The cost of inaction

The costofinaction:
Recognisingthevalueat
riskfrom climate change

1
THE COST OF INACTION:
R E C O G N I S I N G T H E V A L U E A T R I S K F R O M C L I M A T E C H A N G E
© THE ECONOMIST INTELLIGENCE UNIT LIMITED 2015
TABLE OF CONTENTS
2. EXECUTIVE SUMMARY
6. ACKNOWLEDGEMENTS

7. ABOUT THIS RESEARCH

8. METHODOLOGY:
MODELLING THE CLIMATE VAR
The need for further research
13. THE IMPACT OF CLIMATE CHANGE
A destabilising force
Stranded assets
17. ASSESSING PORTFOLIO RISK
Measuring carbon exposure
Motivations for investors to act
20. INVESTING, DIVESTING
AND ENGAGING
A free option on the mispricing of carbon
Engagement
Divesting high-carbon assets
24. THE STATE OF REGULATION
Leadership from developing markets
The beginnings of action
A price on carbon
Correcting market failures
Recommendations for action
31. CONCLUSION
32. APPENDIX

2
EXECUTIVE SUMMARY
The asset management industry—and thus the wider community of investors of all sizes—
is facing the prospect of significant losses from the effects of climate change. Assets can
be directly damaged by floods, droughts and severe storms, but portfolios can also be harmed
indirectly, through weaker growth and lower asset returns. Climate change is a long-term,
probably irreversible problem beset by substantial uncertainty. Crucially, however, climate
change is a problem of extreme risk: this means that the average losses to be expected are not
the only source of concern; on the contrary, the outliers, the particularly extreme scenarios, may
matter most of all.
To highlight the relevance of climate change to the asset management industry and beyond,
this research estimates the value at risk (VaR)1 to 2100 as a result of climate change to the total
global stock of manageable assets (the climate VaR). The world’s current stock of manageable
assets is estimated to be US$143trn.² The resulting expected losses to these assets identified
in our findings, in discounted, present value terms,³ are valued at US$4.2trn—roughly on a par
with the total value of all the world’s listed oil and gas companies or Japan’s entire GDP. This
is the average (mean) expected loss, but the value-at-risk calculation includes a wide range of
probabilities, and the tail risks are far more serious.
Warming of 5°C could result in US$7trn in losses – more than the total market capitalisation
of the London Stock Exchange - while 6°C of warming could lead to a present value loss of
US$13.8trn of manageable financial assets, roughly 10% of the global total.
These values are based on the discount rate of a private investor, a reasonable baseline as the
affected losses mentioned above will be on the privately held pool of global assets. However,
as climate change is also a systemic problem, with issues of wider societal concern, it is often
appropriate to apply a lower discount rate, consistent with public-sector actors that have longer
time horizons than individuals. When the expected losses are considered from the point of view
of a government, employing the same discount rates as the Stern Review,4 they rise dramatically.
From the public-sector perspective, the expected value of a future with 6°C of warming
represents present value losses worth US$43trn—30% of the entire stock of manageable assets.
By way of scale, the current market capitalisation of all the world’s stockmarkets is around
US$70trn.5
Written by
_______________
1 Value at risk measures the size of the loss a portfolio may experience, within a given time horizon, at a particular probability.
2 Our value for the stock of manageable assets is the total stock of assets held by non-bank financial institutions, as estimated by the Financial Stability Board. Bank assets are
excluded as these are, largely, managed by banks themselves.
3 Present value is a common financial metric used to assess the current worth of a future stream of cash flows given a specified rate of return. Future cash flows are discounted at
the discount rate, and the higher the discount rate, the lower the present value of the future cash flows. The cost of capital is commonly applied as a discount rate by both private
investors and public sector bodies.
4 The Economics of Climate Change: The Stern Review. Available at: http://webarchive.nationalarchives.gov.uk/+/http:/www.hm-treasury.gov.uk/independent_reviews/stern_
review_Ûconomics_climate_change/stern_review_report.cfm
5 World Federation of Exchanges.

3
While the value of future losses from the private sector is substantial, this is dwarfed by the forecast harms when considered from a government
point of view. The long time horizon, coupled with private-investor discount rates, can lead to a remarkable tolerance for systemic environmental
risk. The value at risk assessed by this research should be considered the expected losses to global assets if emissions fail to be substantially
reduced, but fortunately, mitigation can greatly reduce these risks. Lower greenhouse gas emissions decrease the probability of temperature
increases and thus the expected harms. Provided that warming from climate change can be kept under 2°C, the average projected losses can be
cut in half, while the extreme losses, identified as tail risks, can be reduced by more than three-quarters. Although the mean projected losses are
significant, the results also show that institutional investors are particularly at risk of lower probability but higher impact losses. Direct impacts
vary geographically; economic sectors and asset classes that are concerned with physical assets or natural resources are the most vulnerable to
climate change, such as real estate, infrastructure, timber, agriculture and tourism.
However, our analysis suggests that much of the impact on future assets will come through weaker growth and lower asset returns across the board.
These indirect impacts will affect the entire economy, even though the direct damage will be more localised. Indirect damage is a particularly
important portion of the overall risk in the more extreme scenarios (those with 5-6°C of warming). Asset managers cannot simply avoid climate
risks by moving out of vulnerable asset classes if climate change has a primarily macroeconomic impact, affecting their entire portfolio of assets.
In effect, total global output will be lower in a future with more climate change, rather than one with mitigation, and accordingly the size of the
future stock of manageable assets will also be lower.
Thirty years is a common time frame for pension funds and other long-term investors. But if
investors wait until these risks actually manifest themselves, then the options they will have to
deal with them will be significantly reduced. This is a vital concern, as the scope of investments
available to a future portfolio will be more limited in a world with severe climate change than in
one which has successfully mitigated climate risks. This means that future pensioners may see
the security of their retirement jeopardised as a result of the climate risk that the asset managers
charged with their investments are currently carrying.
These findings indicate that climate change is likely to represent an obstacle for many asset
owners and managers to fulfil their fiduciary duties. Fiduciary duty requires managers to act
in the best interest of their beneficiaries. In practice this means they need to deliver the best,
risk-adjusted returns possible. Unfortunately, too many investors currently overemphasise
short-term performance at the expense of longer-term returns. If investment managers are
aware of the extent of climate risk to the long-term value of the portfolios they manage, then it
could be argued that to ignore it is a breach of their fiduciary duty. Indeed, fiduciaries arguably
have an obligation to reduce the climate risk embedded in their portfolios. Yet to date few asset
managers have measured the climate-related risks embedded in their portfolios, much less tried
to mitigate them. According to estimates by the Asset Owners Disclosure Project,6 only 7% of
asset owners calculate the carbon footprint of their portfolios, and only 1.4% have an explicit
target to reduce it.
The good news is that there are widespread opportunities to reduce systemic environmental
risks, and many of them are clearly profitable. Some leading investors are already taking the
_______________
6 Global Climate 500 Index 2015, Asset Owners Disclosure Project. Available at: http://aodproject.net/climate-ratings/aodp-global-climate-500-index

4
initiative by investing in projects that finance the transition to a lower carbon
economy. Norway’s Government Pension Fund Global runs an environmental
fund of some NKr50bn (US$6bn) that is largely invested in alternative energy
and energy efficiency; Aviva, the UK-headquartered insurer, is targeting a £500m
(US$780m) annual investment in a low-carbon infrastructure over the next five
years, and Allianz of Germany has committed €2.5bn (US$2.72bn) to renewable
energy investments and plans to at least double its actual exposure in the
medium term.
Others are seeking to reduce long-term climate risks by decarbonising their
portfolios. This need not come at the expense of short-term performance. The
Swedish public pension fund AP4, for instance, has identified the 150 worst
performers, in terms of carbon intensity, in the S&P 500 index and divested its
holdings in them. The remaining 350 stocks track the performance profile of the
index closely but have 50% of its carbon footprint.
While proactive steps addressing climate risk can demonstrate leadership,
isolated activities will ultimately be insufficient. This is a collective action
problem that must be addressed if carbon emissions, and thus climate risks, are
to be reduced. It is clear that government action is required to establish a firm,
clear price that reasonably reflects its externality costs. Rather than opposing
this, institutional investors can collectively influence the companies in their
portfolios to adapt and prepare for a lower carbon future. Moreover, investors
can actively engage with policymakers, encouraging them to address this market
failure as something that is in their collective self-interest.
Although pricing carbon is essential, a carbon price alone is unlikely to
completely solve the problem of climate change; complementary policies are
necessary. The financial services sector has a vital role to play in managing the
tail risks. To do so, better information and more thorough disclosure are needed
by all market participants so that investors can make informed decisions.
Financial institutions, however, have an obligation to manage their tail risks, and
institutional investors specifically must manage their funds with the long-term
benefit of their beneficiaries in mind. For this to be possible, regulators should
issue guidance explicitly recognising climate risks as material. This means that
disclosure of carbon emissions and acknowledgement of climate-related risks by
publicly listed companies should be mandatory. Institutional investors should be
able to assess and, where feasible, mitigate their climate risks accordingly.
• The value at risk to manageable assets from climate change
calculated in this report is US$4.2trn, in present value terms.
• The tail risks are more extreme; 6°C of warming could lead to
a present value loss worth US$13.8trn, using private-sector
discount rates.
• From the public-sector perspective, 6°C of warming represents
present value losses worth US$43trn—30% of the entire stock
of the world’s manageable assets.
• Impacts on future assets will come not merely through direct,
physical harms but also from weaker growth and lower asset
returns across the board. The interconnected nature of the
problem will reduce returns, even on investments unharmed by
physical damage.
• Although direct damage will be more localised, indirect impacts
will affect the entire global economy; accordingly, asset
managers will face significant challenges diversifying out of
assets affected by climate change. Institutional investors need
to assess their climate-related risks and take steps to mitigate
them; very few have begun to do this.
• Regulation has largely failed to confront the risks associated
with climate change borne by long-term institutional investors.
To enable meaningful risk analyses, public companies should
be required to disclose their emissions in a standardised and
comparable form.
• Carbon pricing is crucial to addressing climate change.
Government inaction with respect to this market failure
neglects an issue of systemic risk and global importance.
KEY FINDINGS

5
Imperfect data availability and patchy admissions on climate risk leave both regulators and
institutional investors unable to adequately address these risks. Moreover, effectively co-
ordinated regulation is necessary so that best practice can become standard practice. Without
requirements to recognise climate risks as material, many organisations will choose to ignore
them, creating “free riders” who shirk their own responsibilities while contributing to the long-
term, systemic impact of climate change.
France is going further than issuing guidance. In May 2015 its National Assembly voted to
require French institutional investors to disclose information on sustainability factors in their
investment criteria, and to explain how they take into account exposure to climate risks and
how they measure greenhouse gas emissions associated with assets held in their portfolios.
This makes sense, because just as a particular institution may represent systemic financial risk,
similarly climate risks may be concentrated but poorly assessed by institutional investors;
regulators need clarity as to where these long-term risks are borne. For these assessments to be
meaningful, regulators need to require companies to disclose their carbon emissions and related
risks so that investors can make informed decisions.
To avoid sleepwalking into a climate crisis, large-scale efforts, such as France’s, are needed from
both the public and the private sector. Moreover, to bolster effectiveness and avoid regulatory
arbitrage, there is a clear need for co-ordinated action by national governments, institutional
investors, regulatory bodies and international financial organisations. The UN Climate Change
Conference (Conference of the Parties, or COP21) due to take place in Paris at the end of this
year offers a major forum for governments to address this market failure and to chart a path
towards mitigating climate change. If there are no strong commitments to reduce greenhouse
gas emissions and meaningful actions to price carbon, then this historic opportunity will have
been wasted.

6
ACKNOWLEDGEMENTS
This research endeavour was sponsored by Aviva and supported by the Mlinda Foundation and KPMG; it builds on an original proposal by Steve
Waygood at Aviva Investors. The Economist Intelligence Unit (The EIU) bears sole responsibility for the content of this report. The findings and
views expressed in the report do not necessarily reflect the views of the sponsors. Christopher Watts was the author of the report, and Brian
Gardner was the editor.
The experts below have been kind enough to review the work conducted during the course of this research programme. Regardless, The EIU
maintains full editorial control of this white paper; neither the reviewers nor their organisations necessarily support nor endorse the views
expressed in the course of this report. We sincerely thank them for their time and participation.
Reviewers are listed below in alphabetical order by organisation:
• Howard Covington of the Alan Turing Institute
• Anthony Hobley of the Carbon Tracker Initiative
• James Leaton of the Carbon Tracker Initiative
• Mark Fulton of the Carbon Tracker Initiative
• Jonathon Porritt of the Forum for the Future
• Alex Bowen of the Grantham Research Institute on Climate Change and the Environment at the London School of Economics
• Serena Brown of KPMG
• Vincent Neate of KPMG
• Marloes Nicholls of Meteos
• Sophia Tickell of Meteos
• Richard Azarnia of the Mlinda Foundation
• Olivier Cassaro of Preventable Surprises
• Raj Thamotheram of Preventable Surprises
• Nick Robins of the United Nations Environmental Programme, Inquiry into the design of a sustainable financial system
• Professor Tim Jackson of the University of Surrey
Lastly, we would like to offer a special thanks to Simon Dietz, Charlie Dixon, Jason Eis and Philip Gradwell of Vivid Economics for their contribution
to making this research possible.

7
The Economist Intelligence Unit would like to thank the following experts (listed alphabetically by organisation name) who participated in the
interview programme:
• Tom Wilson, chief risk officer, Allianz,
Germany
• Laurent Clamagirand, chief investment
officer, AXA Group, France
• Felix Hufeld, president, Bundesanstalt für
Finanzdienstleistungsaufsicht (BAFIN,
Federal Financial Supervisory Authority),
Germany
• Gunnela Hahn, head of responsible
investment, Church of Sweden, Sweden
• Philippe Desfossés, CEO, Établissement
de Retraite Additionnel de la Fonction
Publique (ERAFP), France
• Mats Andersson, CEO, Fjärde AP-fonden
(AP4), Sweden
• David Blood, managing partner,
Generation Investment Management, UK
• Al Gore, chairman, Generation Investment
Management, UK
• Stephanie Pfeifer, chief executive,
Institutional Investors Group on Climate
Change, UK
• Sebastian von Dahlen, chairman, Global
Systemically Important Insurers Analysts
Working Group, International Association
of Insurance Supervisors, Switzerland
• Jonathan Bailey, consultant, McKinsey &
Company, US
• Dag Huse, chief risk officer, Norges Bank
Investment Management, Norway
• Yngve Slyngstad, CEO, Norges Bank
Investment Management, Norway
• Odd Arild Grefstad, CEO, Storebrand Group,
Norway
• Tobias Reichmuth, CEO and co-founder,
SUSI Partners, Switzerland
• Lauren Smart, executive director, Trucost,
UK
• Nick Robins, co-director, Inquiry into the
Design of a Sustainable Financial System,
United Nations Environment Programme
(UNEP)
• Rory Sullivan, senior research fellow,
Centre for Climate Change Economics and
Policy, University of Leeds, UK
• Mike Kreidler, commissioner,
Washington State Office of the Insurance
Commissioner, US
ABOUT THIS RESEARCH
The cost of inaction: Recognising the value at risk from climate change is a report by The Economist Intelligence Unit (The EIU). The research
depicts the scope of assets at risk from climate change from the present to 2100. This innovative achievement draws on a modelling endeavour
that combines The EIU’s long-term forecasts with a nuanced, integrated assessment model provided by Vivid Economics. The full methodology
is provided in the appendix to this report. This white paper further discusses the possible consequences of climate change as well as how both
investors and governments are measuring and responding to climate-related risks. The findings of this paper are based on detailed modelling,
extensive desk research and interviews with a range of experts, conducted by The Economist Intelligence Unit.

8
METHODOLOGY:
MODELLING THE CLIMATE VaR
A core responsibility of asset managers and institutional investors is to manage risk, and the
most commonly employed measure to assess it is value at risk (VaR). This measures the size of
the loss a portfolio may experience, within a given time horizon, at a particular probability.
In a pioneering endeavour to highlight the relevance of climate change to the investment
community, this research estimates the VaR to 2100 of the global stock of manageable assets
owing to the impacts of climate change, referred to in this report as the climate VaR.
In particular, the estimates of climate VaR comprise the effect of climate change this century
on the global stock of manageable financial assets, in present value terms. The global stock of
manageable financial assets today is quantified at US$143trn, which is the stock of assets held by
non-bank financial institutions, according to the Financial Stability Board.
To estimate the effect of climate change to 2100 on the changing stock of manageable financial
assets, The Economist Intelligence Unit (The EIU) and Vivid Economics have used a leading,
peer-reviewed forecasting model of the impact of climate change on the economy, the DICE
(Dynamic Integrated Climate-Economy) model. DICE is one of a small number of integrated
assessment models (IAMs) that have been built to estimate the economic cost of future climate
change. These models link economic growth, greenhouse gas emissions, climate change and the
damages from climate change back on the economy, and they do so in an integrated, consistent
framework. They are typically built by adding a simple model of climate change to an existing
framework for modelling the macroeconomy, with carbon emissions and climate damages being
the links between the two. DICE is the most popular of these models, having been used and cited
in thousands of academic studies over nearly three decades. It is publicly available, and several
evaluations have been performed of its forecasts. For example, it has been shown to produce
forecasts of climate change in line with much more complex physics-based models, such as that
held by the UK Met Office.
The traditional purpose of IAMs has been to estimate the size of the climate change externality—
the social cost of greenhouse gas emissions—in order to inform policymakers in setting emission
targets or carbon prices. A famous example of such an exercise is the Stern Review, which
estimated the present value of the future social costs of climate change to be equivalent to
5-20% of global GDP. The US Environmental Protection Agency has also recently used a suite
of IAMs, including DICE, to determine the social costs of carbon for federal regulatory impact
assessments. Since the value of financial assets is intrinsically linked to the performance of the
economy, the innovation of this study is to use the DICE model to estimate the impact of climate
change on financial assets instead.

