The Implications of Climate Change for

Financial Stability

23 November 2020
The Financial Stability Board (FSB) coordinates at the international level the work of national
financial authorities and international standard-setting bodies in order to develop and promote
the implementation of effective regulatory, supervisory and other financial sector policies. Its
mandate is set out in the FSB Charter, which governs the policymaking and related activities of
the FSB. These activities, including any decisions reached in their context, shall not be binding
or give rise to any legal rights or obligations.

Contact the Financial Stability Board

Sign up for e-mail alerts: www.fsb.org/emailalert
Follow the FSB on Twitter: @FinStbBoard
E-mail the FSB at: fsb@fsb.org

Copyright © 2020 Financial Stability Board. Please refer to the terms and conditions
 Table of Contents

Executive summary ............................................................................................................... 1

1. Introduction ..................................................................................................................... 3
2. Climate-related risks to financial stability ......................................................................... 4
2.1. Physical risks ....................................................................................................... 5
2.2. Transition risks ................................................................................................... 12
2.3. Liability risks ...................................................................................................... 16
3. Transmission and amplification mechanisms ................................................................ 17
3.1. Changes in the pricing and management of financial risks ................................. 18
3.2. Potential for pro-cyclical behaviour by market participants ................................. 19
3.3. Self-reinforcing reductions in bank lending and insurance provision ................... 20
4. Cross-border transmission of climate-related risks ........................................................ 25
5. Mitigating climate-related risks ...................................................................................... 28
5.1. Actions by financial institutions ........................................................................... 29
5.2. Financial authorities’ actions .............................................................................. 31
6. Next steps ..................................................................................................................... 32

iii
Executive summary
This report discusses the potential implications of climate change for financial stability. It
investigates channels through which climate-related risks might impact the financial system. It
also examines potential mechanisms within the financial system that might amplify the effects of
climate-related risk as well as the cross-border transmission of risks. The report draws on
existing work by the official and private sector. Such work is, in places, nascent in its
consideration of risks to financial stability. In places, therefore, the report raises issues that go
beyond those discussed in the existing literature.

Risks to financial stability from climate change can be divided into physical and transition risks.
The value of financial assets/liabilities could be affected either by the actual or expected
economic effects of a continuation in climate change (physical risks), or by an adjustment
towards a low-carbon economy (transition risks).

Current central estimates of the impact of physical risks on asset prices appear relatively
contained but may be subject to considerable tail risk. The manifestation of physical risks –
particularly that prompted by a self-reinforcing acceleration in climate change and its economic
effects – could lead to a sharp fall in asset prices and increase in uncertainty. This could have a
destabilising effect on the financial system, including in the relatively short term. Market and
credit risks could also be concentrated in certain sectors of the real economy and geographies.
Disruption could also occur at national level. Some emerging market and developing economies
(EMDEs) that are more vulnerable to climate-related risks, especially those in which
mechanisms for sharing financial risk are less developed, may be particularly affected.

A disorderly transition to a low carbon economy could also have a destabilising effect on the
financial system. This could be brought about by an abrupt change in (actual or expected) public
policy not anticipated by market participants, including that due to the increased materialisation
of physical risks, as well as technological developments. In such a scenario, physical and
transition risks might combine, amplifying their overall effect on financial stability. Central
estimates of the impact on asset prices of a well-anticipated transition to a low carbon economy
are relatively contained, although there are many measurement uncertainties.

Climate-related risks may also affect how the global financial system responds to shocks. They
may give rise to abrupt increases in risk premia across a wide range of assets. This could alter
asset price (co-)movement across sectors and jurisdictions; amplify credit, liquidity and
counterparty risks; and challenge financial risk management in ways that are hard to predict.
Such changes may weaken the effectiveness of some current approaches to risk diversification
and management. This may in turn affect financial system resilience and lead to a self-reinforcing
reduction in bank lending and insurance provision.

The breadth and magnitude of climate-related risks might make these effects more pernicious
than in the case of other economic risks. Moreover, the interaction of climate-related risks with
other macroeconomic vulnerabilities could increase risks to financial stability. For instance,
certain EMDEs that are particularly vulnerable to climate change are also dependent on cross-
border bank lending.

1
There are various actions that financial institutions can take – and are taking – to reduce or
manage their exposure to climate-related risks. However, some authorities find that these are
not applied systematically by most firms. The efficacy of such actions taken by financial firms
may also be hampered by a lack of data with which to assess clients’ exposures to climate-
related risks, or the magnitude of the effects described above. Actions taken by individual
financial institutions may also not by themselves mitigate broader climate-related risks to
financial stability. Robust risk management might be supported by initiatives to enhance
information with which to assess climate-related risk.

2
1. Introduction
This report discusses the potential implications of climate change for financial stability. It
investigates channels through which climate-related risks might impact the financial system. It
also examines potential mechanisms within the financial system that might amplify the effects of
climate-related risk as well as the cross-border transmission of risks. It also builds on existing
work by the FSB and other international bodies.

In doing so, it identifies how risks to financial stability from climate change may differ from risks
from other vulnerabilities or forms of structural change. The effects of climate change may be
far-reaching in their breadth and magnitude, and could affect a wide variety of firms, sectors and
geographies in a highly correlated manner. The impact of such risks may be irreversible. 1
Different types of climate-related risks (e.g. physical and transition) may also crystallise
simultaneously, which might amplify their effect on the financial system. Risks to the financial
system from climate change tend to be particularly uncertain in both their severity and the time
horizon over which they might crystallise. They may also be more dependent on measures taken
by policymakers.

It is difficult to quantify risks to financial stability from climate change precisely. The future path
of climate change and its impact on the financial system are highly uncertain and could be non-
linear over time. There is also a shortage of data through which to measure financial institutions’
exposures to climate-related risks. 2 It is unclear whether the financial system’s responses to past
climate-related shocks have amplified their economic impact. This may partly be because there
has not yet been a climate-related shock of a magnitude sufficient to trigger amplification
mechanisms. This makes it difficult to assess the significance of such amplification mechanisms.

Nonetheless, the analysis that follows attempts to give a sense of scale of risks to financial
stability, and the potential for the financial system to amplify them. Where possible, it does so by
drawing on existing work by the official and private sector. Such work is, however, rather limited
in its consideration of implications for global financial stability, both because it often pertains only
to a subset of jurisdictions and financial institutions, and it gives only limited consideration of tail
risks. In places, therefore, the report’s discussion of financial stability risks raises issues that go
beyond those supported by existing literature.

The report also draws on the FSB’s stocktake of financial authorities’ experience in including
climate-related risks as part of their financial stability monitoring. 3 It also draws on analysis by
both the official and private sectors. This includes that of the Task Force on Climate-related
Financial Disclosures (TCFD), a private sector-led Task Force established by the FSB to develop
recommendations for voluntary and consistent climate-related financial risk disclosures by
companies.

The report does not discuss the possible implications of climate change for the broader macro
economy or the functions and responsibilities of financial authorities beyond those relating to

1
See Breeden, S, Avoiding the Storm: climate change and the financial system,
2
See FSB (2020), Stocktake of financial authorities’ experience in including physical and transition climate risk as part of their
financial stability monitoring.
3
Ibid.

3
financial stability. This includes those functions relating to monetary policy, securities regulation
and portfolio management.

The report proceeds as follows. The next section sets out the ways in which physical and
transition risks from climate change can affect financial stability, as well as the potential for these
to interact. Section 3 examines how climate-related risks might be transmitted across, and
amplified by, the financial system including via its interaction with the real economy. Section 4
examines the degree to which climate-related risks might be transmitted across borders,
including through the exposures of financial institutions. Section 5 examines actions by market
participants and regulators that might go some way towards mitigating climate-related risks to
financial stability. Section 6 sets out some proposed next steps for the FSB’s work in this area.

2. Climate-related risks to financial stability

At a high level, risks to financial stability from climate change are typically divided into: 4

■ Physical risks, that is, the possibility that the economic costs of the increasing severity
and frequency of climate-change related extreme weather events, as well as more
gradual changes in climate, might erode the value of financial assets, and/or increase
liabilities.

■ Transition risks that relate to the process of adjustment towards a low-carbon

economy. Whilst such an adjustment may be a necessary part of the global economy’s
response to climate change, shifts in policies designed to mitigate and adapt to climate
change could affect the value of financial assets and liabilities.

The magnitude of these risks and the relationship between them is likely to depend not
only on the course of climate change, but also the course of action to mitigate it, including
whether any transition to a low-carbon economy occurs in an orderly or disorderly way. 5 For
example, a sudden and unanticipated policy response to climate change could reduce physical
risks, but generate a disorderly adjustment to a low-carbon economy that could prompt the
materialisation of some transition risks in the short-term. Conversely, avoiding or deferring such
an adjustment might prevent such a materialisation in the near term, but the continued increase
in emissions could lead to the increased crystallisation of physical risks (referred to by some as
a ‘hot house world’ scenario). 6

Physical and transition risks might also interact and crystallise in tandem. An increased
materialisation of physical risks could prompt more significant policy action to stem them. This
could then give rise to a disorderly transition to a lower-carbon economy, and the crystallisation
of transition risks (referred to by some as a ‘too-little-too-late scenario’). 7

4
Ibid.
5
See, for example, NGFS (2019), A call for action – climate change as a source of financial risks, p21.
6
Ibid.
7
See, for example, NGFS (2019).

4
The remainder of this section examines physical and transition risks to the financial system in
turn. It briefly examines how these can result in liability risks, when parties are held liable for
losses related to environmental damage.

