current
savings. This is called the substitution effect and leads to a reduction of private
consumption for the current period. But this higher interest rate will also increase the
amount of future consumption. Hence the consumer can afford a higher level of
current consumption without having to reduce future consumption so she tends to
increase private consumption for the current period. In addition, higher interest rate
implies that the future labour income is discounted more heavily so the value of
human wealth decreases. High interest rate makes future income easier to attain by
saving out more current income and consuming less. In conclusion, the interest rate is
negatively correlated with the private consumption. In the table, we can notice that a
1% decrease in the interest rate increases the private consumption by 6 basis points.
Gross domestic product exhibits the same reaction on changes in interest rate as
private consumption. High interest rate will decrease consumption and investment
(considered exogenous in this model) in order to increase savings. It is therefore
surprising that table shows only a 3 basis point change in gross domestic product with
respect to a 1% shock in interest rate.
Alvarez, Lucas et al (2001) analyze relationship between interest rate and inflation.
The Fed meets eight times a year to set short-term interest rate targets in order to
target inflation. As interest rates drop, consumer spending increases, and this in turn
stimulates economic growth. But if consumers’ demand for goods increases faster
than the supply, it will lead to a general increase of prices. The downward shock fits
the theory: a 1% decrease in interest rate will lead to a 48 basis points increase in
price level.
The money demanded reacts negatively on an increase in interest rate as consumers
reduce their spending thus need less money. Inflation is also negatively correlated
with interest rate and affects money demand. A 1 % downward shock on interest rate
implies a 7% increase in money demanded.
4.2.3 Depreciation of Nominal Effective Exchange Rate
We assume a 5% downward shock on exchange rate, which actually represents
depreciation of the US$ with respect to other traded currencies.

Figure 4.4: Effect on the percentage change in oil price
Variable 2006:3 2006:4 2007:1 2007:2 2007:3
WTI 0 5.6 5.6 5.6 5.6
CP 0 -0.33 -0.33 -0.33 -0.33
GDP 0 -0.23 -0.23 -0.23 -0.23
PQ 0 0.28 0.27 0.26 0.27
M1 0 1.18 1.12 1.08 1.12
CAB 0 0 0 0 0
Alhajji (2004) analyzed the impact of a dollar devaluation on the world oil industry.
Dollar devaluation increases demand for oil in countries with non-dollar appreciating

currencies. Regardless of OPEC decisions, dollar devaluation on its own may tighten
supplies, increase demand and keep oil prices high for an extended period of time.
A 5% downward shock on exchange rate increases oil price by 5.6% or to around US$
62.5 per barrel from the US$ 60 per barrel benchmark.
As exchange rate is assumed to directly impact only oil price in our model, it will
affect the other macroeconomic variables through oil price. The impact will be similar
as in the case of the oil supply shock. For instance, private consumption decreases by
30 basis points with a 5% decrease in exchange rate of 5%. This can be explained by
the fact that a bigger part of the income is dedicated to energy consumption.
The devaluation of the US dollar slows down domestic economic activity as
inhabitants consume less. However, on the other hand, domestic economic activity
will be pushed up due to the decrease in imports since the goods valued in other
currency are more expensive for the US inhabitants. GDP decreases by around 23
basis points when the NEER benchmark is depreciated by 5%. Thus, there is a greater
impact on private consumption than on imports.
GDP deflator increases by 28 basis points when exchange rate depreciates by 5%. The
theory confirms our empirical results. Gosh, Ostry et al (1996) studied the relationship
between exchange rate and inflation. If the U.S currency becomes cheaper with
respect to other currencies, the foreign countries will be tempted to buy US goods.
Hence, world demand for the US goods increases, which leads to an increase in the
general level of prices. Similarly, country with high inflation rate is expected to see
the value of its currency with respect to the other ones deteriorate. Thus a negative
correlation is expected between exchange rate and inflation.
As for the two previous shocks, the demand of money is directly affected by inflation.
A 5% decrease in the exchange rate will push up the money demand by 1.2%.
Depreciation of the US currency indicates that consumers need more money to buy
the same quantity of goods and services.
5 Conclusion
The IMF uses its well-known "financial programming" model to derive monetary and
fiscal programs in order to achieve desired macroeconomic targets in countries
undergoing crises or receiving debt relief, mainly developing countries.
We agree with the idea of Agenor and Montiel (1999) that financial programming,
even if applied frequently in policy formulation in such developing nations, is subject
to limitations that must be considered in the policy guidance. Financial programming
models are rather theoretical and empirical implications are very difficult to link with
the real world performance of economic variables.
Mussa and Savastano (1999) emphasize that the instability of behavioural parameters
is a serious concern. As we work with the logarithmic form of these variables, the
instability can lead to even more riskier consequences. However, instability impact is
limited in the model as we work with the US data, which are more accessible and