9
This modelling recognises that, since the present value of a portfolio of equities is just the discounted cash flow of future dividends, then in the
long run—ie, over the course of a century—dividends in a diversified portfolio should grow at the same rate as GDP, because ultimately dividends
are paid for from the output of the economy.7
In well-functioning financial markets the same relationship with GDP growth should hold for cash
flows from other kinds of assets, such as bonds. This relationship may not be observed over a relatively long time period, even decades, owing
to business cycles; for example, corporate profits are currently at historical highs, while GDP growth is low. However, on average, to 2100, this
relationship can be expected to hold up. The DICE model is then used to forecast the effect of climate change on GDP, and in turn on cash flows from
assets.
The appendix further details an alternative approach, which uses estimates made directly by the DICE model of the impact of climate change on the
stock of non-financial capital assets. These are then converted to manageable assets, based on estimates of the share of non-financial assets used
to back financial liabilities, and the ratio of financial liabilities created per US dollar of non-financial assets.
This is an up-to-date version of DICE, which extends the model to incorporate direct damages from climate change to the stock of non-financial
capital assets, as well as the more traditional route of modelling a reduction in the amount of goods and services that can be produced with given
inputs of capital and labour. To estimate the climate VaR at different confidence levels, there are three key uncertainties, which the academic
literature has identified as being particularly determinative of the impacts of climate change, that are assessed as part of this Monte Carlo analysis.
The first is the rate of productivity growth this decade, reflecting uncertainty over general macroeconomic conditions. This sets the magnitude
of growth over the rest of the century, which exerts a strong influence on the size of assets in the future and, through the link between economic
activity and carbon emissions, on the amount of warming along a path of uncontrolled emissions.
After the initial decade productivity growth follows The EIU’s long-term forecasts, which predict
increasing productivity over the long term. An alternative approach based on a decreasing
productivity scenario was used as a check for robustness; further details on this can be found
in the appendix. The decreasing productivity scenario, in line with expectations of secular
stagnation, yields a climate VaR that is even higher than that discussed in the body of this
report.
The second key uncertainty is climate sensitivity, which is by how much the planet warms in
response to a given increase in greenhouse gases in the atmosphere. Climate sensitivity captures
key uncertainties in the climate system, in particular the role of feedbacks in the warming
process, so the probability distribution is calibrated on the latest scientific consensus from the
Intergovernmental Panel on Climate Change (IPCC). The third is the risk of catastrophic climate
change, embodied in DICE’s representation of economic damages. This probability distribution
captures the divergence of views in the academic literature on the possibility of catastrophic
impacts beyond warming of 3°C.
To discount future impacts of climate change on the stock of assets back to the present, two
perspectives are taken. The first is that of a private investor, whose initial discount rate is
representative of the rate of return on a diversified portfolio of assets with some undiversifiable,
systemic risk, in line with the capital asset pricing model. This discount rate then moves in
line with changes in GDP growth in the future, based on a premium to account for bearing
undiversifiable risk. The GDP growth rate without climate change is used, as investors do
not currently consider climate impacts in their asset valuations. This provides a conservative
estimate of losses, as GDP growth with climate change will be lower, which should lead to a lower
_______________
7 The Case for Forceful Stewardship (Part 1): The Financial Risk from Global Warming (SSRN, Covington & Thamotheram,
January 2015) Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2551478

10
discount rate were investors to take this into
account. Average private-sector discount rates
begin at 5.5%, but fall to 4% towards the end of
the century owing to slowing economic growth.
The alternative perspective takes the viewpoint
of a public-sector regulator examining the social
value of climate risk to financial assets. It is
widely accepted that public-sector discount rates
should be below private-sector discount rates, for
reasons including a lower cost of capital as well
as the avoidance of market distortions, such as
taxation and externalities, which means that the
gross returns on private investments are above
their net social returns. This is modelled with the
social discount rate applied by the Stern Review,
which is appropriate for very long-run problems
such as climate change, because it treats the
wellbeing of future generations on a par with
the wellbeing of current generations with some
included uncertainty. The remaining justification
for discounting in this framework is economic
growth and its effect of reducing the marginal
utility of consumption for future generations.
Average government discount rates are initially
3.8% but fall to 2% towards the end of the
century owing to slower economic growth.
The results show that the asset management
industry is particularly subject to tail risks: lower
probability but higher impact losses. Figures 1-4
show damage across the range of temperature
levels over time. The damage increases over time,
especially in the latter half of the century, when
the effect of emissions released today begins to
feed through to the climate. Even the average
(mean) losses are significant. However, it is the
risk in the “tails” of the distribution that is most
severe, and this tail risk increases rapidly over
time.
Our findings, when discounted back to their value
in the present day, expect a mean (average) loss
of US$4.2trn from the perspective of a private
investor. However, this is only the mean; our
value at risk calculation includes a range of
probabilities, and the tail risks are far more
VaR due to climate
change
VaR due to climate
change
Mean
(average)
Mean
(average)
5°C
5°C
6°C
6°C
Present value from the
perspective of a private investor
The perspective of
a private investor
US$4.2trn
50%
US$7.2trn
63%
US$13.8trn
76%
Present value from the
perspective of a government
The perspective of
a government
US$13.9trn
57%
US$18.4trn
71%
US$43.0trn
85%
Table1-Valueatrisk
Table2-Percentagereductioninriskshouldwarmingbekeptwithin2°C
Figure 1: Losses are primarily incurred in the latter half of the century
Present value loss to current manageable assets (trillion $, 2015 prices)
Note: Losses are discounted at private sector discount rates. The losses over time are consistent with paths to the respective temperature levels being
reached in 2105.
Source: Vivid Economics.
6°C loss 5°C loss 4°C loss 3°C loss Average loss
0
2
4
6
8
10
12
14
16
18
2105209520852075206520552045203520252015

11
serious. Warming of 5°C is consistent with
US$7trn in losses—more than the total market
capitalisation of the London Stock Exchange—
while 6°C of warming is consistent with a present
value loss of US$13.8trn to manageable financial
assets, roughly 10% of their global total. Table 1
highlights the extent of these losses.
Tail risks need to be taken very seriously when
considering climate change. As highlighted by
the 2008 financial crisis, tail risks are subject to
miscalculation, but ignoring them can lead to
exceedingly negatively outcomes. The estimated
harms depicted in the modelling are conservative
in nature, and the average result combines a
range of unlikely but less damaging outcomes
as well as the more extreme negative outcomes.
This brings down the expected losses. It should
be recognised that unless climate change is
mitigated, the modelling depicts a permanent
divergence towards a path of lower growth
and diminished prosperity. Accordingly, it is
important to recognise that losses from climate
change do not merely represent market volatility
or business cycles but permanent impairments to
total assets.
Understandably, as the size of the expected losses
is very heavily influenced by the discount rate
employed, a discount rate more appropriate for
the public sector regulators examining systemic
risk would be significantly lower. For comparison,
the climate VaR expected using the discount rates
applied by the Stern Review are more than three
times those of a private-sector investor. This may
explain some of the unwillingness on the part of
the private sector to address these issues. It does
not, however, excuse government inaction on an
issue of systemic risk.
This is particularly the case as there are
opportunities to greatly reduce the risks from
climate change. Table 2 presents the percentage
reduction in the climate VaR of a mitigation
scenario consistent with a 66% probability of
remaining under 2°C of warming, relative to the
Figure 3: Range of expected losses discounted from a government perspective, in accordance with the Stern
Review
0
10
20
30
40
50
60
2105209520852075206520552045203520252015
Figure 2: Private sector percentage loss of future dividends
Percentage loss of value in 2015 manageable assets
Average lossRange of 3°C-6°C loss 5°C loss 4°C loss
0
2
4
6
8
10
12
2105209520852075206520552045203520252015

12
baseline expected losses.8
Increased mitigation
reduces carbon emissions, and thus the likely
array of future temperatures rises, occurring at
different levels of probability. This lowers risks,
and it reduces them most dramatically at the tail
ends of the distribution where the worst potential
outcomes lie.
These results are for total global assets, and DICE
is a globally aggregated model that does not allow
explicit disaggregation by asset class or by region.
Nonetheless, impacts can be expected to vary
by asset class and region. The impacts of climate
change, at least for modest degrees of warming,
can be expected to concentrate in sectors of
the economy sensitive to weather conditions,
for instance agriculture, energy, forestry and
water. However, these sectors are connected
with the rest of the economy through supply
and demand linkages, and shifts in the prices of
goods and services because of climate change will
affect overall spending patterns and household
incomes. Allied to the fact that at higher degrees
of warming the impacts of climate change are
expected to become increasingly economy-wide,
this means that climate change poses a systemic
risk, coming through weaker growth and lower
asset returns affecting the entire portfolio.
As a result, asset managers may struggle to
avoid climate risks by moving out of vulnerable
asset classes and regions. This is because, at
least under lower-probability and higher-impact
outcomes, our findings suggest that climate
change will primarily have a macroeconomic
impact that affects the entire portfolio of assets.
The interconnected nature of the problem is likely
to reduce returns, even on those investments
not actually harmed by physical damage. Given
this result, asset managers will face significant
challenges diversifying out of assets affected by
climate change.
Figure 4: Government sector percentage loss of future dividends
Percentage loss of value in 2015 manageable assets
Average lossRange of 3°C-6°C loss 5°C loss 4°C loss
0
5
10
15
20
25
30
35
40
2105209520852075206520552045203520252015
_______________
8 The minimal mitigation scenario is characterised by future inaction and stalemate on climate policies, which provides an emissions reduction of 6.5% in 2105 relative to a world of
no further action whatsoever. Keeping global warming within 2°C as the “likely” outcome is the stated goal of the IPCC’s mitigation analysis, and it uses the benchmark of at least 66%
chance as the threshold.

13
This report is a major step towards understanding
the impact of climate change on financial assets.
However, there is substantial uncertainty
embedded in any attempt to make such long-
term forecasts, and the model itself necessarily
presents an aggregated view of the world. Rather
than detracting from the findings, which are
significant, this highlights the need for further
research.
While by no means comprehensive, below are
some areas in need of further exploration:
• Geographical and sector-specific risk
assessments. Ideally, this would lead to a
comprehensive, bottom-up model covering
all assets and assessing both direct and
indirect impacts from climate change.
Research should then aim to further
disentangle the level of risk that cannot be
diversified away.
• Investigations into the interaction between
physical and financial assets. While financial
markets have tended to be resilient to
isolated natural disasters, the impact when
affected at systemic level is less clearly
understood.
• Detailed analyses need to explore the
human and social costs of the more extreme
potential outcomes. The substantial GDP and
total asset impacts identified in this research
suggest the likelihood of major dislocations
to economies and societies around the
world. This modelling endeavour is not able
to depict this at a local or national level, but
given the magnitude of the risks involved,
further work is clearly merited.
THE NEED FOR FURTHER RESEARCH
_______________
9 Climate Change 2014: Synthesis Report, Intergovernmental Panel on Climate Change, 2104. Available at:
https://www.ipcc.ch/pdf/assessment-report/ar5/syr/AR5_SYR_FINAL_SPM.pdf
THE IMPACT OF CLIMATE CHANGE
Since 1992, when the United Nations Framework Convention on Climate Change (UNFCCC)
was first negotiated in Rio de Janeiro, much time has been lost without adopting
mitigation policies to reduce the effects of climate change. Now, 23 years later, the 21st
session of the Conference of the Parties to the UNFCCC (COP21), due to take place in
Paris at the end of 2015, may provide a forum for governments to reach a substantive
agreement on collective action to combat climate change.
According to current projections from the Intergovernmental Panel on Climate Change
(IPCC), the earth’s surface temperature is forecast to continue rising in the remainder of
the 21st century; the global mean surface temperature increase for the period 2016–35
relative to 1986–2005 is likely to be in the range of 0.3°C to 0.7°C. Furthermore, heat
waves will occur more often, extreme precipitation events will become more intense and
more frequent in many regions, and oceans will continue to warm and acidify, while sea
levels are expected to rise.9
Nevertheless, the precise impacts of these changes in climate and weather patterns are far
from predictable. According to Rory Sullivan, senior research fellow at the ESRC Centre for
Climate Change Economics and Policy at the University of Leeds, climate change-related
risks must be assessed and managed in the same way as other risks, where attention must
be paid to both probability and consequences.
Consider weather events such as storms. Mike Kreidler, Washington’s state insurance
commissioner, forecasts: “We’re going to be looking at storms with greater intensity; we’re
going to face more problems related to wildfires and droughts; tornadoes that carry more
of a punch; and hail that is larger and more damaging.” Yet pinpointing where and when
these events might occur, and predicting their individual impact, is not feasible with any
degree of precision. And there is some uncertainty concerning the extent and the pace at
which climate change will lead to these risks materialising.
Nick Robins, co-director of the Inquiry into the Design of a Sustainable Financial System at
the UN Environment Programme (UNEP), points out that even as efforts to mitigate climate
change bear fruit, there remains a probability of a highly destructive scenario outcome.
While the risk of that outcome is low, the results may be catastrophic. “We wouldn’t get on
a plane if there was a 5% chance of the plane crashing,” he says. “But we’re treating the
climate with that same level of risk in a very offhand, complacent way.”
The effects of climate change are expected to be long-term, cumulative and probably
irreversible. In North America and Europe, according to the IPCC, key risks include
increased damages from river and coastal floods and from wildfires; in Asia, water and

14
We wouldn’t get
on a plane if there
was a 5% chance
of the plane
crashing, but
we’re treating the
climate with that
same level of risk
in a very offhand,
complacent way.
Nick Robins,
co-director, Inquiry into the Design of
a Sustainable Financial System, United
Nations Environment Programme
food shortages are a significant threat; and in Africa, risks include stress on water resources,
reduced food security and a greater threat of vector- and water-borne diseases.
Keeping warming within 2°C as the “likely” outcome is the stated goal of the IPCC’s mitigation
analysis, and it uses the benchmark of an at least 66% chance as the threshold. Applying this
threshold, the mean projected losses from the climate VaR—the value at risk (VaR) of global
assets under management owing to the impacts of climate change—would be cut in half. More
importantly, our findings indicate that mitigation has the potential to cut the projected harms of
even the extreme losses identified as tail risks by more than three-quarters.
Global warming expected to 2050 is a result of past emissions; in other words, warming will occur
even if further emissions are halted immediately. “Once the greenhouse gases are up in the air,
climate change then is effectively irreversible,” says Mr Robins of UNEP. “Unchecked, climate
change will be very disruptive during this first half of this century. And then into the second half,
it will be very destructive and potentially catastrophic.” Owing to the delayed effect of carbon
dioxide emissions, the full consequences of greenhouse gases emitted today may not be felt until
long after they have entered the atmosphere.
As such, it is possible that any “tipping point” will not be recognised until after it has been
reached. But positive action taken today will lessen the impact of climate change in the second
half of this century. “I wonder whether ... the accumulated burden over time and the lack of
willingness to directly address [emissions] in a meaningful manner is going to leave us in a
position where it could be a catastrophe and we just miss the opportunity,” says Tom Wilson,
chief risk officer at Munich-based insurer Allianz.

15
_______________
10 Our value for the stock of manageable assets is the stock of assets held by non-bank financial institutions, as estimated by the Financial Stability Board. Bank assets are excluded
as these are largely managed by banks themselves. The value of the world’s stockmarkets is taken from the World Federation of Exchanges.
11 American Climate Prospectus: Economic Risks in the United States, October 2014. Available at: http://rhg.com/wp-content/uploads/2014/10/AmericanClimateProspectus_
v1.2.pdf
A destabilising force
The effects of climate change on capital markets have the potential to destabilise the global
financial system. In April 2015 the G20 group of major economies asked the Financial Stability
Board, created by the G20 and hosted by the Bank for International Settlements (BIS) in Basel,
to conduct an inquiry into the potential risks to global financial stability posed by climate
change or regulation aimed at combatting the risks associated with natural disasters. “I believe
natural disasters could affect global financial stability through different channels,” says
Sebastian von Dahlen, chairman of the Global Systemically Important Insurers Analysts Working
Group at the International Association of Insurance Supervisors (IAIS) in Basel.
What is the scope of the economic cost of climate change? This report estimates the present-
day climate VaR to be US$4.2trn. This is the mean (average) from the standpoint of a private
investor. The tail risks are far more serious. Should there be warming of 5°C or 6°C, then the
expected losses would rise to US$7trn or US$13.8trn respectively. These figures represent
the harm to financial assets from the impacts of climate change if warming is not significantly
mitigated.
When these risks are considered from the public-sector perspective, then a future with 6°C of
warming represents discounted, present-day losses worth US$43trn, while 5°C of warming is
consistent with losses of US$18.4trn. By way of scale, the current market capitalisation of all the
world’s stockmarkets is around US$70trn, while the current stock of manageable assets is around
US$143trn, according to the latest estimates.10
The potential for climate-related destruction includes the risk of direct destruction of assets.
In the US alone, climate change and severe weather patterns have the potential to put between
US$238bn and US$507bn worth of coastal real estate under water by 2100 as sea levels rise,
according to one estimate.11
“Anything you can’t move is of concern,” says Mr Sullivan of Leeds
University. “That’s why sea-level rise is probably the most immediate risk for investors to be
concerned about.”
Climate change is expected to cause significant physical damage, in particular owing to the
increased frequency of severe storms and flooding. Sea-level rise, in some of the more extreme
scenarios, is likely to displace millions of people. To some degree this makes sense, as many
of the world’s most densely populated areas are located in low-lying coastal plains. While the
human and the geopolitical implications that such disruption might cause are far from clear, the
possibility of conflicts, refugee crises and widespread social dislocation would almost certainly
increase.