2.1. Physical risks

Economic losses from natural catastrophes have increased in recent decades. The
number of some types of extreme weather events globally has steadily increased (Graph 1, LH
panel). The weight of scientific evidence suggests that such events have become more likely or
more severe due to the anthropogenic effects of climate change, and that further anthropogenic
warming will cause them to intensify. 8 Economic losses associated with such events have also
increased (Graph 1, RH panel). 9

These include losses stemming from changes in physical capital, as natural disasters destroy
infrastructure and divert resources toward reconstruction and replacement. 10 Economic losses
associated with severe weather can also impact natural capital, as, for example, rising sea levels
and other changes in climate can affect the availability of land and reduce agricultural
productivity. 11 The crystallisation of physical risks can also negatively affect human capital,
through deterioration in health and living conditions. 12 Global and regional changes in climate
can also have indirect effects, such as disruption to supply chains and reductions in the
productivity of both human labour and physical assets. They can also reduce investment, given
the prevailing uncertainty about future demand and growth prospects. 13

Absent action to reduce the effects of climate change, physical risks to the global
economy are likely to continue to increase in future. Analysis suggests that the frequency
and severity of extreme weather events might increase non-linearly and become increasingly
correlated with each other over time. 14 While this might impact a range of countries, including
those that are advanced economies, it might also lead to disproportionate losses in emerging
market and developing economies (see Box 1). 15

8
See IPCC (2014), AR5 Synthesis Report and IPCC (2018), Special report: global warming. The extent of physical risks is also
likely to be dependent future official-sector policy, as well as the development of technology that may curb the effects of climate
change.
9
See International Association of Insurance Supervisors (IAIS) and Sustainable Insurance Forum (SIF) (2018), Issues Paper on
Climate Change Risks to the Insurance Sector, July and ESRB (2020), Positively Green: Measuring climate change risks to
financial stability.
10
See IPCC (2018).
11
Ibid.
12
For example, a two-metre sea level rise triggered by the potential melting of ice sheets could displace close to 200 million people
by 2100. See J. Bamber, M. Oppenheimer, R. Kopp, W. Aspinall, and R. Cooke (2019), Ice Sheet Contributions to Future Sea-
Level Rise from Structured Expert Judgment Proceedings of the National Academy of Sciences of the United States of America
166 (23): 11195–200.
13
See IPCC (2018).
14
Some studies suggest that potential impacts could reduce global GDP by 2-3% annually by 2060; see OECD (2015), “The
Economic Consequences of Climate Change”, while IAIS and SIF (2018) note that future climate impacts may be non-linear and
increasingly correlated – with multi-annual recurrence of “1-in-100” year events. Estimates of such impact remain dispersed; see
Network for Greening the Financial System (2019), “Macroeconomic and Financial Stability Implications of Climate Change:
NGFS Technical Supplement”.
15
See G. Abel, M. Brottrager, J. Cuaresma, and R, Muttarak (2019), Climate, Conflict and Forced Migration Global Environmental
Change 54 (January), pp 239–249; and World Bank (2018), “Groundswell - Preparing for Internal Climate Migration”.

5
The impact of physical risks on the global economy has increased in recent
decades Graph 1

Number of natural loss events Estimated global economic loss from natural catastrophe
events1
Number of events US$ bn

1
Not all natural catastrophes enumerated in the chart result from climate change.
Sources: Bank for International Settlements, Banque de France and MunichRe.

Market and credit risks

Increased physical risks could result in both market and credit risks to the financial
system. Market risks – that is, the risk of reductions in the value of financial assets – could result
in losses for banks, asset owners, and other financial institutions. Market risks might also emerge
due to abrupt increases in risk premia due to uncertainty concerning financial assets’ future
payoffs. Physical risks can also give rise to credit losses due to reductions in the income – or
reductions in the profitability – of borrowers. 16 Credit risk might also result from reductions in the
value of assets used as collateral. Together, these effects could have a broad range of impacts
on the financial system, reducing the value of investments, and increasing risks to lenders and
other financial market participants.

Estimates of the impact of physical risks on financial assets vary considerably. All are
based on a number of assumptions and subject to numerous sources of uncertainty. First,
estimates depend on the assumed future path of global emissions. 17 A given increase in global
emissions results in a range of potential increases in global temperatures and associated
increases in the severity and magnitude of extreme weather events. 18 Second, the impact of
such physical risks on the global macroeconomy and financial assets is also highly uncertain
and subject to numerous modelling assumptions. 19 Third, the resulting estimated reductions in
the value of financial assets depend on the rate at which assets’ future cash flows are

16
See S. Vladimir, J. Amado, and R. Connell (2011), Climate Risk and Financial Institutions: Challenges and Opportunities, World
Bank Group.
17
See S. Hsiang and R. Kopp (2018), An economist’s guide to climate change science.
18
The extent of climate change is also likely to depend on official-sector policy, as well as the development of technology that may
curb its effects.
19
See Hsiang and Kopp (2018).

6
discounted. Estimated impacts are much larger if they are discounted at a lower rate, which
might reflect the view of a government that has a longer time horizon than some individual
investors. 20
Table 1: Estimated impact on the value of global financial assets by the year 2105 (percent)1

Initial Percentile
Expected
discount
Study increase in global
rate3
temperature2
(per cent) Mean 90th 95th 97th 99th
7.0 -0.7 -1.6 -4.0
Dietz et 2ºC
4.1 -1.2 -2.9 -9.2
al
(2016) 7.0 -1.0 -2.4 -7.7
2.5ºC
4.1 -1.8 -4.8 -16.9
5.5 -1.5 -1.9 -2.3
2ºC
EIU 3.8 -4.2 -7.5
(2015) 2010 Baseline 5.5 -2.9 -5.0 -9.7
scenario (~4ºC) 3.8 -9.7 -12.9 -30.1
Note: Blank cells indicate where no estimates are available based on these existing studies.
Source: Economist Intelligence Unit (EIU) (2015), ‘The cost of inaction’ and Dietz, S., Bowen, A., Dixon, C. and Gradwell P (2016), ‘Climate
Value at Risk of global financial assets’.
1
This analysis is based on an integrated assessment model (IAM) that links economic growth, greenhouse gas emissions and its effect on
the real economy. Estimates of the impact on asset prices are based on the assumption that, in aggregate and over the long-term, the
payoffs of a diversified portfolio of financial assets should grow at the same rate as the real economy. For further details, see EIU (2015),
and Dietz et al (2016).
2 Estimated reductions in asset values in each row of the table are based on an assumed pathway for future emissions that is consistent
with the expected increase in global temperatures since the pre-industrial era. 2.5/4ºC are expected increases in global mean temperatures
occurring under two different pathways for global emissions. A 2ºC increase in temperatures relates to a pathway for emissions under
which it is likely that increases in global temperatures will remain less than 2ºC, which is defined by the IPCC as a 66% probability of this
being the case (see IPCC. (2014), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change).
3 Discount rates are assumed to correspond to future modelled GDP growth (see note 1) in the absence of climate change plus a premium
to account for the compensation they demand for holding risky assets. Discount rates shown are those applied to near-term asset pay-
offs. A higher discount rate corresponds to a lower present value of the future losses. Under both modelling approaches discount rates
decrease towards the end of the century in line with slowing GDP growth.

Central estimates of the impact of physical risks on asset prices appear reasonably
contained, but vary considerably with the expected degree of global warming. Results of
two studies are shown in Table 1. Under a scenario where the increase in global mean
temperature above pre-industrial levels is likely to remain within 2ºC, estimates of the mean
reduction in global financial asset values are between 0.7 and 4.2% (US$ 1–6tr), depending on
the study and discount rate. 21 Under a ‘baseline’ scenario in which policies to mitigate climate
change that were in place in 2010 are extended indefinitely but there is no additional action to
reduce emissions, 22 the expected temperature increase is around 4ºC and the estimated mean

20
Such lower discount rates might be appropriate for long-run issues such as climate change, because they treat the wellbeing of
future generations on a par with that of current generations; see Economist Intelligence Unit (2015), The cost of inaction. A lower
discount rate is used elsewhere in the literature to reflect the possible longer time horizon of governments; see, for example,
Stern, N (2013), The structure of economic modelling of the potential impacts of climate change: grafting gross under estimation
of risk onto already narrow science models, Journal of Economic Literature, Vol 51, No. 3; and Nordhaus (2017), Integrated
assessment models of climate change, NBER.
21
The estimates of percentage loss in asset value are independent of the initial stock value. Absolute values are estimated by
applying the estimated percentage reductions to a global total $143 trillion in assets managed by non-bank financial institutions
at end-2015 as estimated by Financial Stability Board (2015), Global Monitoring Report.
22
This corresponds to a 6.5% reduction in emissions in 2105, relative to a world of no further action whatsoever.

7
reduction in asset prices is between 2.9% and 9.7% (US$4-14tr), depending on the chosen
discount rate. Other studies based on different methodologies estimate central case outcomes
for a reduction in financial asset values that are of a similar order of magnitude. 23

The uncertainties associated with the future path of climate change and its impact on
asset prices mean that these potential outcomes are subject to considerable tail risk. This
is shown in the right-hand columns of Table 1. For example, while the central estimate of the
reduction in asset prices associated with a 2.5ºC increase in temperatures by 2105 is relatively
modest (i.e. 1.0-1.8%), it is estimated with 5% probability that it could exceed 4.8%, and with 1%
probability that they could exceed 16.9% (US$ 24tr). Under the baseline scenario, the most
severe reported reduction in asset prices are 30.1% (US$ 43tr), which occurs with a 3%
probability by 2105 under an assumed low discount rate.