reliable than the ones of developing countries. All parameters are estimated by formal
econometric techniques (i.e. the error correction model). On the contrary, in
developing countries, the estimation of the parameters is based on rough statistical
applications and results are questionable due to unstable relationships and unreliable
data.
In addition, we list some other restrictions of the model presented here:
(1) We notice the size of errors in the econometric relationships. The level of accuracy
decreases with the increase of the length of the forecast period. Wedging could be
used to correct this effect;
(2) As an effort towards simplification, the decision to consider variables as
exogenous while they obviously contain "endogenous" characteristics could be
optimized as well;
(3) The behavioural equation for money demand, which is critical for fiscal and
monetary policies, should include more sophisticated specifications to better mirror
the reality.
Despite its imperfections, our model adds value to other financial programming
models. The main structural feature is “endogenizing” private consumption, imports,
and inflation, in addition to oil price. In this way, three relevant markets of financial
programming, the foreign exchange market, the money market and the goods market,
are modelled jointly with the energy market. We can then measure the impact of
macroeconomic variables on oil price and vice-versa. More importantly, our model
can also forecast the oil price futures.
We estimate an econometric model of the US production based on the Cobb-Douglas
function. Even though economic literature states that oil price directly influences
gross domestic product, it could not be integrated in production function, as it is not a
productivity factor. Nevertheless, oil price was directly included into the private
consumption function.
We prove that an increase in WTI prices leads people to consume less and vice-versa.
Private consumption positively affects GDP through the accounting identity. Thus, a
rise in oil prices has also a negative effect on GDP. We also confirm that rising oil
prices add to inflationary pressures.
While money supply has been exogenously determined by the FED, money demand
has been defined as a function of national income, interest rate and inflation. As
expected, money demand increases with higher level of income, whereas higher
interest rate stimulates households to hold more other financial assets. Furthermore,
higher inflation results in a higher demand in money. Furthermore, oil prices
indirectly affect money demand as a function of inflation.
Imports can be interpreted as private consumption of foreign goods. We notice that
total consumption is positively related with imports. Moreover, oil price affects both
oil and total imports. Since part of oil consumed in the USA is imported, an increase
in oil price will indicate that a bigger part of income must be used for the same
quantity of oil imports. Consumer should then decrease his consumption of other
goods, which should lead to a general decrease in total imports. However, the United

States imports 58 percent of the crude oil it consumes, which means that 42 percent is
extracted domestically. Crude oil is therefore considered to be domestic as well as
imported good. As we show that imports are positively correlated with price of
domestic goods and negatively correlated with price of imported goods, we find effect
of oil prices on US imports ambiguous. Surprisingly, total imports increase even with
a rise in oil price.
When analyzing effects of macroeconomic variables on oil prices, we conclude that
the US economic situation does not affect oil price. Only the global world situation
can drive oil price through the supply and demand framework. However, this
statement is not completely true since oil price is valued in US$ and we have noticed
the effect of the US exchange rate on oil price. The most important factor seems to be
the gap between the demand and the supply. In the model, oil supply is considered as
an exogenous variable. In this way, we are able to measure the reaction of oil price to
supply shocks.
Speculation might play a big role in oil price rise. We integrate a part of it by using
futures contract but it does not seem to be satisfactory. Another way would be, for
instance, to add a variable, which could measure the confidence of investors.
The integration of oil in the various identities of financial programming enables us to
answer the question whether oil price impacts the U.S economy. However, as we
consider equilibrium coefficients in econometric relationships, the dynamic
possibilities of model are limited. Three major variables are used to link oil price and
U.S economically conditions: the oil supply, the exchange rate and the interest rate.
A decrease in oil supply and a depreciation of the US$ lead to an increase in oil price,
which in turn reduces private consumption. As a result, economic activity slows down
and the purchasing power decreases. At the same time, the rise in oil price pushes up
inflation that leads to a bigger depreciation of the US$ which then reinforces the
increase in oil price. We enter into an inflationary spiral.
The pressure on the Federal Reserve (FED) to cut the interest rate is assumed to
revitalize economic activity. With lower interest rates, private consumption will
increase along with an increase in gross domestic product and improvement in the
purchasing power. However, there is a side effect in the interest rates cut: the inflation
will increase, triggering depreciation of the US$. Such depreciation pushes up oil
price and the spiral is engaged.
This explains the importance of the dynamic component of the model. As a first step,
the FED can cut the interest rates at short-term to rebound economic activity, but this
measure should be followed by an increase in interest rates to fight inflation.
Eventually, the model confirms some general relations between selected
macroeconomic variables and oil price. It opens the gate for some more improvements
in this challenging framework.