16
The modelling endeavour undertaken in this research is only able to provide an aggregate view of the damages. However, at higher levels of
warming the trillions of dollars in present value losses represent economic damages comparable with those seen in wars or civil conflicts. Unlike
isolated events, however, climate change is global in scope and largely irreversible, presenting a permanent divergence towards a path of lower
growth and diminished prosperity.
Similarly, increased frequency of droughts and severe storms will reduce food production and undermine water supplies in some locations. In
aggregate, climate change is expected to undermine global food security, but if kept to a low level, this is likely to be manageable. Facing higher
levels of warming, however, the IPCC’s Fifth Assessment Report highlights that a large faction of the world’s species will face an increased risk
of extinction as a result of climate change.12
For instance, most plant species are unable to shift their geographical range fast enough to adapt
to unmitigated warming. The future following that sort of systemic collapse cannot be fully
modelled.
There is also an increasing understanding that there will be direct impacts on human health,
ranging from heat waves and fires to malaria and other vector-borne diseases. Poorer regions,
less able to pay for adaptation measures, will suffer particularly acutely: the physical impacts
alone will be devastating in many parts of the world.
Crucially, however, it is not just physical assets that are threatened with destruction from climate
change. There is also significant scope for indirect destruction, for example through lower economic
growth that will be a consequence of climate change; lower levels of savings will have a negative effect
on investment levels; and lower returns on assets are likely to lead to poor performance of equity
portfolios across the board. “If disasters are becoming more extreme, then I think it's not so much
the question which asset classes are exposed, but rather, are there any assets which wouldn't be,”
highlights Mr von Dahlen of the IAIS.
Stranded assets
While leaving climate change unchecked will result in large-scale value destruction, it should be
acknowledged that there is also scope for impacts on investor portfolios should major regulatory
efforts to combat climate change be implemented. Among these is the impact of the stranding of
assets—not least the substantial portion of known coal, oil and gas reserves which will probably remain
unburned if climate change is to be limited—a scenario not currently priced into current valuations of
these assets.
Owners and managers of assets in different countries are likely to be exposed to differing degrees
of potential losses in their portfolios. In the UK, for example, the benchmark FTSE 100 share index
holds a very substantial percentage of businesses that rely on carbon-intensive activities or assets,
so that investors would face sharper potential losses if the prices of these assets were to experience
a correction. At the same time, institutional investors in the UK, such as pension fund managers,
favour index tracking more than elsewhere, meaning that ownership of carbon-intensive businesses is
especially widespread.
_______________
12 Intergovernmental Panel on Climate Change, Fifth Assessment Report. Available at: http://www.ipcc.ch/pdf/assessment-report/ar5/syr/AR5_SYR_FINAL_SPM.pdf

17
_______________
13 Global Climate 500 Index 2015–The Asset Owners Disclosure Project. Available at: http://aodproject.net/climate-ratings/aodp-global-climate-500-index
14 Insurer Climate Risk Disclosure Survey Report & Scorecard: 2014 Findings & Recommendations, Ceres, 2014. Available at: http://www.ceres.org/resources/reports/insurer-
climate-risk-disclosure-survey-report- scorecard-2014-findings-recommendations/view
You can ask
pension funds
about traditional
risk metrics, but
ask them about
their exposure
to climate risk
and I’m sure that
90% will say they
haven’t got a clue.
Mats Andersson,
CEO, Fjärde AP-fonden (AP4)
And there is no guarantee that any effects of climate change on the valuation of assets will be gradual
or linear, so that a wait-and-see approach among investors potentially presents more risk to portfolios.
“Regulations and restrictions on coal in 20 years’ time: what would that do in terms of the current value
of those stocks that we hold, and what do we think the path from here to there would look like?” asks
Jonathan Bailey, a consultant at McKinsey & Company in New York. “There may suddenly be a consensus
that shifts pricing.” Inaction may leave investors holding such assets at particularly high risk.
This means that global investors are currently facing a stark choice. Either they will experience
impairments to their holdings in fossil-fuel companies should action on climate change take place,
or they will face losses to their entire portfolio of manageable assets should little mitigation be
forthcoming. Charting a path away from these two options should be a strong motivation for long-
term investors to engage with companies in their portfolios and to shift investments towards a
profitable, low-carbon future.
ASSESSING PORTFOLIO RISK
What are investors doing to assess long-term risk in their portfolios? Only 7% of asset owners are
currently calculating the carbon emissions embedded in their portfolios.13 And in the US, just over
10% of insurers surveyed have published their climate risk management principles, according to
a recent report.14 The report’s authors note that “most of the companies responding to the survey
reported a profound lack of preparedness in addressing climate-related risks and opportunities.”
It helps very little that the financial markets typically operate on a short time horizon. While in
theory asset managers have an obligation to maximise risk-adjusted returns over the long term,
in practice the tendency is to emphasise short-term returns. This is a problem more generally: if
company executives’ focus is too short-term, it can hamper those investments needed to ensure the
long-term health of their companies; however, investors with longer-term horizons are required to
focus on the long-term financial health of a company. But beyond general concerns, short-termism
is a particularly acute problem for climate change owing to its exceedingly long-term nature.
This inevitably has implications for institutional investors with long-term liabilities such as
insurers, pension funds and sovereign wealth funds. Norway’s Government Pension Fund Global,
for example, looks 30 years ahead. “We think of the firm on a much longer horizon than that, but
realistically and practically 30 years is what we operate with,” says Yngve Slyngstad, CEO of Norges
Bank Investment Management, which runs the fund. Even this recognisably long-term horizon is a
problem when addressing climate change, as actions taken in 2070 will be far too late to influence
the level of warming from current trends.

18
As growing numbers of institutional investors address the climate-
related risks in their investment portfolios, they are driven to
act by risk management goals, financial returns and regulatory
mandates. For those that have not yet taken steps to address the
long-term risks they face, here are some reasons to do so:
• Yngve Slyngstad, CEO of Norges Bank Investment Manage-
ment, which runs Norway’s Government Pension Fund Global,
says that the mandate to manage the sovereign wealth fund
clearly stipulates that the fund’s main driver is financial re-
turn. “Given the premise that it’s very long-term, we are now
looking at sustainable business models,” he says.
• Storebrand Group is another asset manager which believes
that sustainability is a prerequisite for securing long-term
financial results. “To understand the risks and the opportu-
nities that companies are facing over the long term is vital
to making good investment decisions,” says CEO Odd Arild
Grefstad. “And we believe that, over time, our sustainability
research will contribute to higher quality and also positive
returns.”
• Increasingly, says Philippe Desfossés, CEO of French pension
fund Établissement de Retraite Additionnel de la Fonction
Publique (ERAFP), asset managers may be called to account
for not incorporating sustainability into the management of
their portfolios, “because now it has really been made obvi-
ous that carbon is a risk, and it’s a risk for business,” he says.
In case of legal dispute, he points out: “It’s now very difficult
for anyone to say ‘Oh, I didn’t know it was a risk’.”
• In some cases, investment management mandates stipulate
that funds must be managed with sustainability in mind but
without sacrificing financial returns—especially in public
pension funds. “We are regulated by law, and the law says
that we should allocate capital on a 30-40-year horizon,”
explains Mats Andersson, CEO of Swedish public pension fund
Fjärde AP-fonden (AP4). “The law also states that we should
take sustainability into account without giving up returns.”
MOTIVATIONS FOR INVESTORS TO ACT
Recognising that climate change represents a significant long-term risk
to their portfolios, some leading institutional investors are starting to
take decisive action—but they remain in the minority. “I think the most
important thing is to get the pension fund asset managers to understand
that climate is a risk,” says Mats Andersson, CEO of Swedish public pension
fund manager Fjärde AP-fonden (AP4). “You can ask pension funds about
traditional risk metrics, but ask them about their exposure to climate risk
and I’m sure that 90% will say they haven’t got a clue.”
In recognition of the long-term systemic risks, a response to climate
change should be incorporated into the legal duties of market participants,
in particular, the fiduciary duty of asset managers. Laurent Clamagirand,
chief investment officer of France’s AXA Group, asserts: “If ‘carbon’ is
considered to be a risk, investors should attempt to identify, measure and
reduce it on behalf of their beneficiaries.”
Consequently, climate change presents a significant challenge to the
ability of institutional investors to fulfil their fiduciary responsibilities.
Institutional investors with a long-term investment horizon are deeply
implicated. “Many financial institutions, such as pension funds, aim to be
around for many decades,” says Mr Robins of UNEP. “And in that context,
their ability to continue to honour the promises they may have made to
their beneficiaries will be profoundly impacted.”
Measuring carbon exposure
As a start, a number of investors are calculating the carbon footprints
of their portfolios. In September 2014 the United Nations-sponsored
Principles for Responsible Investing (PRI) launched the Montréal Carbon
Pledge, whose signatories commit to the annual measurement and public
disclosure of the carbon footprint of their investment portfolios. The goal
is to attract at least US$3trn of portfolio commitment before the COP21
meeting in Paris at the end of this year. Currently 53 firms have signed
the pledge, including Sweden’s Fjärde AP-fonden (AP4), with SEK295bn
(US$36.3bn) of assets under management, and France’s Établissement
de Retraite Additionnel de la Fonction Publique (ERAFP), with €21bn
(US$23.8bn) of assets under management.
For leading investors, the motivation is there (See box: Motivations
for investors to act), and many are calibrating their response. Mr
Wilson, chief risk officer at Allianz, says that his firm is taking action:
“We are exploring whether climate considerations and the regulatory
implications of climate considerations should be influencing our asset-
by-asset determination, not from an ESG [environmental, social and
governance] perspective, but from an enlightened self-interest with
regard to portfolio returns.”

19
It has really been
made obvious that
carbon is a risk,
and it’s a risk for
business. It’s now
very difficult for
anyone to say ‘Oh, I
didn’t know it was
a risk’.
Philippe Desfossés,
CEO of French pension fund
Établissement de Retraite Additionnel
de la Fonction Publique (ERAFP)
Measurement needs to precede any effective mitigation, but this can take several forms. Dag
Huse, chief risk officer at Norges Bank Investment Management, has been making efforts to
understand the meteorological and geographical models used by reinsurers and how these
models are incorporated into their research and their pricing. "All those geo-models produce
a lot of statistics that the reinsurers use for setting their risk thresholds” he says. “Looking
into this could be important when assessing changes in climate-related risk.” Not surprisingly,
perhaps, the reinsurers are reluctant to share proprietary models.
As part of its investment management activities, Storebrand of Norway has introduced a scoring
method to help drive sustainability in its portfolios. An in-house research team compiles data to
then assign a sustainability score from zero to 100 to each of the 2,500 companies in the firm’s
investment universe. “If the company has a higher rating, it will be a more attractive investment
to different teams, whether it’s a bond team or an equity team,” explains CEO Odd Arild Grefstad.
Most investors appear to make use of ESG data sourced externally, often from specialist third-
party information providers such as Corporate Knights, MSCI, South Pole Group, Sustainalytics or
Trucost. For example, Trucost offers metrics to assess companies’ carbon emissions, embedded
carbon emissions from fossil-fuel reserves, water use, pollutants and waste generation. MSCI
offers ESG indices, including low-carbon indices.
Investors caution that external data on sustainability measures are variable. Some industry
sectors have long reported ESG measures, while others have not. Datasets sometimes combine
reported company data and estimates. “We are constantly searching for good data on
companies,” says Mr Grefstad. “You need to know what you’re doing if you’re combining data
from reported and estimated data and from different types of providers.”
Accurate information is important for all companies, but particularly vital for fossil-fuel
businesses and large energy consumers. Understanding the carbon intensity of their reserves
or their consumption, respectively, ought to be clearly conveyed. Unfortunately, there is
no standardised set of metrics or indicators for the disclosure of climate change-related
information. This is despite calls from some leading investors for comparable and consistent
information.
The fact that information on carbon emissions and climate risks is not readily available in a
comparable form limits the ability of institutional investors to manage the risks within their own
portfolios effectively. Despite an existing array of reporting schemes that address environmental
information, most of these initiatives are voluntary in nature and limited in scope. This is clearly
an area where financial regulators should have a role in laying out clear standards for all market
actors.
Engaged asset managers are developing their own metrics for these issues, often relying on
third parties’ analysis or information to make up for the lack of market standards; however, most
investors have not begun to approach climate risks in any systematic way.

20
_______________
15 The Intergovernmental Panel on Climate Change (IPCC), Fifth Assessment Report. Available
at: https://www.ipcc.ch/report/ar5/
16 Climate Change Scenarios—Implications for Strategic Asset Allocation, Mercer, 2011.
Available at: http://www.mercer.com/insights/point/2014/climate-change-scenarios-
implications-for-strategic-asset- allocation.html
TheSwedishpublicpensionfundFjärdeAP-fonden(AP4)isgoverned
bylegislationthatrequiresittotakesustainabilityintoaccount,
butwithoutgivingupreturns.WhenAP4decidedtoscrutinisethe
climate-relatedrisksinitsinvestmentportfolio,itbeganbymeasur-
ingthecarbonfootprintoftheS&P500index.Thepensionfund’s
managersdecidedthat,astheysoughttomitigatetheclimate-related
riskintheirholdingsintheindex,theydidnotwanttotakeanysec-
torstances—specifically,theydidnotwanttotakeanyexplicitbets
againstfossilfuels.
Instead,AP4keptitsS&P500holdingssector-neutralbydivesting
onlythosestocksthatprovedtohavetheheaviestcarbonfootprints
amongbanks,retailers,automotivemanufacturers,oilcompanies,
coalcompaniesandmore.Bytakingoutthe150worststocksona
sector-neutralbasis,leavingAP4’sholdingsat350,thefirmcontinued
totracktheS&P500withatrackingerrorofbetween0.7%and0.8%.
“Thesurprisingpartofthiswasthatwetookdownthecarbonfootprint
by50%,”recallsAP4CEOMatsAndersson.“Sowewillgetafreeoption
ifcarboniswronglypriced,whichIbelieveitis.”Headds:“IfI’mright,
we’vetakendowntheriskforourclients—theSwedishpensioners—in
termsofclimateriskinourportfolio,withoutjeopardisingreturnsin
theshorttermbecausewegetmoreorlessthesamereturns.”
Afurthersurprisewasinstore.AfteroneyearAP4’scarbon-lightS&P
portfoliohadoutperformedthewiderS&P500bysome70-80basis
points.“Istilldon’tknowwhy,”saysMrAndersson.“Onereasoncould
bethatcompaniesthatlookafterthecarbonfootprintwillprobably
lookaftertherestofthebusinessinthesamemanner.Sothisisactu-
allyawaytopickthegoodcompanies.”
AP4hasnowrolledoutasimilarmodelinEuropeandplanstodothe
sameinJapan.Themanagementplanstodecarbonisethefund’sentire
equityportfoliowithintwoyears.“Ithinkthatwewillendupactually
havingwithinthespanof50%to70%lowercarbonintensity,and
thatiswithoutgivingupreturns,”MrAnderssoncomments.“Onthe
contrary—wegetthisfreeoption.”
A FREE OPTION ON THE MISPRICING OF CARBONINVESTING, DIVESTING
AND ENGAGING
For those investors who recognise and measure the long-term climate-
related risks embedded in their portfolio, there are a number of options.
A clear opportunity is to identify and invest in companies that are better
placed to benefit from a transition to a low-carbon economy—and to shift
funds from those that are more heavily dependent on carbon for returns. In
many cases these investment opportunities have the potential to generate
attractive returns, as well as representing an opportunity for institutional
investors to mitigate carbon-related risks embedded in their portfolio. “The
business case to invest in a transition to a low-carbon economy is becoming
extremely compelling,” says David Blood, managing partner of Generation
Investment Management. “And that’s what’s going to drive change.”
Furthermore, substantial financing is required as the world economy
transitions towards a low-carbon future, presenting opportunities for
investors to mitigate their risks, decarbonise their portfolios and diversify
their returns. These opportunities include areas such as renewable energy
and energy efficiency. “In the current low interest rate environment,
investments in renewable energy infrastructure are an attractive investment
proposition with stable distributions for institutional investors,” says
Tobias Reichmuth, CEO and co-founder of SUSI Partners, a Zurich-based
company specialising in investing in sustainable infrastructure.
Indeed, the IPCC estimates that additional investment of between US$190bn
and US$900bn is required annually in the energy sector alone if the rise in
average global temperatures is to be capped at 2°C.15 Based on International
Energy Agency (IEA) data, Mercer, a consulting firm, estimates that additional
cumulative investment in efficiency improvements, renewable energy, biofuels
and nuclear and carbon capture and storage (CCS) could expand in the range of
US$3trn to US$5trn by 2030.16
Felix Hufeld, president of Germany’s Federal Financial Supervisory Authority
(BAFIN), explains that the principles-based approach of new Solvency II rules for
insurers, due to enter into force in the European Union in January 2016, could
help to promote such investment. “All insurance companies which fall under
Solvency II will have to comply with the famous Prudent-Person Principle, which
does offer you more degrees of freedom to invest your money,” he says. “That

21
The business case
to invest in a
transition to a
low-carbon
economy is
becoming extremely
compelling
David Blood,
managing partner, Generation
Investment Management
does include infrastructure projects, which, from an investment vehicle point of view, are usually the
main reactions to cope with climate change.” As such, Mr Hufeld believes that Solvency II regulation is
actually an opportunity for insurers.
Regulations affecting the asset allocation strategies of pension funds certainly play an important
role, according to Mr Reichmuth. “The funds available for investment in clean energy and energy
efficiency projects could expand greatly if infrastructure were more widely recognised as its own
sector. Legislation in Germany has made this easier, but most countries simply treat infrastructure as
part of ‘alternative investments’ despite their very different risk/return profiles”, he states. This can
impose limitations on investor exposure to certain asset classes.
Leading institutional investors are already positioning themselves favourably to gain from growth
in renewable energy and energy-efficiency activities. Norway’s Government Pension Fund Global,
for example, manages an environmental fund of some NKr50bn (US$6bn) that is largely invested in
alternative energy and energy efficiency; Aviva, the UK- headquartered insurer, is targeting a £500m
(US$780m) annual investment in a low-carbon infrastructure over the next five years; and Allianz of
Germany has committed €2.5bn (US$2.72bn) to renewable energy investments and plans to at least
double its actual exposure in the medium term. France’s AXA Group said in May 2015 that it would
triple its green investments to €3bn by 2020 through green bonds, private equity and infrastructure
investments. “Institutional investors are increasing their exposure to this space, but there is a need
to better commodify the opportunities for clean infrastructure investment,” says Mr Reichmuth.
In March 2015 EU finance ministers approved the regulation for the European Fund for Strategic
Investments, which is set to unlock €315bn, the bulk of which is aimed at financing strategic
infrastructure projects of European interest, including in the energy sector. Public institutions
and development banks are set to shoulder some of the risk, paving the way for investment from
pension funds, insurers and other institutional investors, according to Stephanie Pfeifer, chief
executive of the UK-based Institutional Investors Group on Climate Change (IIGCC). “We have had
a lot of discussions with the EU Commission on how to attract more investment into low-carbon
infrastructure in particular,” Ms Pfeifer asserts.