An acceleration in the progression of climate change, and in the manifestation of physical

risks, could have a destabilising effect on the financial system. The studies discussed
above assume that increased physical risks will materialise gradually over time, with the majority
of the impact on asset prices occurring in the latter half of the 21st century. Such a reduction in
asset prices may, however, occur suddenly and be more likely to have a destabilising effect on
the financial system. This could arise in part from how the dynamics of climate change may be
self-reinforcing. 24 Increases in global temperatures may have positive feedback effects, whereby
a small perturbation in temperatures alters the dynamics of climate in a way that causes further
increases in temperature. 25 This raises the possibility of rapid non-linear short-term changes in
physical, ecological and societal systems due to climate risks, whose impacts would be felt far
sooner, and may be greater, than those captured by some central estimates. This has led some
to call for more emphasis to be placed on research about the extreme tails of distributions rather
than central tendencies. 26

Market and credit risks resulting from physical risks could also be concentrated in certain
sectors and geographies of the real economy. Many of the most directly affected firms might
be those in sectors with less movable and capital-intensive infrastructure. 27 Some evidence
already suggests, for example, that the price of real estate exposed to rising sea levels is less
than that of similar properties in other locations. 28 To the extent that the prices of the assets and
liabilities of firms in these sectors are adjusting to reflect physical risks, and/or investors/creditors
are taking action to reflect them, this might reduce risks to financial stability. That said, it also

23
For example, UNEP Finance Initiative (2019), Changing Course estimates a 2.14% reduction in the value of financial assets,
but acknowledge that this would increase significantly if global emissions are not curbed.
24
See T. Lenton, J. Rockström, O. Gaffney, S. Rahmstorf, K. Richardson, W. Steffen, and H.J. Schellnhuber (2019), Climate
Tipping Points — Too Risky to Bet Against, Nature.
25
For example, the shrinkage of icepacks caused by global warming might reduce the reflection of solar radiation, thereby further
increasing global warming. Similarly, the thawing of permafrost at high latitudes due to global warming might itself release
greenhouse gases that in turn lead to further global warming. IPCC (2019), The Ocean and Cryosphere in a Changing Climate
– Summary for Policymakers.
26
Martin L. Weitzman (2011), Fat-Tailed Uncertainty in the Economics of Catastrophic Climate Change, Review of Environmental
Economics and Policy 5 (2): 275–292.
27
For example, UNEPFI (2019) finds that the impact of physical risk on the value of equities of firms in the construction sector
could be twice that of firms in the utilities sector.
28
There is some evidence that homes exposed to sea level rise already sell at a discount relative to observationally equivalent
unexposed properties equidistant from the beach. See A. Bernstein, M. Gustafson, and Lewis (2019), Disaster on the Horizon:
The Price Effect of Sea Level Rise, Journal of Financial Economics 134 (2): 253–272.

8
serves to illustrate that risks might be concentrated in certain parts of the financial system. This
could increase risks to financial stability, particularly if it triggers amplifying behaviours by some
financial institutions (see Section 3).

Significant losses from physical risks could also result in disruption at national level, and
be concentrated in certain countries. The impact of climate change is likely to be felt globally.
Nonetheless, some EMDEs may be particularly vulnerable to physical risks, given their higher
exposure and sensitivity to the effects of climate change (see Box 1). For example, in 2011,
floods in Thailand resulted in a substantial (10%) reduction in GDP, and the value of major Thai
equity indices fell by about 30% in the following 40 trading days. 29 Greater vulnerability to
physical risks might also lead to a deterioration in sovereign creditworthiness, given the effects
of natural disasters on tax revenues and fiscal expenditure (particularly in economies where risk
transfer mechanisms are weaker; see Box 1). 30 To the extent that the creditworthiness of
sovereigns, in some jurisdictions, affects and is affected by, that of financial institutions, this
might also amplify shocks to the financial system (see Section 3.1)

Box 1 – Climate-related risks in EMDEs

Many EMDEs are particularly vulnerable to physical risks from climate change. Africa and the Asia
Pacific – two regions with a large proportion of EMDEs – are amongst the most vulnerable to the
negative effects of climate change, according to the ND-GAIN vulnerability index (Graph A, LH panel). 31
Like other countries, many emerging market jurisdictions are already affected regularly by floods,
droughts and cyclones. 32 Recent analysis suggests that these physical risks have already had severe
impacts on some countries’ economies, as well as on their banking sectors. 33
Countries’ vulnerability to physical risks may be greater where there are less developed mechanisms
through which to share risk, particularly where such risks coincide with broader macroeconomic
vulnerabilities. 34 Graph A (RH panel) suggests that many of the countries that are more vulnerable to
climate change also face large macroeconomic vulnerabilities. Recovery from weather-related natural
disasters may be hindered by more limited public disaster relief funds, fiscal constraints and a lack of
well-developed insurance markets. Their share of uninsured losses in EMDEs from natural disasters is
more than twice that in advanced economies. 35 The lack of recovery funding can also increase firms’
and households’ credit risk, thereby amplifying losses borne by financial institutions.

Countries that are more dependent on fossil fuels may also be more susceptible to transition risks. 36
Exports of fossil fuels account for more than half of all exports for some countries. A disorderly transition
to low-carbon economy – including that initiated by these countries’ trading partners – could therefore

29
See IMF (2020), Global Financial Stability Report. Chapter 5: Climate Change: Physical Risk and Equity Prices, April.
30
One study finds that vulnerability to climate-related risks has already increased the average cost of debt in a sample of
developing countries by 117 basis points, translating into USD 40 bn in additional interest payments on government debt over
the past 10 years. See B. Buhr et al (2019), Climate Change and the Cost of Capital in Developing Countries.
31
Countries’ vulnerability to climate-related risks is proxied by the ND-GAIN vulnerability index – a measure of a country’s
propensity or predisposition to be negatively impacted by climate hazards; see C. Chen, I. Noble, J. Hellmann, J. Coffee, M.
Murillo and N. Chawla (2015), University of Notre Dame Global Adaptation Index, Country Index Technical Report.
32
See World Bank (2005), Natural Disaster Hotspots – A Global Risk Analysis.
33
See, for example, IMF (2019), The Bahamas 2019 Financial System Stability Assessment (other FSAP results are forthcoming)
and UNDP (2019), Human development report.
34
See E. Feyen et al. (2020), Macro-Financial Aspects of Climate Change.
35
See IMF (2020), Global Financial Stability Report April 2020.
36
See World Bank (2020), Diversification and Cooperation in a Decarbonizing World: Climate Strategies for Fossil Fuel-Dependent
Countries.

9
 leave banks and other investors bearing large losses on fossil fuel-related assets (i.e. credit and market
risk). It could also have a broader impact on government revenues and creditworthiness, particularly in
those countries whose governments rely heavily on revenues from fossil fuels. At the same time, some
EMDEs are expanding their reliance on fossil fuel assets, which could also expose those who finance
these activities to transition risks. 37

EMDEs are vulnerable climate-related risks particularly where these

tandem with other macro-financial vulnerabilities Graph A

Climate change vulnerability per region, 2006-20181 Regional and national relationships between climate-
related and other macrofinancial vulnerabilities2

1Red dots indicate the median in each period/region. Red bars represent plus/minus one standard deviation. Vulnerability to the
negative effects of climate change is measured using the ND-GAIN vulnerability index.
2 Data are as at 2017. The Economist Intelligence Unit (EIU) country risk index captures sovereign, currency, and banking risks.
Sources: EIU Credit Risk Model, Notre Dame GAIN.

In the longer run, climate-change might also lead to geopolitical tensions in some regions that might
amplify risks to financial stability. A global increase in temperature and its associated sea-level rise
could have substantial socioeconomic impacts, including increased risk of flooding in coastal cities and
lower productivity in regions close to the equator. 38, 39 Such local changes in climate may cause people
and capital to relocate on a scale that has adverse outcomes for both geopolitical and financial stability.
For example, one study finds that an increase in drought can exacerbate conflict, which in turn explained
a significant portion of asylum seeking in the period 2011-2015. 40

Insurance underwriting risks

More frequent and severe extreme weather events have resulted (and could continue to
result) in underwriting risks: that is, higher-than-expected claims against firms that
provide insurance for physical risks. Claims faced by non-life insurers with respect to certain

37
For example, the US Energy Information Administration (EIA) forecasts a 1% annual growth rate in energy-related carbon dioxide
emissions in non-OECD member countries between 2018-2050, driven by increases in the use of natural gas (+1.7%), petroleum
liquids (+1.2%), and coal (+0.6%). Another estimate projects coal demand in Southeast Asia to grow by more than 5% per year
through 2024. See US Energy Information Administration (2019), “International Energy Outlook 2019, September and
“International Energy Agency Coal 2019, December.
38
See IMF (2018), The effects of weather shocks on economic activity: what are the channels of impact?
39
See S. Hallegatte, C. Green, R. Nicholls, and J. Corfee-Morlot (2013), Future flood losses in major coastal cities, Nature
Climate Change 3, pp 802-806.
40
See G. Abel, M. Brottrager, J. Cuaresma, and R, Muttarak (2019).

10
weather-related catastrophes have increased in recent decades (see Graph 2). While part of this
increase may be due to increases in exposure (i.e. increasing value of property in areas prone
to physical risks), it may also be due to increases in severe weather events (Graph 1). 41 Physical
risks could, over time, also result in greater claims for life and health insurers, if they impact
policyholders’ mortality and demographics. 42

Economic losses resulting from weather-related catastrophes have

increased significantly1 Graph 2

Global insured (and uninsured losses) resulting from weather-related natural catastrophes (2019 prices)
US$ bn

Source: SwissRe (2020).

1 Data may not match those used in Graph 1 due to differences in the scope.

The degree to which insurance underwriting risks act as a channel of risk to financial
stability would depend in part on the degree to which they might undermine insurers’
resilience. Most non-life insurers’ products are short-term and can be repriced or withdrawn as
a result of increased physical risks on an annual basis (or more frequently). This might reduce
the impact of physical risks on the resilience of individual insurance firms. 43 That said, the
increasing frequency and severity of severe weather events has resulted – and could continue
to result – in claims that substantially exceed those witnessed historically. 44 This, combined with
uncertainty concerning the future path of climate change, could undermine insurers’ ability to
price risks adequately, particularly if the severity of physical risks were to accelerate
unexpectedly (see above). 45 This might mean that increased losses due to physical risks could
come to weigh on some insurance firms’ resilience. 46 However, actions taken by individual
insurers to reduce their exposure to climate-related risks could have negative
consequences for the financial system as a whole. If large numbers of firms significantly

41
For example, weather-related catastrophe losses accounted for over 80% of insured catastrophe losses in 2018 globally. See
ECB (May 2019), Financial Stability Review.
42
IAIS and SIF (2018).
43
This is in line with the opinions of some market participants to whom the FSB has discussed this channel of risk. See FSB
(2020), Annex 4.
44
For example, weather-related claims in Canada amounted to an average of around CAD$1bn per year between 2006 and 2016.
However the severity of the 2017 Fort McMurray wildfire brought the total liabilities to about CAD$4.5bn; see IAIS and SIF
(2018).
45
Indeed, the impact on the insurance industry from climate-related risks has been greater where risks are systematically
underestimated. See Herweijer, Ranger and Ward (2009) Adaptation to Climate Change: Threats and Opportunities for the
Insurance Industry, and IAIS and SIF (2018).
46
See Finansinspektionen (2016), Climate Change and Financial Stability.