6 Bibliography
Agenor, Pierre-Richard and Peter Montiel (1999), Development Macroeconomics,
Princeton University Press, 2nd edition
Alhajji, AF (2004), The impact of dollar devaluation on the world oil industry: Do
exchange rates matter ?", in Middle East Economy Survey, 2004
Alvarez, Fernando, Lucas, Robert et al (2001), Interest rates and inflation, Federal
Reserve Bank of Minneapolis, 2001
Barth, Chada (1989), "A simulation model for financial programming", IMF
WORKING PAPER, WP/89/24
Bier, Willem (1992), "Macroeconomic Models for the PC", IMF WORKING PAPER,
WP/92/110
Brown, Cronin (2007), "Commodity prices, money and inflation", European Central
Bank, Working paper series no. 738
Brouwer, Erichsson (1995), "Modelling inflation in Australia", Board of Governors of
the Federal Reserve System, International Finance Discussion Papers, No. 530
Cooke Ronald (2006), "Oil depletion economics", Energy Bulletin, July 2006
Dées, Gasteuil, Kaufmann and Mann (2008), "Assessing the factors behind oil price
changes", European Central Bank, Working paper series no.855
Dées, Karadeloglou, Kaufmann, Sanchez (2005), "Modeling the world oil market:
Assessment of a quarterly econometric model", Elsevier Ltd, Energy policy
Devolder Pierre (2008), Lecture : Finance Stochastique I
Fattouh, Bassam (2007), The drivers of oil prices, Oxford Institute for energy studies"
Frankel, Peter (2006) "Commodity prices and monetary policy", National Bureau of
Economic Research
Geman, Hélyette (2005), Commodities and commodity derivatives, Wiley Finance
Ghosh, Atish, Ostry Jonathan, et al (1996), "Does the exchange rate regime matter for
inflation and growth", IMF WORKING PAPER
Gorton, Rouwenhorst (2004), "Facts and Fantasies about Commodity Futures", Yale
ICF Working Paper No. 04-20, 2004
Krichene, Noureddine (2005), "A simultaneous equation model for world crude
oil&gas markets", IMF Working paper
MacDonald, Ronald, Nagayam Jun (2000), "The long-run relationship between real
exchange rates and real interest rate differentials: A panel sutdy", IMF WORKING
PAPER
Hirsch Robert, Bezdek Roger et al (2005), Peaking of oil production: impacts,
mitigation& risk management, MISI
Mikkelsen, Jan (1998), "A model for financial programming", IMF WORKING
PAPER, WP/98/80, 1998

Roubini, Nouriel and Brad Setser (2004), The effects of the recent oil price shock on
the U.S. and global economy, 2004.
Scherer, Ron (2008), "The rising impact of high oil prices", The Christian Science
Monitor
Sorensen, Whitta-Jacobsen (2005), Introducing advanced Macroeconomics: Growth
& Business Cycles, McGraw-Hill Education
Statistics Department of IMF (2007), "The System of Macroeconomic Accounts
Statistics", Pamphlet Series no.56 [25]
Williamson, Jeffrey (2005), Macroeconomics: 2nd edition, Addison Wesley
Wooldridge (2003), Introductory Econometrics: A modern approach", 2nd Edition
Yönetim ve Ekonomi, Celal Bayar University (1999), "The IMF and World Bank
Approaches to Macroeconomic Management in Developing Countries", Journal of
Economics& Administrative Sciences, No: 5, pp.367-378, 1999
http ://www.imf.org
http ://www.newyorkfed.org
http ://www.opec.org
LeBlanc, Chinn (2004), Do High Oil Prices Presage Inflation? The Evidence from G-
5 Countries, Department of Economics USSC
Cromb, Fernandez-Corugedo (2004), Long-term Interest Rates, Wealth and
Consumption, Bank of England
Thomas R L (1997), Modern Econometrics: an introduction, Addison Wesley
Longman
UC Berkeley Econometrics Lab, "TSP 4.5 Reference Manual"
http ://economics.about.com/cs/macrohelp/a/nominal-vs-real.htm
Shimon, Awerbush (2004), "Exploiting the oil-GDP effect to support renewable
deployments"
Scoffield H. (2008),"Oil prices changing consumer spending", Reportonbusiness.com
OCDE (2007), "Evolution des prix du pétrole : moteurs, conséquences économiques
et ajustement des politiques", Perspectives de l’OCDE n°76
Frimpong, Joseph Magnus and Oteng-Abayie, Eric Fosu (2006), "Aggregate Import
Demand and Expenditure Components in Ghana: An Econometric Analysis", 2006
International Energy Agency (2004), Analysis of the Impact of High Oil Prices on the
Global Economy, 2004
Mussa M. (2003),The Impact of Higher Oil Prices on the Global Economy, IMF
Elekdag, Lalonde, Laxton, Muir, Pesenti (2008), Oil Price Movements and the Global
Economy: A Model-Based Assessment
Zandi, DeKaser (2008), Oil shock hurting U.S.economy, United Press International
Zittel, Schindler (2007), Crude Oil: the Supply Outlook, ASPO International

Financial Programming and Oil Dynamics

More Related Content

What's hot (20)

Similar to Financial Programming and Oil Dynamics (20)

More from HELIOSPADILLAMAYER (13)

Recently uploaded (20)

Financial Programming and Oil Dynamics