22
Engagement
Engagement is an important tool offering investors a path towards risk mitigation. In some
cases, investors are removing the worst performers in each sector from their portfolios and are
engaging with the remaining firms, especially where these are local and easily accessible, and
where carbon-intensive activities make up a minority of the investee firm’s activities. Storebrand
has divested from all pure-play coal companies and some pure-play oil firms, for example, but
remains an active owner of a local company involved in oil projects that are not a dominant
part of their business, according to the firm’s CEO, Mr Grefstad. “We’ve been a strong voice,” he
comments.
Institutional investors can use engagement as a tool to reduce the risk that stranded assets may
hamper returns from the shares in their portfolios. Investors can push their portfolio companies
to take action to mitigate their climate risks—wielding influence both through discussions with
management and through the votes that their shareholdings entitle them to. For the growing
volumes of equity funds under management that are managed passively through index tracking
funds, engagement offers an opportunity for investors to bring influence to bear on strategy and
capital decisions, points out Mr Bailey of McKinsey.
“As a company you are going to have to stop investing in further high-carbon assets and start
either returning capital to shareholders and winding the business down, or pivoting towards
lower-carbon business operations,” explains Mr Bailey. “And so the question is what that path
looks like.” As an institutional investor, it becomes all the more important, Mr Bailey says, to
understand the company’s strategy, to engage with management and to identify the moment
when they will need to pivot.
Investors are in a position to push companies to adapt for long-term growth. This includes
advocating that businesses in their portfolios seek profitable means of reducing their carbon
emissions or carbon intensity. This can be done in private meetings or through passing
resolutions at shareholder meetings if a softer touch proves ineffective. “We see businesses
already adapting and already changing their business models because they see that makes
financial sense and they see they don’t have a choice,” Storebrand’s Mr Grefstad says. While
leading companies are responding to these issues, shareholders can exercise active stewardship
throughout their portfolios.
AP4 is one pension fund that very actively engages with the management of the companies in
which it is invested. “When we at AP4 are on a board nomination committee, we always fight for
the board to have enough competence on sustainability,” says Mr Andersson, the CEO of AP4.
“We make sure we’ve got the boards with the competence to actually drive sustainability, and
we make sure we’ve got the management to do it and a strategy that is aligned with what is
possible.”

23
In turn, institutional investors also have scope to influence the lobbying position that these
companies take, points out Ms Pfeifer of the IIGCC. “Are they are in favour of, for example, reform
of emissions trading schemes to support a robust carbon price or not?” she asks. Investors
should have the right to know the lobbying position of companies in their portfolios and, more
importantly, institutional investors can push to ensure that company activities are in accord with
the broader interests of their shareholders.
The opportunities for engagement extend beyond companies held within an asset manager’s
portfolio. Some leading institutional investors are recognising that it is in their own interest and
that of their beneficiaries to advocate policy action on climate change. The IIGCC, for example,
representing US$24trn in assets, has called on the finance ministers of the G7 to support a long-
term global emissions reduction goal in Paris this year aimed at keeping emissions under 2°C.
17
It
has further called for governments to submit short- to medium-term national emissions pledges
and country-level action plans to achieve them.
Divesting high-carbon assets
Divesting has the potential to send a signal to company management and the public more
broadly. While a company may not miss any single investor, Lauren Smart, executive director and
global head of the financial services business at Trucost, points out that “with a critical mass
of investors, there is a strong argument that the cost of capital for that company may go up.”
At the same time, she argues that divestment sends a “very strong public policy signal” that
policymakers and regulators may tighten rules around companies with a heavy carbon footprint,
or else ease regulation or reduce the cost of capital for competitors to those companies, such as
those developing alternative technologies.
But while divestment provides a clear means for investors to publicly express their concerns with
climate change, it forgoes the substantial influence that share ownership provides. It may be
that divestment is the best option only after engagement has failed or where a company’s core
business model is fundamentally reliant on high carbon emissions.
Motivated investors have announced programmes to shed assets that represent significant long-
term climate-related risk or carbon-related risk. Yet, given the significance of carbon-intensive
assets in institutional portfolios, divesting is hardly straightforward. Bloomberg New Energy
Finance estimated in an August 2014 report18
that the stockmarket capitalisation of companies
in the coal and the oil and gas sectors amounts to nearly US$5trn. Oil and gas companies, which
accounted for US$4.65trn, are particularly large and widely held, making wholesale divestment a
complicated process. As a result, asset managers with a mandate to hold a fully diversified portfolio
may struggle to remove this sector from their portfolios entirely.
_______________
17 IIGCC open letter to the Finance Ministers of the Group of Seven. Available at: http://www.iigcc.org/files/publication-files/Investor_CEO_letter_g7_final.pdf
18 “Fossil fuel divestment: a US$5 trillion challenge”, Bloomberg New Energy Finance, August 2014. Available at: http://about.bnef.com/content/uploads/sites/4/2014/08/BNEF_
DOC_2014-08-25-Fossil-Fuel-Divestment.pdf

24
However the stockmarket, value of coal companies in isolation accounts for less than US$250bn,
and their carbon intensity is substantially higher than that of the oil and gas sector, presenting
a clearer opportunity for investors seeking to divest. Indeed, coal is the most carbon-intensive
energy source, and without carbon capture and storage it is fundamentally incompatible with
emissions-reduction initiatives. In May, AXA Group announced that it plans to sell €500m of
coal-related bonds and equities in order to reduce long-term climate-related risk. Norway’s
Government Pension Fund Global also has divested from a number of companies in recent years
following financial assessments that include environmental and social factors. The list includes
companies involved in palm oil production, coal mining and coal extraction for electricity power
production. “It’s around 130 companies that we’ve taken out of our investment universe,” says
Mr Slyngstad of Norges Bank Investment Management. The majority of those divestments are
related to climate issues.
A nuanced approach to reducing the climate-related risks embedded in investment portfolios
is possible – in particular for investors whose funds track equity indices. “Some [investors] are
tilting their funds away from high-carbon assets with best-in-class and screening [approaches]
and changing their benchmarks,” says Ms Pfeifer. Swedish public pension fund AP4 is
decarbonising its portfolio in this way. (See box: A free option on the mispricing of carbon).
THE STATE OF REGULATION
Regulation of the financial services sector has intensified enormously since the 2008 financial
crisis. While national regulation has at times gone further, the Basel III regulatory framework
covering banks set the new baseline. This raised capital requirements, further defined liquidity
ratios and broadened the regulatory and supervisory review process. These efforts have reshaped
wide swaths of the financial services sector in line with a stated goal by the G20 in 2009 to
“generate strong, sustainable and balanced global growth”. At the EU level, Solvency II has
endeavoured to deliver similar outcomes with a framework covering insurers.
These reforms demonstrate clear regulatory concerns with regard to systemic risk. But in their
focus on the excesses that provoked that crisis, most have overlooked an opportunity to address
emerging risks to long-term financial and economic stability. Regulation has largely failed to
confront risks associated with climate change, and in particular long-term climate-related risk
borne by institutional investors. The insurance sector is particularly exposed as greater outlays
will likely be required to cover natural disasters, while at the same time the overall portfolios
which must fund their liabilities will be diminished.

25
While many developed markets have lagged behind,
China and Brazil have arguably been leaders when it
comes to including systemic environmental risk in
their financial regulation. Both have explicitly recog-
nised the materiality of systemic environmental risks
as they relate to long-term financial stability.
Since 2012 the China Banking Regulatory Commis-
sion (CBRC) has required banks to monitor their bor-
rowers’ compliance with environmental regulations
and to begin implementing loan contract changes so
than environmental violations can trigger acceler-
ated loan repayments. Moreover, the CBRC works to
promote bank lending to environmentally sustain-
able economic activities through its Green Credit
Guidelines.
These regulations integrate a range of environmental
and social issues; they were informed and sup-
ported by the World Bank in 2012 and recognise the
Equator Principles as a framework for international
best practice. Funders are requested to collect data
from renewable energy infrastructure and energy
efficiency-related projects, among others, and turn
them over to the regulator
Similarly, in 2014 the Central Bank of Brazil
implemented a regulation establishing guidelines
for financial institutions to consider the degree of
exposure to the social and environmental risk of
their activities and transactions. This not only covers
financial institutions’ own operations, but also the
environmental and social risks within their asset
portfolios. It further requires, at the risk of penal-
ties, that banks publicly disclose their environmental
and social risks.
LEADERSHIPFROMDEVELOPINGMARKETS
_______________
19 Available at: http://ec.europa.eu/internal_market/pensions/docs/directive/140327_proposal_en.pdf
20 Available at: http://www.parlament.gv.at/PAKT/EU/XXV/EU/03/81/EU_38179/imfname_10493724.pdf
Explicit statutory requirements to assess, disclose and mitigate long-term climate-
related risks are largely absent from the regulatory frameworks of most major developed
economies. Japan, for example, has “no regulations associated with climate change
with regard to Japanese insurance companies currently, since Japanese insurance
companies have low exposure to climate change in their portfolios,” states an official
of Japan’s Financial Services Agency, which oversees the country’s banking, securities
and exchange, and insurance sectors. “We have no plans to introduce changes to the
regulatory framework,” the official affirms.
Nor are there clear plans for the European Insurance and Occupational Pensions
Authority (EIOPA),the EU’s insurance and pension fund supervisory body, to further
scrutinise fund managers’ climate-related risks. A March 2014 draft revision19 of the
EU’s 2003 directive on supervision of institutions for occupational retirement provision
included a requirement that institutions produce a risk evaluation for pensions,
including “new or emerging risks relating to climate change, use of resources and
the environment.” Yet, in the September 2014 draft revision this reference had been
scrapped.The final document is due before the European Parliament before the end of
2015.20
In December 2014 a group of members of the European Parliament wrote an open letter
to Mario Draghi, president of the European Central Bank (ECB) and chair of the European
Systemic Risk Board (ESRB)—part of the EU’s European System of Financial Supervisors
that is hosted by the ECB—urging the body to investigate how the exposure to high
carbon investments the exposure to high carbon investments might pose a systemic
risk to our financial system and what the options might be for managing this potential
threat”. As yet, they have received no official response.
Mr Hufeld of BAFIN believes that risks should be captured through the overall framework
of any particular company’s risk-management procedures and rules and regulations—
and that it would not make sense to single out any specific type of risk. “You need to
have a toolbox to supervise those sorts of challenges that is broad enough and flexible
enough to cope with all those different types of phenomenon happening out there,” he
says.
Meanwhile, both listed companies and many financial institutions are subject to
regulations that require them to disclose their material risks. One possibility is that
climate-related risks may be encompassed in those statutes that already govern

26
companies’ general disclosure requirements. A case in point is the US Securities and Exchange
Commission (SEC), which in 2010 issued guidance stating that the SEC requires companies to
report material risks, which would include material climate-change risks.21 While it did not
amend any existing statutes, the SEC did draw attention to the fact that climate-related risks
may fall under existing rules and regulations.
In particular, the SEC has highlighted that there are an array of requirements on businesses
to disclose their risks. The SEC went on to suggest that companies may consider the impact
of legislation and regulation, international accords, indirect consequences of regulation or
business trends, and the physical impacts of climate change. Ceres, a Boston-based non-profit
organisation advocating sustainability leadership, had previously cited breaches of disclosure
rules when it lobbied the SEC for guidance.22
The SEC has clarified that this was not a rule change, nor did it explicitly alter the reporting
requirements of US-listed companies. Critically, this guidance falls short of recognising that
climate change risk is systematically material. This sharply underlines the reality that a great
deal hinges on the interpretation of what constitutes a material risk.
Climate change presents an array of long-term risks; although the precise scope is
understandably uncertain, this research suggests that it will likely be material for all companies.
Perhaps more importantly, dramatically reducing overall carbon emissions will require the
collective efforts of a critical mass of actors. Suffering free riders may undermine the efforts of
the rest and impede the chances of meaningfully mitigating climate risks.
The beginnings of action
In a clear signal that climate-related risks may be a worry for financial regulators, the Bank of
England has launched a probe into the risks that insurers face through climate change. Paul
Fisher, executive director for insurance supervision at the Prudential Regulatory Authority, the
Bank of England’s financial services regulatory arm, warned last March that insurance companies
“may take a huge hit” if their holdings in oil and gas companies lose value because of action to
halt climate change.23
In December 2014, EU Directive 2014/95/EU on disclosure of non-financial and diversity
information entered into force, amending the previous Accounting Directive 2013/34/EU. This
covers disclosure of non-financial information by large companies and so-called public interest
_______________
21 Available at: https://www.sec.gov/rules/interp/2010/33-9106.pdf
22 Available at: http://www.ceres.org/press/press-releases/investors-environmental-groups-push-the-sec-to-require-full- corporate-climate-risk-disclosure
23 “Confronting the challenges of tomorrow’s world”. Speech given by Paul Fisher, deputy head of the Prudential Regulation Authority and executive director for insurance
supervision. Available at: http://www.bankofengland.co.uk/publications/Documents/speeches/2015/speech804.pdf

27
_______________
24 Available at: http://ec.europa.eu/finance/accounting/non-financial_reporting/index_en.htm
Putting a price on
carbon is the best
way to incorporate
all of these risks
where carbon is
concerned into the
investment process
Al Gore,
chairman, Generation Investment
Management
entities (including banks and insurers) across the European Economic Area. This amendment
requires disclosure of non-financial information including information relating to policies, risks
and outcomes as regards environmental matters. EU member states must transpose the directive
into national legislation by the end of 2016. 24
In other cases, public sector action takes the form of increased requests for information, rather
than compliance with specific regulation. For instance in the US, several states conduct an
annual Insurer Climate Risk Disclosure Survey, which was created by the National Association of
Insurance Commissioners and is mandatory for most sizeable insurers. The survey asks insurance
companies what financial risks they face from climate change, as well as actions they are taking
to respond to those risks. Furthermore, financial examiners in the US have been given guidelines
so they may ask insurers about their exposure to climate-related risks, says Mike Kreidler,
Washington’s insurance regulator and co-chair of the NAIC’s Climate Change and Global Warming
Working Group.
And most significantly, in May, France’s National Assembly voted on a series of amendments of
the country’s Energy Transition Law - providing a first glimpse of a statute explicitly covering
climate-related risk. One amendment will require French insurance firms, pension fund managers
and other institutional investors to disclose “information on the consideration of environmental,
social and governance factors in their investment policy criteria.” Investors will also be required
to explain how they “take into account exposure to climate risks, including the measurement of
greenhouse gas emissions associated with assets held” in their portfolios.
While effective regulation to combat climate change has been largely absent to date, regulatory
attention is increasing and the new legislation in some quarters is beginning to focus more
sharply on climate-related risks. However, while awareness of the issue is growing among
financial regulators, few are taking action. This is despite the fact that few institutional investors
have addressed this risk to date; just 7% are able to measure the carbon footprint of their
portfolios and a mere 1.4% have an explicit target to reduce it. Regulators should require that
companies disclose their carbon emissions so than investors can assess their risks with accurate
and comparable data.

28
Over the long
term we have a
self-interest in
well-functioning
markets
Odd Arild Grefstad,
CEO, Storebrand Group
A price on carbon
Many market participants insist that carbon pricing is essentially what is needed, among them
Al Gore of Generation Investment Management. “Putting a price on carbon is the best way to
incorporate all of these risks where carbon is concerned into the investment process,” he says.
The best way to tackle the externality of carbon pollution “is to attach a price tag so that it can
be more smoothly integrated into the routine assessment of asset values that people conduct
and analyse in their portfolios on a constant basis.” While clearly accurate, emissions trading
schemes have yet to deliver on expectations: getting the pricing right will be crucial.
It is clear that government action is required to establish a firm, clear carbon price that
reasonably reflects its externality costs. It is the responsibility of governments to correct market
failures, and climate change is potentially the world’s most important market failure. Without
an appropriately functioning pricing mechanism it is incredibly difficult for climate risks to be
addressed and for capital to be effectively allocated. This requires rigorous carbon taxation or
carbon trading schemes.
In many jurisdictions this is already happening, either through carbon taxes or cap-and-trade
schemes. To date, emissions trading schemes have not always lived up to their expectations
and in many instances the price set for carbon emissions is too low to meaningfully capture the
negative externalities associated with climate change. This is currently the case of the European
Emissions Trading System, the world’s largest scheme, where the price of a tonne of carbon
languishes below US$10. For his part, Mr Clamagirand of AXA Group says: "It is important that
the regulators start considering a more realistic – i.e. higher – carbon price. This will then allow
market participants to start incorporating it into their valuation models.”
Reasonable assessments of the price needed to meaningfully address emissions have generally
ranged above US$30 per tonne. However, the challenges of several current schemes have more
to do with a tendency of governments to oversupply the market, concede free quotas or provide
other loopholes than with inherent failures of cap-and-trade as a system. Carbon taxes, once
considered politically unpalatable, have also been implemented in markets ranging from Chile to
British Columbia.
The choice of a carbon tax or a carbon trading scheme is less important than the need to ensure
that a price mechanism is established, commensurate with the negative externalities that
climate change is expected to bring. The inherent uncertainties and long-term nature of the
problem make this difficult. However it is clear that for these policy measures to be effective,
they must be firm, long-term and comprehensive. Moreover, by establishing a framework
whereby the stringency will predictably increase over time, market actors will be able to respond
while ensuring that mitigation measures are carried out in a cost-effective manner.

29
Establishing this pricing signal is crucial for companies and investors to properly incorporate
climate-related risks into their decision-making. But international cooperation to combat free
riding and encourage greater action is similarly vital. The Paris Conference of the Parties due to
take place at the end of this year will have wasted an important opportunity if concrete measures
to price carbon emissions do not emerge. The precise tools employed may vary, but the direction
of travel needs to be firmly set by governments at a national level and strongly reinforced though
international agreements.
“The market is the most efficient way to allocate a rare resource, which is capital. If we think
that’s the case, why is it not working?” asks Philippe Desfossés, CEO of ERAFP. “It’s not working
because the market is not getting the right signals, and obviously we know that the problem is
linked to the fact that this negative externality that is carbon is not priced.”
Correcting market failures
Regulation is required to address market failures, and the negative externalities of climate
change clearly constitute a market failure. Moreover, addressing climate change is clearly a
problem of collective action. “While responsible industry players will make commitments this
year to contribute to the transition to a low carbon economy, these efforts may not reach the
necessary scale,” warns Mr Clamagirand of AXA. “It will always be a story of a few responsible
actors doing their best within a broader financial system that is not fully designed for
sustainability.”
While pricing carbon emissions effectively is vital, complementary policies are necessary to
mitigate climate risks. Indeed, reforms to the financial system may be needed to facilitate an
orderly transition even in an economy in which carbon is effectively priced. “You’ll still need
disclosure requirements and markets, you’ll still need to make sure that fiduciary responsibilities
are aligned with climate security or that solvency rules are actually properly refined to enable
long-term allocations to the green infrastructure, for example,” states UNEP’s Mr Robins.
Moreover, Mr Grefstad of Storebrand argues that regulation helps create a level playing field
for investing companies. By imposing standards of behaviour on asset managers uniformly,
regulation has the potential to provide industry with incentives to step up their fight against
climate change.
Mr Grefstad argues further that regulation would be an advantage for the financial industry in
general. “Predictability is good for finance because it’s easy to create returns when you have
predictable financial markets,” he points out. “Climate and other resource and ecosystem
changes are affecting that predictability.” Mr Grefstad reasonably concludes that “over the long
term we have a self-interest in well-functioning markets.”