11
increase premiums or withdraw their coverage of certain climate related risks, this might leave
households and firms without cover. This might amplify the resulting risks to financial stability
(see discussion in Section 3).

Operational risks

Financial institutions may also be exposed to operational risks as a result of climate

change. Extreme weather events could disrupt firms’ operations, and affect other firms (financial
and non-financial) to whom they provide financial services. This also may amplify risks to
financial stability (see Section 3.3).

2.2. Transition risks

Transition risks stem from the possible process of adjustment to a low carbon economy
and its possible effects on the value of financial assets and liabilities. Limiting the rise in
global average temperatures to well below 2°C above pre-industrial levels requires an
unprecedented reduction in global emissions. 47 Such a transition to a low carbon economy would
imply significant structural changes to the economy, including a major reallocation of investment
(Graph 3, LH chart). This could have a significant impact on firms involved in the production of
fossil fuels, such as coal, oil and gas, as well as other sectors whose business models rely on
using such fossil fuels or that are energy intensive (such as utilities, heavy industry, and the
transportation sector). 48 There is already evidence that the market value of equities of firms in
some heavily polluting industries is being impacted by policy measures and market trends
related to a transition to a low-carbon economy. 49

Such changes in asset prices need not, in themselves, pose risks to financial stability.
Rather, they may be symptomatic of a well-functioning financial system that channels investment
towards more climate-neutral investments. In particular, the depreciation of some assets as a
result of transition to a low-carbon economy might also be offset by the positive effects of growth
in less carbon-intensive sectors and firms. In addition, climate policies aimed at achieving the
structural economic change described above might also boost innovation and investment,
including in less carbon-intensive technology. 50 This might benefit some parts of the global
economy, and result in the increase in some asset prices. 51 Evidence of this effect on innovation
is, however, mixed. 52

47
In order to meet a 2°C target for increases in global temperatures, emissions would need to drop by 3% each year until 2030
(and by 8% to reach the 1.5°C target). See UNEP (2019), Emissions gap report 2019, November.
48
See Bank of England (2018), Transition in thinking: The impact of climate change on the UK banking sector, September.
49
For example, the Dow Jones Coal Index fell by 85% in 2011-2018 in line with significant increase in the use of natural gas for
power generation and climate-related policy measures; see Bank of England (2018).
50
See NGFS (2019).
51
This notion is called the ‘Porter Hypothesis’. See Porter and van der Linde (1995), Toward a new conception of the
environmental-competitiveness relationship, Journal of Economic Perspectives, Vol. 9.
52
See, for example, Jaffe, Newell and Stavins (2000), Technological Change and the Environment, Working Paper No. 7970,
National Bureau of Economic Research; Berman and Bui (1998), Environmental Regulation and Productivity: Evidence from Oil
Refineries, NBER Working Paper No. 6776; Gray and Shadbegian (1997), “Environmental Regulation, Investment Timing, and
Technology Choice,” Working Paper No. 6036, National Bureau of Economic Research.

12
Estimated capital relocation resulting from transition risks and potential
effects on equity valuations Graph 3

Capital reallocation in the global energy sector consistent Estimated change in value of global equities and corporate
with a 2°C pathway bonds in sectors most exposed to transition to a low-carbon
economy were a transition to begin in 2021 or 20301
USD trn Per cent

Sources: International Energy Agency World Energy Outlook (2017), Bank of England.
1 Estimated change in asset prices consistent with a rapid economic adjustment consistent with a path for emissions that limits increases
in global temperatures to less than 2°C greater than industrial levels in either 2021 or 2030. Analysis is based on IEA’s New Policies
Scenario 53 and detailed data on insurer and investment fund portfolios.

However a disorderly transition to a low-carbon economy, unanticipated by market

participants, could have a destabilising effect on the financial system. Such a disorderly
transition might be brought about by sudden changes in technology. 54 It could also be triggered
by an unexpected change in public policy. Such radical policy action might arise due to the
increased materialisation of physical risks (see discussion of the ‘hot house world scenario’
above). In such a scenario, physical and transition risks would affect the financial system in
tandem, and thereby be more destabilising. 55

A disorderly transition to a low-carbon economy could also lead to increased credit risk
in some industries. A transition to a low-carbon economy might reduce some borrowers’
capacity to generate sufficient income to service and repay their debts. 56 This could occur as
companies in carbon-intensive industries face higher operating costs (e.g. on carbon-intensive
inputs) or due to reduced demand for certain goods or services, reducing their profitability. Credit
risk could also result from reductions in the value of collateral against which transactions are
secured, particularly when such collateral takes the form of immovable or capital intensive
infrastructure that might be particularly affected by climate-related risks (see above).

53
See International Energy Agency (2019), World Energy Model, November.
54
For example, innovations in photovoltaic technology have led to an 80% fall in the cost of renewable energy from some small-
scale power generators; see OECD (2018), A Chain Reaction: Disruptive Innovation in the Electricity Sector.
55
Such an unanticipated and disorderly adjustment in asset prices as a result of transition risks has been termed a ‘climate Minsky
moment’; see Carney (2015).
56
See Pierre Monnin (2018), Integrating Climate Risks into Credit Risk Assessment Current Methodologies and the Case of Central
Banks Corporate Bond Purchases, Council on Economic Policies, December.

13
Central estimates of the impact on aggregate equity and corporate bond portfolios of a
gradual and anticipated adjustment in the value of assets and liabilities of firms in sectors
most exposed to transition risks appear relatively contained. Such estimates of the impact
of transition risks vary significantly, due to differences in the estimation of exposures to carbon-
intensive production, and in the assumed path of transition to a low carbon economy. They also
differ in terms of the scope of losses they consider. Some studies, for example, consider losses
stemming only from reductions in the value of firms’ existing capital (sometimes referred to as
‘stranded capital’). Others consider broader losses that might result from reductions in expected
future cash flows. Despite these differences, one study, for example, finds that the discounted
loss in global wealth that could result from stranded assets may range from US$1 trillion to
US$4 trillion, or less than 3% of global managed financial assets. 57 Central estimates of the
effect on financial sector aggregate equity and corporate bond portfolios are quite benign -
e.g. 2-4% of aggregate portfolio values were transition to occur by 2030 (Graph 3, RH panel).
This reflects the fact that the equities and bonds issued by firms included in this study comprise
only 7% of global listed equity and bonds. Reductions in the value of these securities alone
could, however, reach around US$ 2 trillion, or 50% of their current market value. 58 Were such
reductions in prices to occur gradually, then they might be less likely to have material implications
for financial stability.

Similarly, central estimates of some financial institutions’ exposures to the most climate-
sensitive sectors appear, in aggregate, fairly contained. In the case of the euro area, for
example, banks’ exposures to more carbon-intensive firms (as measured by firms’ carbon
emissions to total sales) are less than 5% of their large exposures (Graph 4, LH panel). 59 The
median exposure to the most carbon-emitting firms, either through lending, bond or equity
holdings, did not exceed 9% of assets in 2017. 60 Similar analysis suggests that euro-area
insurance companies, pension funds and investment funds have exposures to carbon-sensitive
sectors of between 2% and 8% of their overall portfolios (Graph 4, RH panel). 61 This might
indicate that, at a sectoral level, risks to the financial system, at least from an orderly transition
to a low-carbon economy, are relatively contained.

However, a transition to a low-carbon economy may not occur in as gradual and

anticipated a way as in the central case estimates in the studies referred to above. Policy,
technology and/or consumer preferences may shift more rapidly than is modelled in many
transition scenarios. One way this might occur, for example, is via the possible interaction of
transition risks with physical risks (see above). Unlike for physical risks, there are, however, few

57
See J. Mercure, H. Pollitt, J. Vinuales, N. Edwards, P. Holden, U. Chewpreecha, P. Salas, et al. (2018), Macroeconomic Impact
of Stranded Fossil Fuel Assets, Nature Climate Change 8 (7): 588–93.
58
Bank of England calculations using data from MorningStar, Global Data and WardsAuto.
59
See ECB (2019). Large exposures refer to firms reporting of large exposures under the European Capital Requirements
Regulation (CRR), which requires that firms report every exposure to clients or groups of connected clients where its value is
equal or exceeds 10% of the eligible capital of the institution.
60
Other studies of bank exposures in other jurisdictions also find bank exposures to most climate sensitive sectors to be relatively
small. For example, a study by the Bundesbank did not identify a significant carbon risk factor for the majority of enterprises with
which German banks and insurers have a credit relationship; see Deutsche Bundesbank (2019), Financial stability report.
61
A recent analysis of U.S. insurance companies found that exposure to the most carbon-intensive companies represented on
average 1.1% of their total balance sheet, including 1.6% of their equity portfolio and 3.5% of their corporate bond exposure;
see IMF (2020), “United States: Financial Sector Assessment Program-Technical Note-Risk Analysis and Stress Testing the
Financial Sector”.

14
existing studies that attempt to assess tail risks – including those arising from such interactions
of different risks – around these central case estimates.

Headline exposures to more polluting sectors appear fairly contained Graph 4

Distribution of large exposures of euro-area banks to firms Euro-area investment exposures to climate-sensitive
with different carbon intensities sectors
Per cent Per cent EUR bn

Source: ECB.

Moreover, the widespread effects of climate change mean that analysis of exposures at a
sectoral level might underestimate the true level of emissions involved in firms’ overall
value chains. Firms’ vulnerability to transition risks is a function not only of their operations, but
also those of their suppliers and customers. Even if a firm’s operations are not themselves
carbon-intensive, the widespread effects of climate change might mean that a transition to low
carbon economy might increase the cost of its supplies, or decrease its customer base, thereby
affecting its profitability. To the extent that these suppliers and customers are in different sectors,
this might mean that sectoral analysis underestimates firms’ true exposure to transition. This
might help explain why estimates of financial institutions’ exposures to policy-related transition
risks – though they are of a similar order of magnitude – vary to some degree, depending on the
nature of the data on which they are based. 62

62
For example, the DNB finds that fossil fuel producers represent 6% of Dutch pension fund portfolios and 1% of insurance
portfolios. EIOPA found that climate-sensitive exposures of European insurers amount to 10-13% of total assets in 2017.
Similarly, the Banque de France/ACPR estimated that approximately 10% of French insurers’ investments were in carbon-
intensive sector in 2017; see FSB (2020).