30
Governments must enact comprehensive
carbon-pricing mechanisms that reflect their
externality costs. Lacking a realistic price, efforts
by market participants to limit climate change
are inherently handicapped. Addressing market
failures is fundamentally the responsibility of
governments.
Financial regulators need to ensure that
best practice becomes standard practice. This
means recognising climate risk as material and
requiring the disclosure of carbon emissions
by market actors. Standards for comparable
information are necessary to identify free riders
and concentrations of systemic risk.
Stock exchanges should require disclosure of
greenhouse gases by all listed companies. Clear
accounting of carbon intensity is needed. Without
accurate information, integrated into financial
reporting, investors cannot manage their risks
appropriately.
Institutional investors must integrate climate
change into their risk management. Assessing
and measuring the risks in their own portfolios
is a necessary first step. This can lead to
adjustments in investment strategy or to deeper
engagement with company managers. Advocating
that policymakers address market failures is in
their collective self-interest. Complete inaction is
a failure to act in the long-term interest of their
beneficiaries and could risk future litigation.
Pensioners should insist that the fund
managers responsible for their retirement savings
seriously address the full spectrum of long-term
risks they are facing. Concrete measures vary
from promoting corporate engagement and
public policy advocacy to potential legal action
demanding that fiduciary responsibilities be met.
RECOMMENDATIONS FOR ACTION
“The market is very good at innovation. But it doesn't deliver universality. So there is a role
for policymakers to identify where there are market failures, identify best or good practice
and turn it into customary practice," affirms Mr Robins. "We have extensive experience
with voluntary disclosure, for example, but the market on its own will not deliver universal
reporting,” he says. “That’s where regulation can step in.”
Regulators should explicitly recognise climate risks as material. This means that listed
companies should be required to assess their own carbon emissions and disclose them.
Institutional investors need to analyse the climate risks in their portfolios and actively
look for ways of mitigating their exposure. This may be through shifts in their investment
strategy or engagement with company management, but simply ignoring climate change
risks failing to meet their fiduciary obligations.
One aspect of this is more effective co-ordination among the institutions that govern
the world’s financial systems, including regulatory institutions and international bodies.
“We need really to have better systems for risk management. We need to integrate issues
like climate change, particularly into the prudential regulation of financial institutions,
banks, insurers, pension funds,” according to Mr Robins.
The COP21 meeting, due to take place in Paris at the end of this year, may provide
further impetus. The objective of the 2015 conference is to reach a binding and
universal agreement on climate change by all participating nations. Regardless of the
precise outcome, it is important that this chance to bring together a range of different
stakeholders all concerned about the impact of climate change not only on themselves but
on society and the economy more broadly, must not be wasted. Leading organisations need
to do more than change their own investment strategies: they need to campaign both for a
price on carbon and for a more level regulatory playing field.

31
CONCLUSION
The potential impact of climate-related change on the assets owned and managed by institutional
investors is significant. Our estimates indicate that asset managers can expect present-day losses
of US$4.2trn to the US$143trn of current manageable assets as a result of climate change by
2100 at a private-sector discount rate, equivalent to the entire GDP of Japan. It is important
to recognise that this is not a risk of volatility or temporary price movements but of permanent
impairments and capital losses. Perhaps more seriously, the expected losses would more than
triple (US$13.8trn) should global warming reach 6°C. Few institutional investors have addressed
this risk to date; only a modest minority are even able to measure the carbon footprint of their own
portfolios.
Governments have no excuse for inaction. The potential harm from a public-sector point of view
is significantly above the private-sector average expectation should some of the more extreme
outcomes be realised with present value damages of US$43trn consistent with a 6°C scenario.
While the likelihood of that much warming is low, the results would be catastrophic. Regulation has
not generally required owners and managers of assets to measure, disclose or manage the climate-
related risks embedded in their own portfolios, despite the fact that these risks are both material
and systemic. However, regulators should take note of the bold action taken by France, which in
May of this year amended the law requiring asset managers to do so.
Even with little regulatory pressure, a number of leading institutional investors are beginning to
address climate-related risks. Some are making efforts to engage with policymakers, encouraging
them to address this market failure. A few are measuring the carbon footprint of their investment
portfolios. Others still are divesting investments in carbon-intensive industries or engaging with
companies to exert influence over their management. And more excitingly, a wide array of asset
managers and owners are finding profitable opportunities to finance the transition to a lower
carbon economy, including investments in renewable energies, new technologies and energy
efficiency.
Despite the efforts of the private sector, there is substantial scope for regulators and governments
to make a contribution. Not least, many market participants are looking for agreement on a
meaningful price on carbon at the COP21 meeting, due to take place in Paris at the end of the year.
However, even if a substantive agreement were to emerge, there is a clear need for co-ordinated
action by regulators, governments and institutional investors in order to address the long-term,
systemic risks at play. Climate risks need to be assessed, disclosed and, where feasible, mitigated.

32
APPENDIX
APPENDIX
THE VALUE OF GLOBAL ASSETS UNDER MANAGEMENT AT RISK DUE TO
CLIMATE CHANGE
Final Technical Report
1. Introduction
Climate change can affect the asset management industry by putting assets at risk of direct
destruction from events such as storms and floods, or by indirectly affecting asset returns.
While climate change has long been seen as a threat to economic activity, the risks it poses to assets
are not well understood. This is partly because the outcomes of climate change are subject to large
uncertainties, and partly because the issue has not been brought to the attention of the industry. Yet
the industry exists to manage risk, and has a commonly-employed measure to assess it: value at risk
(VaR). In a pioneering attempt to highlight the relevance of climate change to the asset management
industry, this report estimates the value at risk of global assets under management (AuM) due to the
impacts of climate change, or the climate VaR.
VaR measures the size of the loss a portfolio may experience, within a time horizon, at a particular
probability or confidence level, p. So, if for instance the VaR of a portfolio is $1bn over a one-year
horizon at a confidence level p=95%, that means there is a 5% probability that the loss on the portfolio
is at least $1bn. The climate VaR is the loss that can be attributed solely to the impact of climate
change on the stock of assets. That is to say it compares the value of assets in a world with climate
change relative to the same world without climate change.
VaR is a natural way of thinking about the impacts of climate change. This is because there are
great uncertainties in the estimation of climate impacts (IPCC, 2014b; Millner, Dietz, & Heal, 2013;
Tol, 2012; Weitzman, 2009). For example, future greenhouse gas emissions are uncertain, as is the
climate response to emissions, and the impact on the economy of climate change. Furthermore, these
uncertainties feed into each other, creating a ‘cascade’ of increasing uncertainty. This report extends
typical analysis of climate impacts, to consider the impact of climate change on assets, rather than
the more commonly studied impact of climate change on GDP. Therefore uncertainty must be clearly
acknowledged, which VaR does by presenting possible losses at certain levels of probability, so is an
appropriate metric for the impacts of climate change on the asset management industry. In particular,
by considering events that occur ‘at the tail of the distribution’, such as events with a 5% or 1%
probability of occurring, a VaR estimate focuses on the high impact, low probability outcomes that
could result from climate change.
The climate VaR in 2100 is the focus of this report. Climate change is a long- term problem, with a
considerable lag between the release of emissions and economic impacts. The full consequences of
greenhouse gases emitted today will not be felt until around the end of the century. While this date
may seem distant for many in the industry, the scale of future impacts, even in present value terms, is
likely to be significant, and it is only by taking action today, far in advance of the worst effects, that we
can prevent them from occurring.
This report focuses on the broader stock of manageable assets, rather than assets currently under
management. Manageable assets are those the industry could potentially manage, rather than the
share that is currently managed, which is estimated to have been 22-25% of the potential market in
recent years (McKinsey & Company, 2012). The report takes this focus because its aim is to consider
how the industry’s overall prospects are affected by climate change, rather than on its ability to attract
customers and increase its market share in the face of climate change.
Written by

33
APPENDIX
2. Methodology
2.1 General approaches
The impact of climate change on the financial sector, let alone the asset management
industry, has been the subject of limited research to date (Arent et al., 2014). The sector
suffers from limited awareness of, or interest in, the issue (Covington & Thamotheram, 2014)
and few modelling frameworks exist to assess the impacts (Vivid Economics, 2013). Indeed, the
Prudential Regulation Authority in the UK was motivated by this lack of knowledge to request
information from industry in June 2014 (Prudential Regulation Authority, 2014). However the
concept of climate VaR has recently been introduced and demonstrated in a simple, yet powerful,
way (Covington & Thamotheram, 2015).
There are two types of approaches that could be taken to estimating the climate VaR:
bottom-up or top-down. A bottom-up approach would be built around a relatively detailed
portfolio analysis model, which takes as its input various kinds of macroeconomic variables
and goes on to model the returns to different asset classes in different countries or regions.
This approach can be described as bottom-up1 because the analysis of climate VaR must be
built outwards by situating the portfolio analysis model within a system of other, linked models
capable of providing it with the inputs it needs. In particular, this would include the effects of
climate change on economic outcomes, as well as, perhaps, the direct effect of climate change
on portfolio performance, via its effect on the co-variances between assets (in this way, the
interaction between climate risk and other sources of risk is taken into account).
A bottom-up approach is model- and data-intensive, and it is unclear whether the
uncertainty inherent in providing this high level of detail would provide particularly accurate
estimates. One of the principal reasons for this is the famously poor state of knowledge of
climate impacts (Pindyck, 2013; Nicholas Stern, 2013). For many of the causal processes in need
of estimation as part of the bottom-up approach (for example, the effect of climate change on
the covariance between equities and corporate bonds in the United States), there are simply
no data, and it is unclear how to obtain convincing data. Given this, and the need, at this point
in time, to explore the order of magnitude of climate impacts on AuM rather than the detailed
consequences, a simpler, albeit less detailed, top-down approach is valid.
A top-down approach uses a simple macro-economic model that has been integrated with
emissions and climate modules. A small number of so-called integrated assessment models
(IAMs) of climate change exist, which have been built in order to estimate the economic cost of
climate change. These link economic growth, greenhouse gas emissions, climatic changes and
damages from climate change back to the economy in an integrated, consistent framework.
________________
1 Bottom-up is not to be confused with other senses in which the term bottom-up modelling has been used in related fields, for instance energy modelling, where it
indicates that the model is built up from representations of individual energy technologies.

34
APPENDIX
Famous examples include the DICE model of William Nordhaus (1993, 2008, 2014) as well as
the PAGE model of Chris Hope (2006, 2013), which was used by the Stern Review (2007). These
models have been built with public -policy objectives in mind, above all the quantification of the
externality that climate change constitutes, for example the Stern Review’s estimate that the
costs and risks of climate change will be equivalent to losing 5-20% of global GDP each year, now
and forever. This enables an assessment of what reasonable costs should be incurred to bring
climate change under control, in total, such as the percentage of GDP to spend on mitigation, or
at the margin, such as the level of a carbon tax per tonne of CO2.
IAMs are not without their limitations. IAMs can be, and are, fairly criticised for not describing
the full range, richness and uncertainty of scientific knowledge regarding climate impacts,
for making value judgments, particularly in choices over discount rates, and for not using a
sufficiently sophisticated representation of the economy (King, Schrag, Dadi, Ye, & Ghosh, 2015;
Vivid Economics, 2013). They are also generally considered to be conservative in their estimates
of damage. In this report we take great care to acknowledge and incorporate the scientific
uncertainty, in particular regarding catastrophic climate change, lay out the value judgments
that are made and test different discount rates, and extend the economic sophistication of the
analysis to consider financial assets. Despite this, we acknowledge that the modelling faces
limits and that the results should be interpreted as a guide to the likely magnitude of impacts.
However, IAMs remain the only way to quantify the aggregate economic impact of climate
change within a consistent framework. There are no alternative modelling tools yet developed
to quantify the economic cost of climate change within a consistent framework, as is required for
this project (Vivid Economics, 2013). Other studies of the impacts of climate change in general,
and especially work on the effects on financial assets, do not quantify the economic impacts; for
examples see (AVOID, 2015; King et al., 2015; Mercer, 2015). Often the focus is on the physical
impacts, such as the extent of flooding, or non-monetary losses, such as health impacts, or a
qualitative discussion of economic impacts is provided. Such approaches avoid the criticisms of
IAMs but also forgo the expositional benefit of aggregating impacts into a single dollar value, the
magnitude of which can be easily understood – and these are currently unavoidable trade-offs.
While IAMs have not before been used to consider the impact of climate change on assets,
they contain useful information to estimate the climate VaR. Depending on the precise
approach to estimation, IAMs can estimate the impact of climate change on two parameters
useful for estimating the climate VaR: i) GDP growth, which can be linked to the growth of
dividends from manageable assets, and ii) the stock of capital. This gives rise to two approaches:
the dividend approach and the capital approach.

35
APPENDIX
These two approaches address different issues: the present value reduction in future
dividends, in the dividend approach, and the effect of climate change on the future capital
stock, in the capital approach. They are not alternative measures of the same thing. The
dividend approach considers the effect climate change will have on the growth of dividends from
the current stock of manageable assets, which was $143 trillion in 2013,2 and the consequent
reduction in the value of this stock, given that its value equals the discounted flow of future
dividends. The capital approach considers the effect that climate change will have on the future
stock of non-financial assets,3 the economic capital used to produce output, which underpins
many of the financial assets managed by the industry. These approaches are explained in more
detail in the following Sections 2.2 and 2.3.
DICE is the most suitable model for our purposes. Among IAMs set up to estimate climate
impacts, DICE is one of the very few based on the so-called ‘Ramsey’ growth model.4 This makes
it suitable for the task at hand as it has two relevant features that other established IAMs,
such as FUND and PAGE, do not. First, the growth rate is determined within the model rather
than assumed, which means that the effect of climate change on the growth rate is explicitly
modelled. Second, DICE explicitly models the capital stock, which underpins the stock of
manageable assets. DICE is widely used in academia and policy-making, in its original form and
as the basis for other models, such as WITCH, and has been extensively peer-reviewed.
The impact of climate change on assets is distinct from ‘stranded assets’, which describes
an effect of mitigation policy on assets. Recently, much attention has focused on the effects
of climate change mitigation policy on the financial sector, as policy to limit emissions may
prevent the currently -expected value of emission-intensive assets from being realised, thereby
‘stranding’ the asset. However, this is not the focus of this report, which is instead on the effect
on assets of the physical impacts of climate change, should mitigation policy not succeed in
limiting emissions.
________________
2 Our value for the stock of manageable assets is the stock of assets held by non-bank financial institutions, as estimated by the Financial Stability Board. Bank assets
are excluded as these are, largely, managed by banks themselves.
3 Our value for the stock of non-financial assets is based on the stock in DICE, the model used to estimate the results of this report, with additional refinement based
on EIU data.
4 More precisely, this model has a lineage that begins with Ramsey (1928) and includes modern developments by Cass (1965) and Koopmans (1965), among others.

36
APPENDIX
2.2 The dividend approach
In a new paper, Covington & Thamotheram (2015) present an approach to estimating the
climate VaR using information on GDP growth under climate change. They calculate AuM as
the present value of the flow of future dividends, and make a key assumption: in a diversified
portfolio of equities, the growth rate of dividends is exactly equal to the growth rate of GDP.
Then, drawing on other studies, they calculate the loss of GDP due to climate change at a range
of confidence levels. Through the relationship between dividends and GDP, they then calculate
the impact of climate change on assets at particular confidence levels, thereby estimating the
climate VaR. Their work suggests that, in their worst case, ‘the value at risk in 2030 may be
equivalent to a permanent reduction of between 5% and 20% in portfolio value compared to
what it would have been without warming’.5
Our dividend approach builds on this paper. We take Covington & Thamotheram’s key insight,
that there is a relationship between the growth rates of absolute dividends and GDP, and use
this relationship to find the change in dividends arising from the change in GDP due to climate
change, which DICE models directly. We assume that dividends from manageable assets grow
at the same rate as GDP. This assumption, the same as Covington & Thamotheram’s, is made
because dividends are ultimately paid for from the output of the economy. So, if the share of
output paid in dividends is, on average, constant in the long run, then GDP and dividends will
grow at the same rate. This relationship may not be observed in data over a relatively long time
period, even decades, due to business cycles; for example corporate profits are currently at
historic highs while GDP growth is low. However, to 2100, the relationship should be expected to
hold on average.
We provide two main advancements on Covington & Thamotheram’s estimate. First, we
estimate climate and GDP outcomes within an integrated framework, to give a consistent and
more rigorous estimate of the VaR at different confidence levels. Second, as described below, we
consider a broader portfolio of assets than equities.
The dividend approach considers manageable assets to be all non-bank financial assets.
We take a broad definition of manageable assets, to cover all assets held by non-bank financial
institutions, which was $143 trillion in 2013, as estimated by the Financial Stability Board
(2014), as all these assets could, potentially, be managed. Bank and Central Bank assets, on the
other hand, are largely managed by the institutions themselves and so are excluded. This broad
definition not only provides a more accurate estimate of the magnitude of impacts, but it also
makes the assumption that dividends from manageable assets grow at the same rate as GDP more
appropriate, as the broader definition controls for cyclical differences in relative performance
across asset classes.
_______________
5 However, when discounting their results to 2015 present values, Covington & Thamotheram find that the expected present value at risk in 2015 in their worst case is one-fifth of the value
at risk in 2030: that is to say, it is equal to a permanent reduction of between 1% and 4% in portfolio value.