15
Exposures to firms with high emissions could be concentrated within
individual counterparties or financial institutions Graph 5

Firm-level emissions within economic sectors and loans Euro-area banks’ exposures to 40 firms with highest carbon
emissions
Emissions intensity index EUR bn
Electricity, Gas, Steam
Mining and Quarrying
Manufacturing
Transportation and Storage
Wholesale and Retail Trade
Water Supply; Sewerage
Construction
Administrative and Support Services
Prof., Scientific and Technical Activ.
Information and Communication
Accommodation and Food Service Activ.
Finance and Insurance
Real Estate Activities
Agriculture, Forestry and Fishing
Human Health and Social Work
Other Service Activities
Education
Arts, Entertainment and Recreation
1
0.1 10
0.01 100
0.001 1,000
10,000
100,000

Source: ECB.

Analysis based on sectoral exposures might also conceal concentrated exposures to a

given counterparty, or borne by certain financial institutions. Analysis based on sectoral
exposures also overlooks substantial variation in emissions within a given sector. Some firms in
sectors that might generally be considered highly exposed to transition risks have relatively low
emissions (e.g. mining), and some in sectors that might be considered less polluting have
relatively high emissions (e.g. information and communication) (Graph 5, LH panel). Such
differences might arise, for example, from variation in the degree to which firms’ production
processes and technological investment are consistent with a reduction in emissions. Exposures
to more polluting firms might also be concentrated within some counterparties (Graph 5, RH
panel). Overall, euro-area banks’ exposures to the twenty firms with largest carbon emissions
are equivalent to 20% of these banks’ reported large exposures. 63

2.3. Liability risks

Liability risks might arise when parties are held liable for losses related to environmental
damage that may have been caused by their actions or omissions. 64 To the extent that this
might reduce the value of such firms’ liabilities, it might also have implications for the financial
system. Liability risks are of particular relevance to insurance firms. This is because some such
risks can (at least in part) be transferred by means of liability protection insurance. 65

63
See ECB (2019).
64
See Bank of England (2015), The impact of climate change on the UK insurance sector, which distinguishes three types of legal
risks: failure to mitigate, failure to adapt and failure to disclosure.
65
See M. Scott, J. van Huizen, and C. Jung (2017).

16
3. Transmission and amplification mechanisms
This section examines how shocks due to the crystallisation of climate-related risks might be
transmitted through, and amplified by, the financial system.

The breadth of climate-related risks – combined with the uncertainty concerning their
timing and magnitude – might reduce the degree to which market participants are able to
properly price and manage them. For example, the widespread and uncertain nature of climate
related risks (see Section 2) might lead to increases in risk premia across a broad range of
assets. This may change – and in places, increase – the degree of co-movement between asset
prices, and reduce the degree to which financial market participants were able to diversify
exposure to climate-related risks. It might also reduce the efficacy of other channels through
which financial market participants seek to insure against climate risks (e.g. via some derivatives
markets).

It might also increase the degree to which such risks are amplified within the financial
system. The materialisation of physical or transition risks, and their effects on financial
institutions and markets, could give rise to ‘feedback loops’ within the financial system, or
between the financial system and real economy. These arise when actions taken by agents in
their individual interest interact in ways that together amplify effects on the financial system and
real economy. Such amplification has occurred in response to other types of economic shocks. 66
However, the breadth of climate-related risks might increase their perniciousness compared to
that elsewhere. For example:

■ If widespread increases in risk premia were to materialise as a result of climate-

related risks, they might cause financial market participants to behave
procyclically. Large-scale shifts in investor portfolios might amplify changes in asset
prices. This effect might be particularly prevalent where there are substantial
commonalities between investors’ portfolios or concentrations of exposures both
through certain financial products (such as derivatives). It might also be amplified by
changes in collateral values, as well as changes in perceptions of counterparty
creditworthiness and liquidity.

■ The behaviour of financial institutions – including reductions in lending by banks,

and in coverage by insurance firms – could lead to widespread reductions in their
support to the real economy. This might amplify the effects of climate-related risks on
the non-financial sector, and lower economic growth could in turn increase losses faced
by financial institutions.

Uncertainty concerning the scale of climate-related risks – and the difficulties some
financial intermediaries might face in assessing their impact – might also magnify these
effects on financial institutions.

66
During the 2008/9 financial crisis, for example, falling asset prices and slowing economic activity – combined with growing
uncertainty concerning bank creditworthiness – led to a withdrawal of bank funding. This caused further asset sales by banks
and reductions in lending, which in turn led to further falls in asset prices and slowing economic activity. This feedback
mechanism has been dubbed ‘the Paradox of Prudence’; see Markus K. Brunnermeir and Yuliy Sannikov (2016), The I Theory
of Money, August.

17
These effects are examined in turn in the sub-sections that follow.

3.1. Changes in the pricing and management of financial risks

The widespread and uncertain nature of climate-related risks mean that their
manifestation might lead to an increase in risk premia across a wide range of assets. They
might also lead to changes – and potential sharp increases – in the co-movement of asset prices.
Uncertainty around the future path of climate change, as well as its economic impact, 67 might
mean that this was unanticipated by market participants. 68 Such an unanticipated shift in the co-
movement of asset prices might challenge the effectiveness of market participants’ ability to
diversify their exposure to climate-related risks.

Possible reductions in the efficacy of insurance obtained via financial markets

Some financial markets play an important role in facilitating the diversification of investor
portfolios, even in the face of the widespread materialisation of climate-related risks. The
use of weather derivatives contracts, for example, allows risks to be borne by economic agents
that are best placed to monitor and bear them. Climate-related risks can also be transferred via
the issuance of insurance linked securities (ILS), such as catastrophe bonds of collateralised
reinsurance. 69 The size of the global market for such instruments has increased substantially in
recent decades (Graph 6), though remains small relative to the size of economic costs from
weather-related catastrophes (Graph 2). Other financial instruments may also help in managing
climate risks. Credit, equity and oil derivatives, for example, can serve to hedge the risks
associated with stranded assets, as well as the physical risks of extreme weather.

It is possible, however, that the widespread and uncertain effects of climate-related risks
could reduce the efficacy of such risk-sharing mechanisms. The efficacy of such
mechanisms relies in part on the ability/willingness of holders of ILS securities and providers of
reinsurance to continue to bear climate-related risks, and on the creditworthiness of derivatives
counterparties. Uncertainty concerning the scale of risks might also reduce the market
participants’ ability to price such financial products.

67
This difficulty may be exacerbated where firms’ business models rely on long and complex supply chains, where exposures
depend non-linearly on the severity of climate-related risks and on the capacity of firms to adapt to them, and/or materialise over
a long time horizon. See P. Bolton, M. Despres, L.A. Pereira Da Silva, F. Samama, and R. Svartzman (2020) The green swan -
Central banking and financial stability in the age of climate change, Bank for International Settlements and Banque de France,
Section 3.3.
68
See FSB (2020).
69
See Bank of England (2015), Insurance and financial stability.

18
The global market for catastrophe bonds Graph 6

Catastrophe bond & ILS risk capital outstanding Catastrophe bonds & ILS risk capital outstanding by risk or
peril
USD bn Per cent

Source: ECB.

3.2. Potential for pro-cyclical behaviour by market participants

Widespread increases in risk premia – and unanticipated changes in asset price co-
movement – that might result from the materialisation of climate-related risks might also
be amplified by the resulting sales of affected assets by investors. 70 Whilst this dynamic is
not specific to climate-related risks, it could be more pernicious if the effects of climate-related
shocks were more widespread and correlated across assets than investors had previously
expected. 71

The potential for large-scale simultaneous sales of assets in response to the

manifestation of climate-related risks might be greater if there are substantial similarities
across investors’ portfolios. One way of measuring the commonality of exposures to climate-
related risks is to compare the holdings of securities issued by firms in sectors with the highest
carbon emissions, which might be most exposed to transition risks. Graph 7 (LH panel) shows
the securities held by different types of euro-area financial institutions that are issued by firms in
three carbon-intensive sectors (manufacturing, transport and electricity supply) 72 as a proportion
of these financial institutions’ total holdings of securities issued by non-financial corporates.
Euro-area investment funds, insurance firms and banks all appear to have common and
economically significant exposures to these sectors. This may mean that a climate-related shock
affecting the value of such securities could be increased if there were widespread simultaneous
selling of securities across financial institutions in different sectors. Such common sectoral

70
See G. Girardi, K. Weiss Hanley, S. Nikolova, L. Pelizzon, and M. Getmansky Sherman (2020), Portfolio Similarity and Asset
Liquidation in the Insurance Industry.
71
See Blackrock (2019), Getting physical: assessing climate risks, and IMF (2020).
72
An analysis of euro-area firms found these sectors to have the highest carbon emissions; see ECB (2019), Chart A.3.

19
exposures also appear to be present in several jurisdictions: Graph 7 (middle and RH panels)
shows the same exposures for investment funds and insurance firms in the four largest euro-
area jurisdictions by GDP.

Commonality of exposures to securities issued by carbon-intensive sectors

by euro-area financial institutions
Percentage of holdings of securities issued by non-financial corporates in selected sectors. Graph 7

…across the entirety of euro-area …by investment funds in selected …by insurance firms in selected
financial institutions countries countries
Per cent Per cent Per cent

Source: ECB.

It is unclear whether such analysis under- or overstates the scope for shocks to the financial
system to be amplified by common holdings across sectors. On the one hand, use of sectoral
data might overstate the degree to which selling of securities in one sector results in sales of
securities issued by different firms in the same sector. Differences in production process may
mean there are substantial differences in exposure to climate-related risks of firms in the same
sector – for example, firms reliant on renewable versus non-renewable electricity. On the other
hand, sector-level analysis might understate the potential commonality of exposures across
firms in different sectors, particularly if the materialisation of climate-related risks lead to a
widespread increase in risk premia across firms whose future prospects had previously seemed
relatively unrelated (see above).

3.3. Self-reinforcing reductions in bank lending and insurance provision

The widespread nature of climate-related risks could also trigger self-reinforcing
feedback loops to arise whereby losses suffered by the financial system cause a
reduction in the financing of the real economy. 73 The intensification of climate-related risks
could trigger a widespread reappraising of the creditworthiness of large portions of the real
economy. This might reduce the willingness of firms to provide financial services, reducing
access to (or raising the cost of) bank lending, corporate finance and insurance. By further

73
See, for example, Lael Brainard (2019), Why Climate Change Matters for Monetary Policy and Financial Stability, November.

20
depressing macroeconomic prospects, this could then result in further losses for the financial
system, which in turn could lead to a reduction in finance. 74

Potential for self-reinforcing reductions in bank lending

If banks were to suffer widespread losses as a result of the crystallisation of physical

risks, this could cause an increase in their leverage and reduction in their lending
(prompted by a need to preserve their solvency). This could then amplify the shock to the
real economy. To the extent that this, in turn, results in larger losses for banks, this effect could
also be self-reinforcing, resulting in a further retrenchment in bank lending. This effect might be
exacerbated by any increase in banks’ cost of funding and reductions in profitability, which might
further reduce bank solvency and prompt reductions in lending. 75

It is unclear whether the financial system – either globally or in any particular jurisdiction
– has experienced a climate-related shock of a scale and breadth sufficient to trigger such
a feedback effect. Some existing studies suggest that bank lending has remained reasonably
resilient in the face of past natural disasters, at least where banks’ exposures are relatively
diversified across geographic regions. 76 There remains, however, the possibility that a larger
shock (or series of shocks) – for example, that affecting a large number of institutions and/or
jurisdictions and that might imply a longer-lasting impact than those of individual natural disasters
– might trigger a negative feedback loop between bank lending and the real economy.

Existing data on the behaviour of bank capital, lending, and the macroeconomy can be used to
simulate how banks might amplify a future, more severe, shock emanating from the
crystallisation of climate-related risks. Graph 8 shows the effects of two shocks arising from the
crystallisation of transition risks in the euro area. One shock involves an abrupt introduction of
policies aimed at mitigating carbon emissions. The other involves a technological breakthrough
that allows the share of renewable energy production to double over a five-year period. 77 The
policy shock results in a reduction in banks’ capital ratios and lending to the real economy. Both
effects are economically meaningful, but gradually fade over a five-year period. These results
are much less negative in the case of the innovation shock, because new technology leads to
higher potential growth and banks are willing/able to provide finance to these profitable projects.
Nonetheless, the global nature of these shocks, their potential propagation across borders
– together with the possibility that exposures are concentrated within certain financial institutions
(see Section 2) – could result in larger effects than those estimated in this euro-area exercise.

74
See NGFS (2019).
75
These dynamics are described in detail in ECB & ESRB (2020), Positively green: measuring climate change risks to financial
stability; IMF (2011), Durable Financial Stability-Getting There from Here; and Sujit Kapadia, Matthias Drehmann, John Elliott,
Gabriel Sterne (2012), Liquidity risk, cash flow constraints, and systemic feedbacks.
76
For example, Ulrich Schüwer, Claudia Lambert, Felix Noth (2019), How Do Banks React to Catastrophic Events? Evidence from
Hurricane Katrina, Review of Finance, 23(1): p75–116 show that some banks’ risk-based capital ratios increased following
Hurricane Katrina, due to the appreciation of their holdings of US government securities. Romero Cortés and Strahan (2017),
Tracing out capital flows: How financially integrated banks respond to natural disasters, show that some large banks
accommodate an increase in the demand for credit in areas affected by natural disasters, and do not reduce lending in other
areas. On the other hand there is evidence that following the August 1999 earthquake in Turkey, the average cost of Turkish
bank’s domestic lending increased as a result of losses; see Yusuf Soner Basakaya and Ṣebnem Kalemi-Ozcan (2016),
Sovereign Risk and Bank Lending: Evidence from 1999 Turkish Earthquake, June.
77
These results are based on ESRB (2020), Positively green: Measuring climate change risks to financial stability. The methodology
and scenarios are described in Vermeulen et al. (2018), “The Heat is on : A framework measuring financial stress under
disruptive energy transition scenarios”.

21
Estimated effect of climate-related shocks on euro-area GDP and bank capital
and lending
Percentage changes Graph 8

Euro area GDP Core Equity Tier 1 capital ratio Loans to the non-financial private sector

Source: ESRB.

Some climate-related risks could have wider effects on the financial system that increase
this amplification effect. For example, physical risks could materialise in a way that impairs
financial institutions’ operations (see Section 2). This might further reduce their ability to extend
credit or process customer transactions. This might occur if, for example, severe weather events
were to affect their premises, systems, communications and/or staffing (or those of their
outsourced business functions). This might further amplify the contraction in the supply of credit
to the real economy, over and above that caused by the initial losses suffered by financial
institutions. Such operational vulnerabilities might also affect banks’ customers’ ability to meet
their credit obligations, leading to a further contraction in credit. 78

A widespread manifestation of climate-related risks could also give rise to an increase in

the credit and liquidity risk of financial institutions. 79 Widespread reductions in borrower
creditworthiness – combined with reductions in collateral issues – might lead to a widespread
deterioration in financial-sector solvency. Losses due to the materialisation of climate-related
risks for one financial institution could result in deterioration of its solvency that then leads to
losses for other financial institutions. One estimate suggests that, in the case of European banks,
such second round of losses that arise via the lending between financial institutions might lead
to losses that far exceed those caused directly by the initial climate-related shock. 80

78
See M. Daly (2019), Why Climate Change Matters to Us, Speech, Federal Reserve Bank of San Francisco, November.
79
These effects are discussed in more detail in German Ministry of Finance (2016), Final report: potential impact of climate change
on financial market stability.
80
See S. Battiston, A. Mandel, I. Monasterolo, F. Schütze and G. Visentin (2017), A climate stress-test of the financial system,
Nature Climate Change 7: pp 283-288.

22
The effects of reductions in insurance provision

If the materialisation of climate-related risks were to lead to large increases in insured

losses from physical risks (see Section 2), this might reduce the degree to which
households and financial and non-financial firms can insure against these risks. This
might occur either because insurers no longer offer such cover, or increase its price to a degree
that is not economic for end-users. 81

A range of factors determine the level of insurance coverage purchased by end-users of the
financial system. There is, however, already evidence that increases in severe weather events
in some areas has left households unable to obtain insurance against such risks. 82 Globally there
remains a substantial proportion of economic losses resulting from natural catastrophes that are
uninsured (Graph 9). Such a ‘protection gap’ is not only due to the effects of climate change. 83
Nonetheless, underinsurance may worsen if more extreme climate events crystallise.
Representatives from some insurance firms have said that under certain scenarios for climate
change, exposures to some physical risks would no longer be insurable. 84 The availability of
reinsurance – through which most insurers transfer portions of their risk – could, like general
insurance also decrease (or become prohibitively costly).

Such a reduction in insurance cover might leave risks borne by households, firms and
other financial institutions in a manner that is detrimental to growth and financial stability.
A lack of insurance might reduce the viability of some businesses, and have a detrimental effect
on investment. For example, there is some evidence that recent reductions in the availability of
insurance against severe weather events in some areas has led to a reduction in mortgage
lending and a decrease in house prices beyond that commensurate with damage caused by
weather itself. 85 Reductions in the availability of insurance could also increase risks to other
financial institutions. For example, if physical assets (e.g. property) are uninsured, this may
increase the credit risk borne by banks that use such assets as collateral to secure their
lending. 86

81
In 1992 after Hurricanes Andrew and Iniki hit the US, the price of insurance against extreme weather events increased sharply,
with an a rise in premiums of up to 40% in some areas; see Carney (2015).
82
Ibid.
83
Protection gaps also occur due to lack of consumer awareness, a lack of penetration of financial services (particularly in EMDEs),
possible imperfections in reinsurance markets, and a lack of enforcement of mandatory insurance; see Kenneth Froot (2008),
On the Pricing of Intermediated Risks: Theory and Application to Catastrophe Reinsurance and Swiss Re Institution (2019),
Sigma.
84
See FSB (2020).
85
In some extreme cases, householders in the Caribbean have found storm patterns render them unable to obtain private
insurance cover, prompting mortgage lending to decrease and house values to fall; see Carney (2015).
86
For example, damage to assets serving as collateral could create losses that prompt banks to restrict their lending in certain
regions which could put downward pressure on property values, further exacerbating the financial impacts of physical events;
see M. Scott, J. van Huizen, and C. Jung (2017).

23
Insured and uninsured losses from natural catastrophes Graph 9
USD bn in 2019 prices Per cent

Source: SwissRe.

Lack of adequate private insurance provision has led some governments to step in and
assume some risks. One early example is the National Flood Insurance Program in the United
States. This program was created in 1968 due to the lack of availability of private insurance and
sustained increases in federal disaster assistance due to floods. 87 Another example is FloodRe
– a joint initiative between the UK government and insurers, which allows insurers to pass the
flood risk element of home insurance policies to the government for a fixed price. 88 This ensures
that households at high risk of flooding continue to have insurance cover. Some governments
may, however, be discouraged from assuming such risks given constraints on their fiscal
expenditure and potential implications for sovereign creditworthiness (see Section 3.1).

Interaction of financial institutions’ and sovereigns’ creditworthiness

Climate-related risks could also lead to a self-reinforcing deterioration of prospects for

sovereigns as well as the financial sector. 89 Significant uninsured losses from physical risks
have – and could in future – result in economic disruption at a national level, reducing tax
revenues and increasing fiscal expenditures (see Section 2). 90 Increased sovereign default risk
can lead to losses for banks and insurers via their holdings of government bonds and of other
credit to governments and government agencies. 91 The resulting deterioration in the outlook for
the financial sector might cause a further deterioration in sovereign creditworthiness (i.e. a

87
See NAIC (2020), Flood Insurance, October.
88
Note that Flood Re is intended to be transitional (until 2039), by which point insurance pricing will need to reflect flood risks
(rather than being subsidised). Other examples include Consorcio de Compensacion de Seguros in Spain, the Natural
Catastrophe Insurance of Iceland, the Danish Storm Council, or the Florida Hurricane Catastrophe Fund.
89
See Serhan Cevik and João Tovar Jalles (2020), This Changes Everything: Climate Shocks and Sovereign Bonds, IMF (June),
show that countries that are vulnerable to climate change, as well as developing countries with weak adaptation capacity, have
higher sovereign bond yields and spreads.
90
See IMF (2016), Small states’ resilience to natural disasters and climate change – role for the IMF, December, finds that natural
disasters have an increasingly negative impact on growth, fiscal balances and trade in smaller states.
91
See Stefano Battiston and Irene Monasterolo (2019), A Climate Risk Assessment of Sovereign Bonds, estimate that, under
some scenarios, a disorderly introduction of climate policies, not fully anticipated and priced in by investors, would have a non-
negligible impact on insurers’ sovereign bond portfolios.