37
APPENDIX
The dividend approach estimates the present value loss of future dividends from the current
stock of manageable assets. This impairs the value of the current stock in 2015, given that
its value equals the discounted flow of future dividends. As the dividend approach considers
effects to a portfolio of manageable assets, the loss of future dividends is discounted from the
perspective of a private investor, although it can also be discounted using government discount
rates. These issues are discussed further in Annex 5.4.
2.3 The capital approach
The capital approach explicitly estimates the impact of climate change on the stock of assets.
As mentioned, DICE explicitly models the stock of non-financial assets, known as ‘capital’ in the
economics literature, so the impact of climate change on this stock can be estimated. In DICE,
a generic capital asset is created when households save, so a stock of assets is built over time
as savings accumulate, and declines with depreciation. Assets earn a return, which incentivises
savings. In this report, the initial stock of non-financial assets is based on the stock in DICE,
updated based on EIU data, and is $207 trillion in 2015. In our version of DICE, climate change
affects the stock of non-financial assets in two ways: by lowering the rate of return and by
directly destroying assets.
Climate change indirectly affects the stock of non-financial assets by lowering investment.
The impacts of climate change drive a wedge between the output that could potentially be
realised with given capital and labour stocks, and actual output. This lowers the output produced
per unit of asset, thereby reducing the attraction of investment, which leads to a lower stock
of assets. For example, climate change may destroy a fraction of the output of a farm, and as a
result the assets employed on the farm will earn a lower return, meaning that the farmer has less
incentive to hold a large stock of assets. Furthermore, since less output is available, households
will have less to save, and therefore the stock of assets will build more slowly.
Climate change directly affects the stock of assets by destroying them. Dietz and Stern (2015)
propose an extension to DICE, whereby, as well as driving a wedge between actual and potential
output with capital and labour inputs given, a portion of damages also directly reduces the
capital stock. An example of this kind of effect is storm damage to infrastructure. This extension
is incorporated in the model used in this report.
Impacts to non-financial assets can be converted to impacts to manageable assets.
Non-financial assets can be used to back financial liabilities. For example a corporation can take
out a loan secured against a factory, or the flow of income the factory will provide. The financial
liability of a debtor is a financial asset for a creditor, and this financial asset could be managed.
So, if the stock of non-financial assets is reduced due to climate change, the stock of manageable
assets will also decrease. To estimate the impact on manageable assets from the reduction in
non-financial assets, two conversions are required. First, the share of non-financial assets that
are used to back financial liabilities must be estimated.6 Second, the financial liabilities created
per dollar of non-financial asset must be estimated. The product of these two ratios converts a
dollar of lost non-financial asset to a value of lost manageable assets.
________________
6 Not all types of non-financial assets are typically used to back financial liabilities. For example, public infrastructure is rarely used to back government liabilities, and non-financial non-
corporate businesses, such as sole traders, rarely issue manageable securities.

38
APPENDIX
However, results for the capital approach are presented as a percentage reduction in the
capital stock. As explained in Annex 5.1, this conversion does not cover all manageable assets
and the ratios may vary over time. As a consequence, results in the capital approach are not
converted into losses of manageable assets, but presented as a percentage reduction to the
capital stock. If manageable assets are, on average, as vulnerable to climate change as the
capital stock in general, then a similar percentage reduction to manageable assets would be
expected.
3. Key issues in the estimation of the climate VaR
Estimation of the climate VaR raises a number of issues. Before proceeding to the results, it is
important to describe a number of key issues in the estimation of the climate VaR. These issues
concern: uncertainty, long- term prospects for productivity growth, discounting and mitigation
scenarios. A short summary of these issues and their treatment in the modelling is provided
here, with extensive discussion provided on each in the Annex.
The probability of a loss in the climate VaR is generated by uncertainty over three factors:
the level of productivity growth, the climate sensitivity and the risk of catastrophic
climate change. There are many uncertainties in the estimation of climate impacts. However,
the academic literature has identified three parameters to which outcomes are particularly
sensitive. The first, the level of productivity growth achieved in the initial decade, which sets
the magnitude of growth for the century, reflects uncertainty over general macroeconomic
conditions. After the initial decade, the trend then evolves according to a scenario: its growth
rate either increases or decreases, explained below. The second, the climate sensitivity, defines
the temperature increase for a doubling of greenhouse gas concentrations, which is uncertain
due to the complexity of the climate system. The third, the risk of catastrophic climate change,
reflects the divergence of views in the academic literature on the possibility of catastrophic
impacts beyond a certain degree of warming. In this report, a probability distribution for each of
these three parameters is defined using the best available evidence and Monte Carlo analysis is
used to estimate impacts across the range of confidence levels. Further discussion is provided in
Annex 5.2.
Scenarios are used to describe different prospects for long term productivity growth. The
long term growth rate of productivity can either be increasing over time, so the global economy
continues to grow at a relatively high rate, which is the EIU’s view; or decreasing, so global
growth slows, which reflects the idea of ‘secular stagnation’, the default setup in the DICE model.
We use the EIU’s projections of increasing productivity growth as the base case. However, as
these alternatives present different prospects for future economic growth, and therefore the
stock of assets, the secular stagnation view is tested as an alternative scenario. So the trend in
productivity growth is treated as uncertain, as described above, but then whatever the trend, it
either increases or decreases in the long run, depending on the scenario. Further discussion is
provided in Annex 5.3.

39
APPENDIX
The dividend approach is discounted from both the perspective of a private investor and a
government. As the dividend approach considers effects to a portfolio of manageable assets,
a discount rate is applied as is standard practice for a private investor. The discount rate is a
function of the future GDP growth rate without climate change plus a premium to account for
holding risky assets relative to a risk-free asset. The GDP growth rate without climate change is
used as investors do not currently consider climate impacts. This results in a more conservative
estimate of losses, as GDP growth with climate change will be lower, leading to a lower discount
rate if investors took this into account. Average private sector discount rates are initially 5.5%
but fall to 4% towards the end of the century, due to slowing economic growth.
The government discount rate used in this report follows the approach of the Stern Review.
In addition to discounting from the perspective of a private sector investor, government discount
rates are also applied to the dividend approach. This sensitivity is explored because governments
have a duty to safeguard financial assets on behalf of society and regulate to fulfil this duty,
so the value at risk from a government perspective is also relevant. It is widely accepted that
public-sector discount rates should be below private-sector discount rates, for reasons including
a lower cost of capital as well as the avoidance of market distortions, such as taxation and
externalities, which mean the gross returns on private investments are above their net social
returns. The government discount rate used in this report follows the approach of the Stern
Review. This approach is appropriate for very long-run problems such as climate change, because
it treats the wellbeing of future generations on par with the wellbeing of current generations
with some included uncertainty. The remaining justification for discounting in this framework
is economic growth, and its effect of reducing the marginal utility of consumption for future
generations. Average government discount rates are initially 3.8% but fall to 2% towards the
end of the century, due to slowing economic growth.
The capital approach provides undiscounted, future values. Results are not discounted in the
capital approach, as the focus is on the state of the future capital stock, for which the calculation
of present value equivalents is conceptually difficult.7 Further discussion on discounting is
provided in Annex 5.4.
A mitigation scenario consistent with a likely chance of remaining under 2°C of warming
is compared with the base case of minimal mitigation. The base case is a scenario of minimal
mitigation, as the climate VaR is the loss that can be attributed solely to the impact of climate
change on the stock of assets, rather than the impact of (less) climate change and (more)
mitigation. That is to say the base case isolates the effect of climate change by comparing the
value of assets in a world with climate change relative to the same world without climate change;
where both worlds have the same level of minimal mitigation. This analysis can be added to by
finding the climate VaR in a world with high mitigation (and therefore less climate change than
the base case) relative to the same, high mitigation, world without climate change. These two
________________
7 This difficulty arises because the value of the future capital stock cannot be discounted as if it were a private investment, unlike the future flow of dividends from
manageable assets, as it contains items such as public infrastructure. The appropriate discount rate for the capital stock is therefore complex. On the other hand,
presenting future, undiscounted, values in the capital approach is simple and clear – and therefore utilised.

40
APPENDIX
climate VaRs, the base case and the mitigation scenario, can then be compared to show how
the climate VaR is reduced due to mitigation. The base case has a minimal level of mitigation,
an emissions reduction of 6.5% in 2105 relative to a world of no action whatsoever, to reflect
existing policies, but can otherwise be characterised as a scenario of complete future inaction
and stalemate on climate policies. The mitigation scenario is consistent with a likely chance
of keeping temperature change below 2°C, the most stringent mitigation scenario the IPCC
considers to be feasible (IPCC, 2014a), where a likely chance is defined by the IPCC as a 66%
probability of occurrence. Further discussion is provided in Annex 5.5.
4. Results
The dividend and capital approaches estimate distinct climate VaRs. These approaches are
not alternative measures of the same thing, but consider two issues that are of common concern
to the asset management industry. The dividend approach considers the effect climate change
will have on the growth of future dividends from the current stock of manageable assets and the
consequent reduction in the value of this stock, given that its value equals the discounted flow
of future dividends. The capital approach considers the effect that climate change will have on
the future stock of non-financial assets, which underpins many of the financial assets managed
by the industry. Due to this distinction, the results should be considered as measures of different
impacts and not as alternatives, and are presented separately, with primary emphasis placed on
the dividend approach as this is the approach presented in the EIU paper. The secular stagnation
productivity scenario is also presented separately, as a sensitivity to the base case assumptions.
The losses are primarily incurred in the latter half of the century. Figure 1 shows how the
presentvalue loss to current assets evolves over time along temperature consistent paths, as well
as the average loss over time. To illustrate, if we were only concerned with losses over the next
40 years, to 2055, and believed that the world was on a path to 6°C of warming by 2100, then
the consistent present value loss to current assets to 2055 would be $2 trillion (discounted using
private sector rates); this is where the red line crosses the lowermost y-axis gridline at 2055.
Losses after 2055 would of course still occur, but, given the 40 year time horizon, they would,
in this example, be disregarded. As the figure shows, the losses are primarily incurred in the
latter half of the century, as there is a lag between the release of emissions and temperatures
increasing. It is only later in the century that the climate consequences of the majority of
emissions are expected to be felt, although there is scientific debate over whether the climate
response will be faster (Frölicher & Paynter, 2015).

41
APPENDIX
4.1 Dividend approach results
Climate change is expected to cause permanent, present value losses to current
manageable assets of 3% on average and up to 10% at extreme outcomes, when
discounted at a private sector discount rate. Climate change is expected, on average,
to cause $4.2 trillion of present value losses to current assets, equal to 3% of current
assets, when discounted at a private sector discount rate. Losses are much higher if more
extreme outcomes are expected to occur, with $7.2 trillion and $13.8 trillion of present
value losses consistent with warming of 5°C and 6°C respectively. This means that private
investors are overvaluing manageable assets today by 3% of their current value if they
consider the expected average impacts of climate change; while if they believe that 5°C or
6°C of warming will be reached by 2100, then assets today are overvalued by 5% and 10%
respectively.
From a government perspective, current manageable assets are on average 10%
overvalued, and up to 30% at extreme outcomes. When using Stern Review discount
rates, which are lower than private sector discount rates, the mean average expected
loss is $13.9 trillion, equal to 10% of current assets. At more extreme outcomes, losses
from the government perspective are very severe. If 6°C of warming is expected, then the
consistent level of present value losses at government discount rates is $43 trillion, equal
to 30% of current assets.
These results are summarised in Table 1 and an explanation of the average and
temperature consistent losses terminology is provided in Box 1.
Present value loss to current
manageable assets
Average
loss
5°C
loss
6°C
loss
Private investor
perspective
$4.2 trillion $7.2 trillion $13.8 trillion
Government
perspective
Table1-Thepresentdayvalueoflossestocurrentmanageableassetsduetoclimatechange
is3—10%onaveragedependingondiscountrate,and10—30%at6°Cofwarming
Note:Thecurrentstockofmanageableassetsis$143trillion.Dollarvaluesarein2015dollars.
Lossesarethecumulativelosstomanageableassetsto2100discountedtothepresentusing
privatesectordiscountrates(5-7%,decliningwiththegrowthrate)andgovernmentdiscount
rates,aspertheSternReview,respectively.
Resultsforextremeoutcomesarepresentedasdamages
consistentwithtemperaturelevels.Theoriginalpurpose
ofthisreportwastopresentdamagesatparticularlevelsof
probability,suchasthe95%confidencelevel,asisstandard
practicewhenestimatingaValueatRiskfigure.However,
forexpositionalpurposes,temperaturelevelsarepresented
insteadofprobabilitylevels.Thetemperaturelevels,
suchas5°Cand6°C,occurwithparticularprobabilities,
approximately10%and3%respectively.Thedamages
occurringwiththeseprobabilitiesarethenpresented
atthesetemperaturelevels.Thesedamagelevelsare
consistentwiththesetemperatures,astheyoccurwiththe
sameprobability.However,wewanttobeclearthatdamage
levelsarenotuniquetotemperaturelevels.Forexamplethe
worstcasedamagesat4°Ccouldbeequal,orhigher,than
thedamagesthatoccurwiththesameprobabilityas5°Cof
warming(butthatisbecauseworstcasedamagesat4°Chas
alessthan10%probabilityofoccurring).Thepresentation
ofdamagesconsistentwithtemperaturelevels,rather
thanprobabilitylevels,issimplyachoiceaboutexposition:
thereisnomodificationofresults–justadifferentchoice
oflanguage.Whilethispresentationmaycausesome
confusionforanyreadercomfortablewithprobabilitylevels,
theeditorialguidanceisthatthisisoutweighedbythe
benefitofreachingreaderswhomorereadilyidentifywith
temperaturelevels.
Theaveragelossistheexpectedvalueofdamagesacross
thefullrangeoftemperaturelevels.Thisreportestimates
aprobabilitydistributionovertherangeoflossesto
manageableassets.Thatistosaythateachlevelofpossible
damagehasanassociatedprobabilityofoccurring.The
averagelossistheexpectedvalueofthedistribution,which
isfoundbyweightingeveryoutcomeinthedistributionby
itsassociatedprobabilityandsummingallofthesevalues.
Forexample,theexpectedvalueofadiceis3.5,whichis
thesumof1/6thofthenumbers1to6.Theaverageloss
doesnothaveanassociatedtemperaturelevel,butisthe
(probabilityweighted)averagelossthatcanbeexpected
acrossthefullrangeoftemperaturelevels.
EXPLAINING THE TERMINOLOGY
OF THE RESULTS

42
APPENDIX
Mitigation consistent with a likely chance of remaining under 2°C of warming reduces average
losses by at least half and losses at extreme outcomes by at least three quarters. Losses to
manageable assets will occur even if warming is limited to 2°C. However, these losses will be
greatly reduced if mitigation action is taken. Average losses from the private investor perspective
are halved to $2 trillion. Losses at extreme outcomes are reduced even further, with damage
consistent with 6°C of warming reduced by three quarters. The reduction in losses at extreme
outcomes is very important for asset managers that must, due to regulation, be resilient to tail
risks, such as insurance firms. The benefits of mitigation are even greater from the government
perspective, as the lower discount rate increases the benefit of avoiding damage, in the same
way that it increases the cost of suffering damage.
Thelossesareprimarilyincurredinthelatterhalfofthecentury.Figure1showshowthepresentvalue
loss to current assets evolves over time along temperature consistent paths, as well as the
average loss over time. To illustrate, if we were only concerned with losses over the next 40
years, to 2055, and believed that the world was on a path to 6°C of warming by 2100, then the
consistent present value loss to current assets to 2055 would be $2 trillion (discounted using
private sector rates); this is where the red line crosses the lowermost y-axis gridline at 2055.
Losses after 2055 would of course still occur, but, given the 40 year time horizon, they would,
in this example, be disregarded. As the figure shows, the losses are primarily incurred in the
latter half of the century, as there is a lag between the release of emissions and temperatures
increasing. It is only later in the century that the climate consequences of the majority of
emissions are expected to be felt, although there is scientific debate over whether the climate
response will be faster (Frölicher & Paynter, 2015).
Figure 1: Losses are primarily incurred in the latter half of the century
Note: Losses are discounted at private sector discount rates. The losses over time are consistent with paths to the respective temperature levels being
reached in 2105.
0
2
4
6
8
10
12
14
16
18
2105209520852075206520552045203520252015

43
APPENDIX
4.2 Capital approach results
Climate change is expected to, on average, reduce the capital stock in 2100 by 9%, and up to 28%
at extreme outcomes. The capital approach estimates the reduction in non-financial assets in 2100
due to climate change. This capital stock underpins many of the managed financial assets. Climate
change is expected, on average, to reduce this stock by 9% in 2100. Losses are much higher if more
extreme outcomes are expected to occur, with a 14% and 28% reduction in the capital stock in 2100
consistent with warming of 5°C and 6°C respectively. These results are summarised in Table 2.
Average
loss
5°C
loss
6°C
loss
9% 14% 28%Percentage reduction in
the 2100 capital stock
Table2-Climatechangecouldreducethe2100capitalstockby9%onaverage,and28%at6°Cofwarming
Note:Thelossduetoclimatechangeofthecapitalapproachmeasureisnotdiscounted,butpresentedas
thereductiontothecapitalstockin2100.
Source:VividEconomics
Mitigation consistent with a likely chance of remaining under 2°C of warming reduces
average losses by two thirds and losses at extreme outcomes by 85%. The benefits of
mitigation are greater in the capital approach than in the dividend approach because the capital
approach focuses on the future capital stock, rather than present value losses, so the benefit is
not discounted.

44
APPENDIX
Figure 2: Losses to the capital stock are, like losses to current financial assets, primarily incurred in the latter
half of the century
Damage to future capital stock (% loss in year due to climate change)
Note: Losses are not discounted. The losses over time are consistent with paths to the respective temperature levels being reached in 2105.
6°C loss 5°C loss 4°C loss Average loss3°C loss
0
5
10
15
20
25
30
35
40
2105209520852075206520552045203520252015
Losses to the capital stock are, like losses to current financial assets, primarily incurred in
the latter half of the century. Figure 2 shows how the damage to the capital stock evolves over
time along temperature consistent paths, as well as the average loss over time. Damage to the
capital stock is not discounted, that is to say that the values represent the loss of capital in the
relevant year. To illustrate, if we were concerned with losses to the capital stock in 2055, and
believed that the world was on a path to 6°C of warming by 2100, then the consistent damage
due to climate change would be a 4% reduction in the capital stock. As the figure shows, the
losses are primarily incurred in the latter half of the century, for the same reason as in the
dividend approach: that there is a lag between the release of emissions and temperatures
increasing.

45
APPENDIX
4.3 Secular stagnation scenario
Damages in the secular stagnation scenario are higher. Table 3 compares the results for the
dividend approach between the base case and the secular stagnation case. The damages are
high in the secular stagnation case for two reasons. First, as lower productivity leads to lower
economic growth, the discount rate is lower, and therefore the present value of damages is
higher. Second, as explained below, the temperature consistent damages are more extreme.
The temperature consistent damages are higher in the secular stagnation scenario because
high temperatures are less likely to occur and so consistent damages are more extreme.
Lower economic growth in the secular stagnation scenario leads to lower emissions. As a
consequence, it is less likely that a scenario of secular stagnation will result in high temperature
increases. Temperature consistent damages are therefore higher as they too are more extreme.
To illustrate, 5°C and 6°C warming have a 10% and 3% respective probability of occurring in
the base case, but a 5% and 1% respective probability of occurring in the secular stagnation
scenario. Therefore the damages consistent with 6°C of warming occurring in the base case are
damages occurring with 3% probability, while those consistent with 6°C of warming occurring in
the secular stagnation case are damages occurring with 1% probability.
Present value loss to current
manageable assets
Average
loss
5°C
loss
6°C
loss
Increasing productivity
scenario (base case)
Decreasing productivity
scenario (secular stagnation)
Table3-Thesecularstagnationscenariohashigherpresentvaluelossesaslowergrowthleadstolower
discountrates,andbecausetheprobabilityatwhichhighertemperaturesoccurishigher,sotheconsistent
levelofdamageismoreextreme
Note:Thecurrentstockofmanageableassetsis$143trillion.Dollarvaluesarein2015dollars.Lossesare
thecumulativelosstomanageableassetsto2100discountedtothepresentusingprivatesectordiscount
rates.

46
APPENDIX
4.4 Conclusions
The results show that the asset management industry is, from a private investor perspective,
at risk of extreme outcomes and that the government has a role in safeguarding society
from average losses due to climate change, let alone extreme outcomes. The results suggest
that private investors are overvaluing manageable assets today by 3% of their current value if
they consider the expected average impacts of climate change. However, if they consider the
tail risks, that 5°C or 6°C of warming will be reached by 2100, then assets today are overvalued
by 5% and 10% respectively. These are present value losses, discounted at private sector
rates. From a government perspective, with a lower discount rate, even the expected average
impacts to manageable assets are of major concern, let alone extreme outcomes. The mean
average expected loss is $13.9 trillion, equal to 10% of current assets, while if 6°C of warming is
expected, then the consistent level of present value losses at government discount rates is $43
trillion, equal to 30% of current assets.
These results are for global aggregate assets, but impacts will vary by asset class and region.
DICE, as a top-down model, can only estimate results for global aggregate assets. The few
studies that have considered the impact of climate change at a more disaggregated level, often
at the economic sector, asset class or regional level, tend not to quantify the impact, due to the
greater uncertainty inherent in more disaggregated analysis. However, they find that, in general,
economic sectors and asset classes that are concerned with physical assets or natural resources
are the most vulnerable to climate change. For example, Mercer (2015) finds that real estate,
infrastructure, timber and agriculture are the most sensitive to climate impacts, while Vivid
Economics (2013), which surveys a range of models, also finds that tourism may be negatively
affected by climate change. Developing, and to an extent, emerging economies, are more
vulnerable to climate change, not just because these countries are, in some cases, more likely to
face greater climatic changes, but because, due to their lower income, they are less able to adapt
to climate impacts (IPCC, 2014b). This means that, in these countries, large scale economic
consequences may follow from relatively small scale climate change. As a result, financial assets
in these regions may underperform other regions, for example Mercer (2015) identifies emerging
market global equities as a particularly vulnerable asset class. Furthermore, while around 70% of
current manageable assets reside in Europe and North America (Financial Stability Board, 2014),
the growth of emerging markets means that many future assets will be in these regions.

47
APPENDIX
Our modelling suggests that much of the impact to manageable assets will come through
lower asset returns, affecting the entire portfolio, rather than direct damage, which would
be more localised. The capital approach can decompose the percentage reduction in the
capital stock due to climate change into direct damage, from events such as storms and floods,
and indirect damage, due to a decrease in asset returns, which reduces the accumulation of
assets relative to a world of no climate change. Indirect damage, with 5% probability in 2100,
is responsible for around 60-70% of the loss of capital, with the lower and upper values in the
range from the decreasing and increasing productivity scenario respectively. While this result is
driven by assumptions about which there is limited evidence8
, it indicates that asset managers
may face a challenge in avoiding climate risks by moving out of vulnerable assets classes and
regions. This is because, at least under lower probability, higher impact outcomes, the results
suggest that climate change will have a primarily macroeconomic impact, which will affect the
entire portfolio of assets.
In conclusion, this report finds that the impacts of climate change on the asset management
industry are of a significant order of magnitude when put within its own risk management
framework. The industry has limited awareness of the risks it faces due to climate change,
despite an institutional focus on risk management and clear metrics for risk measurement. This
report speaks to that ethos by putting climate risks into a VaR framework. It finds, by adapting
an established economy-climate model, that the VaR due to climate change is significant.
Climate change is expected to cause permanent, present value losses to current manageable
assets of 3% on average and up to 10% at extreme outcomes, when discounted at a private sector
discount rate, and puts at risk 9% of the 2100 capital stock on average, and 28% at 6°C.
From a government, and therefore regulator, perspective the present value losses are even
higher, at 10% of current manageable assets on average and up to 30% at extreme outcomes.
However, mitigation consistent with a 2°C target reduces average losses by at least half and
losses at extreme outcomes by at least three quarters. So action on climate change is likely to be
in the industry’s interest, as a way to manage its own risks.
________________
8 In particular, it is assumed that 30% of climate damage affects the capital stock, but there is currently limited research on the appropriate share (Dietz & Stern,
2015). This assumption is related to, but not equal to, the share of direct damage resulting from the model; for example the share of direct and indirect damage to
capital is approximately equal at higher probability outcomes.

48
APPENDIX
5. Annexes
These annexes provide detailed discussion on the main conceptual and modelling aspects
of the report. These are: the definition of assets, key uncertainties, productivity scenarios,
discounting, mitigation scenarios and the parameter values of key modelling assumptions.
5.1 Definition of Assets
This report focuses on the stock of financial assets that could potentially be managed.
Manageable assets are financial assets that can be professionally invested on behalf of an
asset owner for a fee. This stock is the potential market size of the asset management industry,
rather than the stock of assets currently managed, which is estimated to have been 22-25% of
the potential market in recent years (McKinsey & Company, 2012). We take a broad definition
of manageable assets, to cover all assets held by non-bank financial institutions, which
was $143 trillion in 2013, as estimated by the Financial Stability Board (2014). Bank and
Central Bank financial assets are excluded as these are, largely, managed by the institutions
themselves. Non-bank financial institutions are: insurance companies, pension funds and other
financial intermediaries (such as money market funds, investment funds), plus public financial
institutions, as their assets can be professionally managed by a third party. This definition is
appropriate as this report focuses on how the industry’s overall prospects are affected by climate
change, rather than on its ability to attract customers and increase its market share in the face
of climate change.
Financial assets, broadly speaking, consist of all financial claims9, such as bonds, and shares
or other equity in corporations. A financial asset is created by raising a liability that will be
paid off from a flow of output. At a fundamental level, output results from a production process
where (technology and human capital augmented) labour and non-financial assets, commonly
referred to as economic capital, are combined. Non-financial assets can be fixed assets, such
as machinery, natural resources, such as water, and ideas, such as patents. In a developed
economy, the share of output earned by non-financial, fixed assets is typically 30%, and the
remaining 70% is earned by labour.
The stock of financial assets depends on the flow of output, which in turn depends on the
stock of labour and non-financial assets. For example a corporation can issue a bond and
use its machinery to produce output sufficient to liquidate the bond when it becomes due, or
a household can take out a car loan and repay it with wages. Financial assets can therefore
be created from corresponding financial liabilities that are backed by non-financial assets or
labour income, but not all non-financial assets or labour income need be used to create financial
liabilities and assets.
________________
9 The payment or series of payments due to the creditor by the debtor under the terms of a liability.

49
APPENDIX
The relationship between non-financial and financial assets has two important implications
for this report. First, the results of the capital approach, the impact of climate change on non-
financial assets, can be converted into an effect on financial assets, as described below. Second,
the assumption in the dividend approach: that dividends from manageable assets grow at the
same rate as GDP, is based on the logic that the stock of financial assets depends on the flow
of output, that is to say: GDP, which in turn depends on the stock of labour and non-financial
assets.
The dividend approach takes into account all financial assets – those backed by both capital
and labour – while the capital approach only considers financial assets that are backed by
capital. As the relationship between dividends and GDP growth relates to dividends from all
financial assets, the dividend approach estimates results for the full stock of manageable assets.
Indeed, the broader the definition of financial assets, the more appropriate the assumption is,
as this controls for cyclical differences in relative performance across asset classes. The capital
approach, on the other hand, estimates the impact of climate change on the capital stock, so
can only provide information on the impact of climate change to those financial assets that are
backed by non-financial assets in the capital stock. This difference is illustrated by Figure 3. As
explained below, not all non-financial assets can be used to create a financial asset, and not all
financial assets are backed by capital, such as car loans or household mortgages10, where the
liability is met by labour income.
________________
10 Note that while household mortgages are secured by a non-financial asset, real estate, the liability is metfrom the income of the mortgage holder. For example if a
mortgaged house is destroyed, the mortgage still exists, and must be paid by the mortgage debtor.
Figure 3: The capital approach focuses on a sub-set of manageable assets, those that are backed by
non-financial assets, while the dividend approach focuses on all manageable assets
Note: The list of assets in the figure is not exhaustive.
Non-financial
assets
Natural
resources
Non-corporate
business
Public
infrastructure
Car loans
Household
mortgages
Sovereign debt
Capital
approach
manageable
assets
Non-financial
corporate debt
and equity
Household
wealth
Financial assets
In the dividend
approach, all non-bank
financial assets are
manageable assets

50
APPENDIX
In the capital approach, impacts to non-financial assets can be converted to impacts to
manageable assets. To calculate the effect on manageable assets of a dollar loss of non-
financial assets requires two conversions. First, the share of non-financial assets that are used
to back manageable financial liabilities, and concurrent financial assets, must be estimated.
This includes the non-financial corporate sector, which issues corporate debt and equity, and
financial instruments arising from household wealth, such as pensions. It excludes the non-
financial assets of the financial sector, which holds non-financial assets for risk and regulatory
purposes, non-corporate businesses, as these rarely issue manageable securities against their
non-financial assets, and (the majority of) government debt, as discussed below. Second, the
financial liabilities, and therefore concurrent financial assets, created per dollar of non-financial
asset must be estimated. The product of these two ratios converts a dollar of lost non-financial
asset to a value of lost manageable assets. To illustrate, in the US, on average over the last
decade, the non-financial corporate sector owned 25% of non-financial assets, and $1.7 of
financial asset was created per $1 of non-financial asset in this sector (Board of Governors
of the Federal Reserve, 2015). So a $1 loss of non-financial assets, will lead to a $1.7 loss of
non-financial corporate manageable assets, 25% of the time, or a $0.43 loss of non-financial
corporate manageable assets on average for every $1 of non-financial assets destroyed.
Data is insufficient to identify the value of government debt used for investment in non-
financial assets. Some government debt is used to finance investment in non-financial assets,
so would be included in the scope of manageable assets under consideration in the capital
approach. However, a large proportion of government debt is used to finance consumption, the
repayment of which is premised on future claims to tax income. Data on the share of government
expenditure to acquire non-financial assets is available, for example IMF (2014). From this
data, one could assume that the same share of government borrowing is used to finance this
investment. However, data is not available over a long, or complete, time series, and suffers
from significant variation over time, due to the financial crisis and ensuing recession. Therefore
it does not provide a reliable enough basis to make assumptions about future patterns of
government spending, especially to 2100.
There are two main issues with this conversion of loss to non-financial assets into
manageable assets. First, it only estimates a sub-set of manageable assets; it does not capture
the effect of climate change on financial assets that are backed by labour income, such as
household mortgages, and taxes on this income, as many government bonds are. This is in
contrast to the dividend approach, which considers all manageable assets. Second, data on
the conversion rates is not available globally, and rates are likely to vary over the next century.
As a consequence, results in the capital approach are not converted into losses of manageable
assets, but presented as a percentage reduction to the capital stock. If manageable assets
are, on average, as vulnerable to climate change as the capital stock in general, then a similar
percentage reduction to manageable assets would be expected.

51
APPENDIX
5.2 Key uncertainties
There are a large number of uncertainties associated with modelling the impacts of climate
change. As IAMs combine three separate simplified models: economy, emissions and climate,
they require many different inputs. There is no consensus on a single, definitive value for a wide
range of these inputs and so it is appropriate to treat them as uncertain, i.e. define a probability
distribution over the range of possible values, when estimating the impacts of climate change.
Indeed, the most comprehensive study of uncertainty with regards to the DICE model randomises
all 51 input parameters (Anderson, Borgonovo, Galeotti, & Roson, 2014).
Research suggests that only a small subset of the parameters in DICE have a significant
impact on the key results. Large variations in most parameters do not significantly change the
overall impacts of climate change. This suggests that, for computational simplicity, it is sensible
to confine our attention to those that do. Dietz and Asheim (2012) and Nordhaus (2008) identify
eight parameters as important for uncertainty analysis while other well-known studies, such
as the US Government’s InterAgency Working Group on the Social Cost of Carbon (2010), have
focused on only one, climate sensitivity.
When estimating climate VaR, arguably three parameters are key and as a result, only these
are treated as uncertain in this report. The first parameter is the level of productivity growth
achieved in the initial decade, because it sets the magnitude of growth for the century.
It therefore has a direct and considerable impact on the size of the future economy and thus, the
income available to invest in assets. The second and third parameters impact the magnitude of
damage that climate change causes. Climate sensitivity, the equilibrium increase in global mean
temperature following a doubling in the atmospheric concentration of greenhouse gases, defines
how much the planet warms in response to emissions, which is uncertain due to the complexity
of the climate system. The third parameter, (part of) the curvature of the damage function,
defines how much of output and capital is destroyed by a given temperature increase. So, taken
together these two parameters define the amount of damage a given level of emissions causes.
Uncertainty in each of these three parameters is introduced into DICE by modelling each as a
random variable.
Figure 4: The distribution over initial productivity growth rates varies somewhat between the productivity
scenarios
Cumulative probability (%)
Note: The figure shows the cumulative probability of an annual rate of productivity growth being achieved in the initial decade, after which, this trend
will evolve according to a scenario: its growth rate either increases or decreases.
0
20
40
60
80
100
Increasing productivity scenarioDecreasing productivity scenario
4%3%2%1%0%1%-2%
Annual rate of productivity growth for initial decade

52
APPENDIX
The initial rate of growth in productivity is defined as a normal distribution. In a new paper,
Dietz, Gollier and Kessler (2015) calibrate uncertainty around productivity growth based on
nearly 200 years of data from the UK and US. They find that it is best modelled as a normal
distribution with a mean of 0.84% annual growth and standard deviation of 0.59%. After the
initial decade, the trend then evolves according to a scenario: its growth rate either increases or
decreases, described further in Annex 5.3.
The distribution over initial productivity varies somewhat between the productivity
scenarios. Figure 4 shows the cumulative probability of an annual rate of productivity growth
being achieved in the initial decade, after which, this trend will evolve according to a scenario:
its growth rate either increases or decreases. The cumulative distribution functions are
different for each scenario. The distribution for the decreasing productivity scenario is based
on Dietz, Gollier and Kessler (2015), as described above. The increasing productivity scenario
is a modification of this former distribution. Specifically, it is truncated to have no negative
values, resulting in a shift of some probability weight to positive values. This is because, in this
scenario, an initially negative productivity growth rate would become increasingly negative over
time. Such an outcome would imply that the global economy, of its own accord (rather than due
to climate change), has a recession that leads to a negative spiral of decline, resulting in the
collapse of the economy. This is (relatively) implausible and inconsistent with a scenario where
productivity is increasing, i.e. negative productivity growth should not persist in such a scenario.
Climate sensitivity is defined as a log-logistic distribution. Parameters for uncertainty
over the climate sensitivity are determined by large-scale climate models. We define climate
sensitivity as a log-logistic distribution, following Dietz and Stern (2015) and Dietz, Gollier
and Kessler (2015). It results in a probability distribution with an ~80% probability that value
for climatesensitivity is between 1.5 and 4.5, consistent with the latest IPCC report (IPCC,
2013). The distribution has a mean of 3.6309°C of global average warming for a doubling in the
atmospheric concentration of greenhouse gases and a standard deviation of 1.4215.
The risk of catastrophic climate change is modelled to take into account the range of views
in the academic literature. Damage to the economy due to climate change is estimated by a
damage function. This is a polynomial equation that converts a global average atmospheric
temperature increase into economic damage. This equation is fitted to estimates from impact
studies in the literature. It is very hard to estimate the effect on the modern economy of high
levels of temperature increase, as there has been no experience of this. However, it is highly
likely that damage will be increasing with temperature. The question is whether damage
increases slowly or quickly, in which case the curvature of the damage function will be shallow or
steep respectively. Different authors have taken different approaches to modelling the curvature
of the damage function, particularly at higher temperature increases. A key issue is whether high
temperature increases could lead to catastrophic climate change, which would result in a very
steep damage function at these temperature increases.

53
APPENDIX
In our version of DICE, the damage function is shallower or steeper depending on a random
variable. In DICE, the damage function is defined as a polynomial of atmospheric temperature as
set out in the equation below. Dietz and Asheim (2012) account for uncertainty in the curvature
of the damage function through a random coefficient on a higher-order term: where is the
atmospheric temperature at time t, are coefficients used to calibrate the function on impacts
studies and is the random parameter. The rationale behind this approach is that can be used to
effectively ‘turn on or off’ a catastrophic climate impact responsible for steeply increasing losses
in GDP beyond a certain degree of warming. Weitzman has been the key proponent of the idea of
catastrophic climate impacts at a macroeconomic level (Weitzman, 2012), but Nordhaus tends to
dismiss it. Given that there is no compelling empirical evidence with which to discriminate (Tol,
2012), the calibration in Dietz and Asheim was undertaken such that the distribution of values of
spanned the views of these two scholars. Since then, Dietz and Stern (2015) have introduced, as
a ‘high’ scenario, a damage function of the same form as Weitzman (2012) but that is even more
pessimistic. This damage function is used prominently in Covington and Thamotheram (2015),
and is also the damage function used in this report.
We model the curvature of the damage function as a normal distribution. To estimate the
climate VaR, we take the broad approach of Dietz and Asheim in using to span the existing
literature, but with Nordhaus and Dietz and Stern (2015) as the end-points of the range of views.
This results in a normal distribution for with a mean of 0.12417 and a standard deviation of
0.04139.
5.3 Productivity scenarios
As productivity is critical to the main results, it is important to recognise that uncertainty
exists about its evolution in two ways. We are uncertain about both the trend of productivity
growth, and how this trend might evolve in the longer term. As discussed in Annex 5.2,
productivity growth has a large and cumulative effect on future output and can significantly
alter the climate VaR as a result. Thus, it is important to explore the impact that each of these
uncertainties might have.
Uncertainty regarding the trend of productivity growth is bounded and so can be modelled as
a random variable. As a large dataset of past growth rates exists, we can identify a reasonable
range of values that the future growth rate could take. From this, it is also possible to estimate
how likely it is that the growth rate will take specific values within this range. In the context of
DICE, this allows us to define uncertainty over the initial rate of productivity growth by fitting a
probability distribution over past data, as described in Annex 5.2.