24
sovereign-bank nexus). There is evidence that, in some jurisdictions, this dynamic has taken
hold in response to previous large-scale shocks to the financial system.

The amplifying role of uncertainty

Difficulties in assessing the impact of climate-related risks might increase the

perniciousness of the dynamics described above. Complex and low probability risks may
not be fully internalised and reflected in the actions of financial intermediaries. 92 This could lead
to excessive exposures and mispricing of those risks, with potential for significant impacts and
asset price falls when risks materialise. Similar dynamics could also arise if banks or other
financial intermediaries were using off-the-shelf models to assess their exposure to climate-
related risks. This might lead them to underestimate risks, and/or lead to greater correlation in
their exposures and selling behaviour, 93 particularly where models are based on past data. 94 For
example, there are questions whether climate-related risks are properly built into banks’ credit
approval processes, particularly with respect to longer term lending. 95

Lack of information about financial institutions’ exposure to climate risk could further
contribute to the amplification of risks. A financial intermediary that is uncertain about the
impact of climate shocks on its counterparties might decide to withdraw from existing
relationships when facing heightened risk. 96 Similarly, incomplete/asymmetric information about
exposures can give rise to contagion where different sectors or intermediaries are bundled
together by other market participants due to a lack of information to properly differentiate across
counterparties. 97

4. Cross-border transmission of climate-related risks

Shocks to the financial system from climate change can also be transmitted across
borders. This could give rise to global contagion that further amplifies the consequences of
these shocks through real and financial channels.

One way in which climate-related shocks could be transmitted across borders is via the
co-movement in risk premia on assets exposed to climate-related risks in different
jurisdictions. Such spillovers might be particularly prevalent in the case of climate-related risks,
where local physical climate risks can indicate global trends, and policy measures taken in one

92
See N. Gennaioli, A. Shleifer, and R. Vishny (2012), Neglected risks, financial innovation, and financial fragility.
93
According to Bank of England (2018) only 10% of firms are taking a ‘strategic approach’ to climate risks that involves ‘actively
assessing and mitigating risks at portfolio or individual counterparty level by assessing the sensitivity of their clients to the
financial risks from a transition to a low carbon economy’.
94
Supervisory guidance from one financial authority specifically acknowledges that past data are unlikely to be a good predictor
of climate-related risks, and so expects firms to go beyond using historical data in assessing such risk; see Bank of England
(2018).
95 See, for example, N. Garbarino and B. Guin (2020), High water, no marks? Biased lending after extreme weather, Bank of
England find that lender valuations do not ‘mark-to-market’ against falls in asset prices that arise from climate-related risks, nor
do they adjust for such risk by adjusting the cost or quantity of their lending.
96
There is evidence that this was the case during the 2008-09 financial crisis; see R. Caballero and A. Krishnamurthy (2008),
Collective risk management in a flight to quality episode.
97
See G.B. Gorton and E.W. Tallman (2018), Fighting Financial Crises.

25
country can foreshadow the future path of policy in others. The materialisation of physical risks
in one jurisdiction might, for example, trigger a change in market expectations as to the
prevalence of related risks in another jurisdiction – either as to the likelihood of future physical
risks or of policy action to stem them (i.e. transition risks). Similarly, a disorderly transition
towards a low-carbon economy in one jurisdiction might prompt expectations of a similar
transition in another jurisdiction. An unexpected acceleration in policy to stem climate change in
one jurisdiction might, for example, trigger beliefs that this will be mirrored in other jurisdictions.

Another channel through which climate-related risks can be transmitted across borders
is via the exposures of financial institutions. 98 Financial institutions that provide cross-border
finance face the risk that climate-related risks in the jurisdiction to which they have provided
finance are a source of credit and market risk (see Section 2). Those receiving finance from
overseas face the risk that climate-related risks arising in their own jurisdiction give rise to the
withdrawal of cross-border finance, leaving them without funds. Cross-border investments by
insurance companies and other asset owners could also create financing strains for local equity
or bond issuers should they choose to divest from climate-vulnerable locations.

The cross-border transmission of climate-related risks via financial institutions could

serve both to amplify – and to mitigate – risks to financial stability. On the one hand, it
could give rise to diversification, by transferring risks to those best placed to bear them globally. 99
For example, a global pool of insurers is likely to absorb the losses stemming from large-scale
climate events more easily. Similarly, a diversified set of foreign lenders might be well positioned
to absorb country-specific losses. On the other hand, cross-border exposures can also amplify
risks. Cross-border lending might lead to risks affecting one country being concentrated in
another jurisdiction. To the extent that such jurisdictions, and the financial systems within them,
are not resilient to such risks, this might create financial stability risks.

Some evidence suggests that cross-border financial exposures to climate-related risks

tend to lie with those countries that are more resilient to such risks. Graph 10 (LH panel)
compares jurisdictions’ cross-border liabilities (such as those arising via lending, foreign direct
investment and portfolio investment) with a measure of their resilience to climate-related risks. 100
This analysis suggests that countries’ with larger cross-border liabilities tend to be those that are
more resilient to climate-related risks. In aggregate, therefore, according to this analysis, cross-
border exposures do not appear to be a source of amplification of climate-related risks.

Graph 10 (middle panel) compares the cross-border exposures of each jurisdiction’s

banking system (blue bars), to the same exposures weighted by borrower jurisdictions’
relative vulnerability to climate-related risks (red bars). 101 The difference between the two
approximates the degree to which banks in one country are exposed to climate-related risks via

98
These spillovers might materialise even without obvious financial linkages. The manifestation of transitions risk in a country, say
more stringent domestic environmental regulation, might also affect asset prices in other countries, as governments or the public
there might be more inclined to pursue environmental goals. Changes in perceptions of global climate risks can also affect asset
prices worldwide without cross-border financial flows or common investor base.
99
For further discussion, see FSB (2019), FSB Report on Market Fragmentation.
100
Countries’ resilience to climate-related risks is measured by the ND-GAIN Country Index. This index includes a measure of a
country’s vulnerability to climate change and other global challenges in combination with its readiness to improve its resilience
to such risks; see University of Notre Dame, ND-GAIN index.
101
Cross-border bank exposures are given by the BIS International Banking Statistics. Vulnerabilities to climate related risks are
given by the inverse of the ND-GAIN index; see above.

26
their cross border lending, relative to those in other countries. If the weighted exposures (red
bars) exceed the unweighted exposures (blue bars) for a given country, then banks in that
country are more exposed to climate risks than are banks on average across jurisdictions in
which banks are cross-border lenders.

Cross-border transmission of financial risks from climate change Graph 10

Countries’ resilience to climate-related Total cross-border bank claims, and cross-border bank claims weighted by
risks and cross-border liabilities vulnerability of borrowers to climate-related risks
By lending country By region of borrower
Log10 (USD million) USD trn Per cent

1Total cross-border liabilities consistent with countries’ international investment position. Includes direct investment, portfolio investment and
other investment in other countries. Latest available data for each country. 2Countries’ resilience to climate-related risks is given by the ND-
GAIN index (see above). 3 Total cross-border claims of banks in relevant country, given by BIS Consolidated Banking Statistics. 4 Total
cross-border bank claims weighted by the inverse of ND-Gain index of country that the recipient of funds. Values are scaled so that aggregate
values correspond to total (unweighted) cross border lending of each country. Bank lending data are as of Q4 2019; ND-GAIN index of 2018.
Sources: Notre Dame Global Adaptation Initiative; BIS, Consolidated banking statistics; IMF.

According to this analysis, cross-border bank lending does not seem to result in material
concentrations of risks in those countries whose banks are engaged in such lending. The
difference between the nominal exposures of banking systems in these countries (blue bars in
Graph 10, middle panel) and those weighted by the recipient countries’ vulnerability to climate-
related risks are relatively small, albeit with some exceptions. Cross-border bank lending might
therefore play a role in diversifying, rather than amplifying, climate-related risks across a range
of lending countries.

However, climate-related vulnerabilities might be concentrated in certain (less developed)

economies that are the recipients of cross-border bank lending (Graph 10, RH panel).
Cross-border exposures to countries weighted by their vulnerability to climate-related risks far
exceed unweighted exposures for jurisdictions in Africa and Latin America. This is consistent
with cross-border bank exposures to jurisdictions in these regions being a source of climate-risk.
The opposite is true in North America and Advanced Europe, where exposures weighted by
climate-related vulnerabilities are smaller than unweighted exposures. These differences may

27
reflect differences in economic development, and the relative prevalence of mechanisms to
transfer financial risk (see Box 1).

Cross-border lending may therefore serve to amplify climate-related risks in countries

that are the recipients of cross-border border lending. The crystallisation of physical risks in
such countries might prompt the large-scale withdrawal of funding from foreign investors. This
effect might be amplified in the case of developing economies by macroeconomic vulnerabilities,
including rapid exchange rate depreciation and wider capital outflows.

5. Mitigating climate-related risks

This section considers actions that financial institutions and authorities are taking and/or could
take to mitigate risks from climate change. It also considers the extent to which these have been
adopted by firms. While a growing number of financial institutions are considering, and taking
some action to address, climate-related risks, some authorities find that most institutions in their
jurisdiction are not yet taking a sufficiently comprehensive, strategic, and/or long-term approach
to addressing climate risks. 102

Actions taken by individual firms likely go some way in reducing their direct exposures
to climate-related risks. However, they may not reduce risks to the financial system as a
whole. As described in Section 3, some actions taken by financial institutions that are in their
individual interest could merely shift risks, or amplify risks to the wider financial system. For
example, reductions in finance and/or underwriting – including those designed to reduce
financial institutions’ exposure to climate-related risks – could worsen prospects for economic
sectors most exposed to climate risk. In addition, even if a given financial institution has mitigated
its direct exposure to climate-related risks, its exposures to other financial institutions – which
may themselves have greater direct exposures to climate-related risks – could act to amplify
risks to financial stability (see Section 3.2).