54
APPENDIX
Figure 6: In the decreasing productivity scenario, the rate of growth in productivity decreases, leading to lower
overall growth
Index of productivity (2015= 1)
Note: Figure is for the decreasing productivity scenario, where the growth rate of productivity decreases over time.
0
1
2
3
4
5
2105209520852075206520552045203520252015
1% probability outcome95%-5% probability range Mean outcome
Figure 5: In the increasing productivity scenario, the rate of growth in productivity increases, leading to higher
overall growth
Index of productivity (2015= 1)
Note: Figure is for the increasing productivity scenario, where the growth rate of productivity increases over time.
0
2
4
6
8
10
12
14
16
1% probability outcome95%-5% probability range Mean outcome
2105209520852075206520552045203520252015

55
APPENDIX
Scenarios are used to describe different prospects for long term productivity growth.
The long term growth rate of productivity can either be increasing over time, so the global
economy continues to grow at a relatively high rate, which is the EIU’s view; or decreasing, so
global growth slows, which reflects the idea of ‘secular stagnation’, the default setup in the
DICE model. We use the EIU’s projections of increasing productivity growth as the base case.
However, as these present different prospects for future economic growth, and therefore the
stock of assets, they are presented as scenarios: the increasing productivity scenario (the base
case scenario), where productivity continues to grow at an increasing rate; and the decreasing
productivity scenario (the secular stagnation scenario), where the growth rate of productivity
decreases over time. The increasing productivity scenario is calibrated to EIU projections of the
capital stock, so that the capital stock in the mean outcome in 2100 aligns with EIU projections.
The decreasing productivity scenario follows the default setup in DICE. So, the overall treatment
of productivity is that the trend in productivity growth is uncertain over the initial decade, but
then, whatever the trend, it either increases or decreases in the long run, depending on the
scenario.
As Figure 5 and Figure 6 show, these scenarios produce very different levels of productivity
in the future. In the increasing productivity scenario, there is a 5% probability that productivity
will increase by at least ~7.5 times by 2100 – this probability being generated from the three
uncertain parameters described in Annex 5.2, while in the mean case it increases by at least ~3
times by 2100. This is in contrast to the decreasing productivity scenario, where there is a 5%
probability that productivity will increase by at least ~3 times by 2100, while in the mean case it
almost doubles.
5.4 Discounting
To quantify the present value of the climate VaR, the future climate VaR must be discounted
at an appropriate rate. Discounting is one of the most controversial issues in the economics
of climate change. In this report, however, the primary point of view taken is that of a private
investor, attempting to value the possible impacts of climate change on his/her asset portfolio.
As a private investment problem, many discounting controversies can be avoided, because they
relate to the social discount rate to be applied by governments to public investment.
One difficult issue that remains is the fact that the impacts of climate change are potentially
‘non-marginal’, requiring endogenous discounting. This means that impacts can be so large
as to affect the rate of economic growth (Dietz & Hepburn, 2013; Gollier, 2012). The rate of
economic growth is intrinsically linked to the discount rate: the faster the economy grows, the
higher the discount rate, and vice versa. The difficulty presented by climate change being a
non-marginal problem is then that no single discount rate will be appropriate for all scenarios,
rather different discount rates will be appropriate for each scenario, depending on economic
growth in that scenario. As the discount rate within a scenario depends on the economic growth
rate of that the scenario, the discount rates are known as endogenous rates. The need to use
endogenous rates rightfully precludes exogenous approaches to discounting, such as declining
discount rates.

56
APPENDIX
Endogenous discounting creates a paradox in the case of the capital approach. Climate
change has the potential to significantly reduce the future stock of capital, compared with a
counterfactual scenario absent climate change. But in this bleak future the economy grows much
more slowly, meaning that, to calculate the present value of the climate VaR, a lower discount
rate should be applied. The result is that, perversely, the present value of the stock of capital can
be higher under climate change than without. This is not incorrect – it merely says that a smaller
future stock of capital is more valuable today, given the prospect of a slow-growing world.
However, it naturally presents difficulties in presentation and understanding, which means it can
be more effective to simply estimate the future climate VaR in undiscounted terms.
The same issue affects the dividend approach, but can justifiably be avoided. In this case
one can take the empirically-supported view that portfolio managers will nonetheless derive
the present value of the climate VaR using a single discount rate for all scenarios, because that
is standard practice (i.e. their practice does not take into account that climate change could be
non-marginal). In particular, the discount rate tends to be based on historical returns, although
it can also be calibrated on a growth path consistent with the DICE model.
The discount rate used in the dividend approach is calculated from the perspective of a
private investor. The discount rate is a function of the future GDP growth rate without climate
change, calculated each decade, plus a premium to account for holding risky assets relative to a
risk-free asset. The GDP growth rate without climate change is used as investors do not currently
consider climate impacts. This results in a more conservative estimate of losses, as GDP growth
with climate change will be lower, leading to a lower discount rate if investors took this into
account. Average private sector discount rates are initially 5.5% but fall to 4% towards the end
of the century, due to slowing economic growth.
However, as a sensitivity, government discount rates are also applied in the dividend
approach. This sensitivity is explored because governments have a duty to safeguard financial
assets on behalf of society and regulate to fulfil this duty, so the value at risk from a government
perspective is also relevant. The government discount rate used in this report follows the
approach of the Stern Review. As noted above, such an approach is not without its challenges,
but at least the challenges of the Stern Review approach to discounting are well-documented
and, largely, well-understood. In the Stern Review approach, the government discount rate is a
function of the future GDP growth rate with climate change, as the government should consider
climate impacts, plus a pure rate of time preference of 0.1%, which means that impacts on
future generations are only discounted according to the probability that future generations will
not exist to experience the impacts.11 The pure rate of time preference is also multiplied by a
marginal elasticity of utility of 1, which means that impacts to people within a time period are
valued equally. Average government discount rates are initially 3.8% but fall to 2% towards the
end of the century, due to slowing economic growth.
________________
11 This takes into account the low but non-negligible probability that civilisation ends through some means unrelated to climate change.

57
APPENDIX
5.5 Mitigation scenarios
The base case in the analysis follows a path of minimal emissions mitigation. The emissions
trajectory in the base case is taken directly from the published version of DICE-2010 and
represents a scenario in which no new climate change policies are adopted (Nordhaus, 2010);
this can be interpreted as ‘complete [future] inaction and stalemate on climate policies’. This
scenario results in a reduction of 6.5% of total emissions in 2105 relative to a world in which
there were not only no new climate policies but also no existing climate policies.
This is the relevant scenario as it allows the isolation of the impacts of climate change. When
trying to estimate the climate VaR, we are trying to estimate the total assets that are at risk
due to climate change in the future. The variable of interest is therefore the potential impact of
climate change and not the response to this potential impact, that is to say greater mitigation.
Therefore, it is most appropriate to compare scenarios with and without climate change, but also
both without additional mitigation. If scenarios both with additional mitigation were to be used,
it would not be possible to disentangle the impact of climate change itself from the offsetting
impact that additional mitigation would have through reduced emissions. So the base case
results present the value at risk in a world of minimal mitigation with climate change relative to a
world of minimal mitigation without climate change.
Nonetheless, examining scenarios with additional mitigation can provide useful insights
and is explored as a sensitivity. The scenario with additional mitigation is based on the ‘LimT’
scenario in the published version of DICE-2010 (Nordhaus, 2010) but it is recalibrated such that
it limits global average temperature increase to 2°C with a 66% probability. This is consistent
with a ‘likely chance’ of keeping temperature change below 2°C, the most stringent mitigation
scenario the IPCC considers to be feasible (IPCC, 2014a)
The sensitivity calculates the difference in the climate VaR between the base case and the
2°C scenario employing the same methodology as the main analysis. For the 2°C scenario,
the climate VaR is calculated for both the capital approach and dividend approach. Similar to
the main analysis, this is done by comparing the value of the capital stock and of total dividends
respectively in a world with and without climate change. The absolute value for the climate VaR
is then compared with the equivalent climate VaR in the base case scenario. Therefore, this
difference indicates how much less capital would be at risk from climate change if mitigation
efforts were sufficient to limit the global average temperature increase to 2°C.

58
APPENDIX
5.6 Key modelling assumptions
Name Value Source Comment
Climate sensitivity
Depreciation rate
(% per annum)
Curvature of the damage
function
Initial yield on portfolio
(% per annum)
10%
Normal distribution:
Mean = 0.0084
Std Dev = 0.0059
(truncated at 0)
Normal distribution:
Mean = 0.0084
Std Dev = 0.0059
Log-logistic distribution:
Mean = 3.6309
Std Dev = 1.4215
(truncated at 0.75)
2.76%
Dietz and Stern
(2015); Dietz,
Gollier and Kessler
(2015); IPCC (2013)
Nordhaus
(2010)
Dietz and Asheim
(2012); Nordhaus
and Dietz and Stern
(2015)
Dimson, Marsh, &
Staunton (2011)
Distribution results in a probability distribution
withan ~80% probability that value for climate
sensitivity is between 1.5 and 4.5, consistent with
the latest IPCC report.
Distribution reflects the wide range of views
on the appropriate coefficient for the ‘high-
order’ damage term seen in the literature.
Standard assumption in the literature and
follows original calibration of DICE-2010.
Estimates for the representative yield and
rate of return on equities and bonds are
taken from the real annualised yields and
returns over the period 1900-2010. These are
calculated as the geometric mean of the time
series data.
Productivity growth rate in
initial decade
Share of damage
to capital
Capital share of income (%)
Initial rate of return
on portfolio (% per annum)
30%
30%
5.50%
Dietz, Gollier and
Kessler (2015)
Dietz & Stern
(2015)
Nordhaus
(2010)
Dimson, Marsh, &
Staunton (2011)
Distribution is fitted to 200 year dataset on UK
andUS productivity. The distribution is truncated
at 0 for the increasing productivity scenario.
Table4-Table4.KeyassumptionsinDICEforthecalculationoftheclimateVaR

59
APPENDIX
6. References
Anderson, B., Borgonovo, E., Galeotti, M., & Roson, R. (2014). Uncertainty in climate change
modeling: can global sensitivity analysis be of help? Risk Analysis, 34(2), 271–293.
Arent, D. J., Tol, R. S. J., Faust, E., Hella, J. P., Kumar, S., Strzepek, K. M., … Yan, D. (2014).
Key economic sectors and services. In C. B. Field, V. R. Barros, D. J. Dokken, K. J. Mach, M. D.
Mastrandrea, T. E. Bilir, … L. L. White (Eds.), Climate Change 2014: Impacts, Adaptation, and
Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the
Fifth Assessment Report of the Intergovernmental Panel of Climate Change (pp. 659–708).
Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press.
AVOID. (2015). Climate change impacts. Retrieved from http://www.avoid.uk.net/research/
climate-change-impacts/
Board of Governors of the Federal Reserve. (2015). Financial Accounts of the United States.
Retrieved from http://www.federalreserve.gov/releases/z1/Current/z1.pdf
Cass, D. (1965). Optimum Growth in an Aggregative Model of Capital Accumulation. The Review of
Economic Studies, 32(3), 233–240.
Covington, H., & Thamotheram, R. (2014). How Should Investors Manage Climate-Change Risk?
Rotman International Journal of Pension Management, 7(2). Retrieved from http://ssrn.com/
abstract=2497514
Covington, H., & Thamotheram, R. (2015). The case for forceful stewardship. Part 1: the financial
risk from global warming. Cambridge, UK.
Dietz, S., & Asheim, G. B. (2012). Climate policy under sustainable discounted utilitarianism.
Journal of Environmental Economics and Management, 63(3), 321 – 335.
Dietz, S., Gollier, C., & Kessler, L. (2015). The climate beta. London.
Dietz, S., & Hepburn, C. J. (2013). Benefit-cost analysis of non-marginal climate and energy
projects.
Energy Economics, 40, 61–71. http://doi.org/10.1016/j.eneco.2013.05.023
Dietz, S., & Stern, N. (2015). Endogenous growth, convexity of damages and climate risk: how
Nordhaus’ framework supports deep cuts in carbon emissions. Economic Journal, 125(583),
574–602.
Dimson, E., Marsh, P., & Staunton, M. (2011). Equity Premia Around the World, 2011(July), 1–19.
Financial Stability Board. (2014). Global Shadow Banking Monitoring Report 2014.
Frölicher, T. L., & Paynter, D. J. (2015). Extending the relationship between global warming and
cumulative carbon emissions to multi-millennial timescales. Environmental Research Letters,
10(7), 75002. Retrieved from http://stacks.iop.org/1748-9326/10/i=7/a=075002
Gollier, C. (2012). Pricing the Planet’s Future: The Economics of Discounting in an Uncertain
World. Princeton University Press.
Hope, C. (2006). The Marginal Impact of CO2 from PAGE2002: An Integrated Assessment Model
Incorporating the IPCC’s Five Reasons for Concern. The Integrated Assessment Journal, 6, 16–56.
Hope, C. (2013). Critical issues for the calculation of the social cost of CO2: why the estimates
from PAGE09 are higher than those from PAGE2002. Climatic Change, 117(3), 531–543.
IMF. (2014). Table W7. Major Categories, by Levels of Government. Retrieved from http://
elibrary-data.imf.org/DataReport.aspx?c=25159164&pars=
IPCC. (2013). Working Group I Contribution to the IPCC Fifth Assessment Report: Summary for
Policymakers. In Climate Change 2013: The Physical Science Basis. IPCC.
IPCC. (2014a). Climate Change 2014: Mitigation of Climate Change. In Contribution of Working

60
APPENDIX
Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change.
Cambridge, United Kingdom and New York, NY, USA.: Cambridge University Press.
IPCC. (2014b). Summary for Policymakers. In Climate Change 2014: Impacts, Adaptation,
and Vulnerability. Contribution of Working Group II to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. IPCC.
King, D., Schrag, D., Dadi, Z., Ye, Q., & Ghosh, A. (2015). Climate Change: A Risk Asessment.
Retrieved from http://www.csap.cam.ac.uk/media/uploads/files/1/climate-change--a-risk-
assessment-v9-spreads.pdf
Koopmans, T. C. (1965). On the concept of optimal economic growth. In The Econometric
Approach to Development Planning. Amsterdam: North Holland.
McKinsey & Company. (2012). The Hunt for Elusive Growth: Asset Management in 2012.
Mercer. (2015). Investing in a time of climate change.
Millner, A., Dietz, S., & Heal, G. M. (2013). Scientific ambiguity and climate policy. Environmental
and Resource Economics, 55(1), 21–46.
Nordhaus, W. (2008). A Question of Balance: Weighing the Options on Global Warming Policies.
Yale University Press.
Nordhaus, W. D. (1993). Rolling the “DICE”: an optimal transition path for controlling
greenhouse gases. Resource and Energy Economics, 15(1), 27–50.
Nordhaus, W. D. (2010). Economic aspects of global warming in a post-Copenhagen environment.
Proceedings of the National Academy of Sciences of the United States of America, 107(26),
11721–6. http://doi.org/10.1073/pnas.1005985107
Nordhaus, W. D. (2014). Estimates of the social cost of carbon: concepts and results from the
DICE-2013R model and alternative approaches. Journal of the Association of Environmental and
Resource Economists, 1, 273–312. http://doi.org/10.1086/676035
Pindyck, R. S. (2013). Climate change policy: What do the models tell us? Journal of Economic
Literature, 51(3), 860–872.
Prudential Regulation Authority. (2014). PRA response to Defra on climate change adaptation
reporting. Retrieved from http://www.bankofengland.co.uk/pra/Documents/about/
praletter020614.pdf
Ramsey, F. (1928). A mathematical theory of saving. The Economic Journal, 38(152), 543–559.
Stern, N. (2007). The Economics of Climate Change: the Stern Review. Cambridge: Cambridge
University Press.
Stern, N. (2013). The structure of economic modeling of the potential impacts of climate change:
grafting gross underestimation of risk onto already narrow science models. Journal of Economic
Literature, 51(3), 838–859.
Tol, R. S. J. (2012). On the uncertainty about the total economic impact of climate change.
Environmental and Resource Economics, 53(1), 97–116.
US Interagency Working Group on Social Cost of Carbon. (2010). Technical Support Document:
Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866.
Vivid Economics. (2013). The macroeconomics of climate change, report prepared for Defra.
Weitzman, M. L. (2009). On modeling and interpreting the economics of catastrophic climate
change. Review of Economics and Statistics, 91(1), 1–19.
Weitzman, M. L. (2012). GHG targets as insurance against catastrophic climate damages. Journal
of Public Economic Theory, 14(2), 221–244.

Economist Intelligence Unit
The Economist Intelligence Unit (EIU) is the world’s
leading resource for economic and business research,
forecasting and analysis. It provides accurate and
impartial intelligence for companies, government
agencies, financial institutions and academic
organisations around the globe, inspiring business
leaders to act with confidence since 1946.
Copyright
© 2015 The Economist Intelligence Unit Limited. All rights
reserved. Neither this publication nor any part of it may be
reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without prior
permission of The Economist Group. Whilst every effort
has been taken to verify the accuracy of information
presented in this document, neither The Economist Group
nor its affiliates can accept any responsibility or liability
for reliance by any person on this information.

LONDON
20 Cabot Square
London
E14 4QW
United Kingdom
Tel: (44.20) 7576 8000
Fax: (44.20) 7576 8500
E-mail: london@eiu.com
NEW YORK
750 Third Avenue
5th Floor
New York, NY 10017, US
Tel: (1.212) 554 0600
Fax: (1.212) 586 0248
E-mail: newyork@eiu.com
HONG KONG
6001, Central Plaza
18 Harbour Road
Wanchai
Hong Kong
Tel: (852) 2585 3888
Fax: (852) 2802 7638
E-mail: hongkong@eiu.com
GENEVA
Rue de l’Athénée 32
1206 Geneva
Switzerland
Tel: (41) 22 566 2470
Fax: (41) 22 346 9347
E-mail: geneva@eiu.com

The cost of inaction

More Related Content

Viewers also liked (17)

Similar to The cost of inaction (20)

More from The Economist Media Businesses (20)

Recently uploaded (20)

The cost of inaction