The efficacy of actions taken by firms to mitigate their exposure to climate-related risks
may be reduced by a lack of data with which to assess their clients’ exposures to climate-
related risks. Substantial progress has been made at the international level to establish
voluntary frameworks for disclosure of climate-related risks, including amongst non-financial
firms. The TCFD found, however, that while the proportion of companies disclosing climate-
related information has increased in recent years, it is still fairly low in absolute terms. 103 The
availability of data related to climate risks, along with any potential data gaps, are not discussed
further here, but will be the subject of further work by the FSB (see Section 6).

In addition, even where individual financial firms take actions to mitigate risks, climate-
related risks to financial stability might be sustained by a lack of disclosure by financial
firms of their own financial exposures to climate-related risks and risk management

102
See Bank of England (2018) and Banque de France (2019), Climate change: what are the risks to the French financial sector?
Bulletin de la Banque de France 225/8; Banca d’Italia (2020), The climate risk for finance in Italy, Questioni di Economia e
Finanza (Occasional Papers) Number 545.
103
In a survey conducted by the TCFD none of its disclosure recommendations had been implemented by more than 50% of
firms; see TCFD (2019).

28
processes. For example, a recent TCFD status report found that only 12% of banks surveyed
disclose details on their climate scenario analysis. 104 Even if financial firms are taking steps to
mitigate their individual exposures to climate-related risks, gaps in information about financial
institutions’ exposures to climate risk could further contribute to the amplification of risks. This
might have wider implications for financial stability (e.g. by causing their creditors to withdraw
from lending to exposed institutions) (see Section 3.2).

5.1. Actions by financial institutions

Some financial institutions have incorporated – and could in future continue to
incorporate – their exposure to climate risks in their investment, lending and underwriting
decisions, and integrated these risks into their broader risk management processes. This
might go some way towards limiting firms’ exposures to both physical and transition risks (see
Section 2). Examples of actions taken by some financial institutions include:

■ Heightened due diligence or categorical rules. Some financial institutions undertake

negative screening in their selection of firms to whom they lend, in which they invest or
whom they underwrite. Many financial firms that implement some sort of exclusion
policy include industries with high exposure to climate-related risks as part of such
policies. 105

■ Engagement with investees and clients. Many investors engage collectively with
companies in order to encourage them to reduce their emissions. 106 Some banks also
engage clients in the energy sector to do so. 107 Insurers are increasingly taking steps
to reduce their exposure to climate-related risks by providing incentives for customers
to mitigate such risks and guidance as to how they might do so. 108

■ Use of metrics that track (and in some cases reduce) financial firms’ exposures
to climate risk. 109 Within the TCFD framework, such metrics focus on transition risk
and include, for example, firms’ carbon footprint, carbon intensity, or financed
emissions. That said, only a minority of financial institutions define exposure limits
based upon these metrics, or use an explicit framework to measure emissions (though
some are in the process of developing such frameworks). 110 Specific environmental and
climate-risk scores have also proliferated in recent years and gained traction as a tool
to support investment decisions and portfolio allocation.

104
See Task Force on Climate-related Financial Disclosures (TCFD) (2020), 2020 Status Report, October
105
See J. Colas, I. Khaykin, and A. Pyanet (2019), Climate Change: Managing a New Financial Risk, Oliver Wyman.
106
See UN PRI (2020), PRI Climate Snapshot 2020.
107
See Boston Common Asset Management (2019).
108
See Geneva Association (2018), Climate Change and the Insurance Industry: Taking Action as Risk Managers and Investors
and IAIS and SIF (2018) pp 18-20.
109
See IIF and EBF (2020) and UN PRI (2020), PRI Climate Snapshot 2020.
110
See IIF and EBF (2020) and J. Colas, I. Khaykin, and A. Pyanet (2019).

29
 ■ Integrating climate-related risks into their assessments of borrower credit risk
and investment decisions. 111 In one survey of 45 financial institutions, most
respondents that capture climate-related risks in their assessments of borrower
creditworthiness reported doing so by adjusting other variables in their credit risk
models or making a qualitative override to their internal credit ratings. 112 A minority
directly integrate climate-specific variables into their models of credit risk. In a separate
survey of 28 large banks, only around a third integrate climate change at all stages of
the credit risk assessment process. 113, Newer climate-based models and techniques
have also emerged to support management of climate risk in investments. These
include climate Value-at-Risk (VaR), which aims to provide a forward-looking and
return-based valuation assessment to measure climate related risks and opportunities
in an investment portfolio. 114

■ Integrating climate risk management into institution-wide governance, strategies

and risk management frameworks. Many financial institutions, particularly larger
firms, place accountability and management responsibility of climate-related risks with
their Board and/or senior management. 115 Many also integrate consideration of climate-
related risks into their business strategy. 116 An increasing number of financial
institutions are integrating climate-related risks into their overall risk management
frameworks, but only a minority do so fully. 117

In 2018 in the UK, the Prudential Regulation Authority found that only 10% of banks
were taking a strategic, forward-looking view of climate risk, grounded in their long-term
interests. 118 The ECB also recently reported that “the vast majority of banks have not
yet established internal processes that allow them to systematically identify and
manage climate-related risks”. 119

■ Financial institutions may use scenario analysis to identify and assess the
potential implications of climate change, and to test the resilience of firms’
overall approach to climate-related risks. 120 Scenario analysis is typically based on
a range of pathways for the progression of climate change as well as firms’ exposures
to climate-related risks. 121 Scenarios are hypothetical constructs, not designed to
deliver precise outcomes or forecasts, but to let organisations consider how the future

111
For an overview, see NGFS (2020), Overview of Environmental Risk Analysis by Financial Institutions.
112
See J. Colas, I. Khaykin, and A. Pyanet (2019).
113
See Moody’s (2020).
114
See Asset Owners Disclosure Project (2018) Got it Covered? Insurance in a Changing Climate. It is unclear how widespread
the use of such techniques is at present, however.
115
See CDP (2019), Global Climate Change Analysis 2018; UN PRI (2020) and GARP Risk Institute (2020) Second Annual
Global Survey of Climate Related Risk Management.
116
See CDP (2019); UN PRI (2020); Boston Common Asset Management (2019); ShareAction (2020) Banking on a Low-Carbon
Future II and GARP Risk Institute (2020) Second Annual Global Survey of Climate Related Risk Management.
117
IIF and EBF (2020) and Bank of England (2018).
118
See Bank of England (2018).
119
See ECB (2020), ECB report on banks’ ICAAP practices.
120
See Task Force on Climate-related Financial Disclosures (TCFD) (2019), 2019 Status Report.
121
For an overview of these approaches, see BIS (2019), Financial Stability Institute Insight on policy implementation No.20:
Turning up the heat – climate risk assessment in the insurance sector.

30
 might look if certain trends continue or conditions are met – for example, how various
combinations of climate-related risks may affect its businesses, strategies, and financial
performance over time. According to some surveys, 16-30% 122 of financial institutions
are currently implementing some degree of climate scenario analysis and about the
same proportion are working on doing so in the near term. 123

The major credit rating agencies (CRAs) also now consider environmental, social and
governance (ESG) factors, including climate risks, in their ratings, to varying degrees. 124
Climate risk is a driver of ratings changes in certain sectors, including for sovereigns exposed to
climate risk, utilities and power companies, and fossil fuel sectors. 125 However, other evidence,
such as work by the Principles for Responsible Investment, has identified challenges to the
effective integration of climate risks into credit ratings. 126 These include long and uncertain time
horizons, the variability of potential policy measures, and lack of standardised or consistent
disclosures by companies.

5.2. Financial authorities’ actions

Some financial authorities are exploring possible approaches to reduce climate-related financial
risks. They are generally less developed than tools to address other types of financial risk,
however.

Microprudential policy

Financial authorities are beginning to assess how climate-related risks are managed by
financial firms, and to take actions to encourage firms to mitigate such risks. 127 Most
financial authorities have engaged in awareness raising of climate risks and many have surveyed
the institutions they supervise to understand how they manage climate risk. Some authorities
have issued or are in the process of issuing supervisory expectations, which tend to cover some
set of institutional risk management elements (i.e., governance, strategy, scenario analysis,
and/or risk management; see Section 5.2). Some financial standard-setting bodies are also
starting to work on supervisory guidance related to climate risk.

122
See FitchRatings (2020), ESG Credit Trends 2020; Colas et al (2019); UN PRI (2020).
123
See J. Colas, I. Khaykin, and A. Pyanet (2019) and TCFD (2019).
124
See FitchRatings (2020), Climate Change Impacts on Sovereign Ratings: A Primer; Moody’s Investment Services (2018),
Heat map: 11 sectors with $2.2 trillion debt have elevated environmental risk exposure; S&P Global Ratings (2017), How
Environmental and Climate Risks and Opportunities Factor Into Global Corporate Ratings - An Update.
125
One credit rating agency identified 10 sectors where carbon transition risk is already a significant credit consideration and a
further 22 sectors with moderate credit exposure to environmental and climate risks. Moody’s Investment Services (2018)
Heat map: 11 sectors with US$2.2 trillion debt have elevated environmental risk exposure. Another found that climate change
materially influenced approximately 10% of ratings actions in recent years; see S&P (2019), Environmental, Social, And
Governance: Delays In Addressing Global Warming And The Longer-Term Ratings Implications.
126
See UN PRI, Credit risks and ratings initiative.
127
See FSB (2020).

31
Macroprudential policy

Three-quarters of financial authorities are considering climate risks in their financial

stability monitoring and some are starting to quantify those risks. 128 Scenario analysis (see
above) can be used by financial authorities to quantify the totality of exposures of financial
institutions to climate-related risks within their jurisdiction (this is sometimes called a ‘climate
stress test’). 129 Such estimates are similar to firm-specific scenario analysis described above,
but are based on a common scenario (or scenarios) specified by financial authorities, which
allow for a comparison of results across firms. To date, such scenario analysis is being
developed by the Bank of England, Banque de France/ACPR, and the European Central
Bank. 130 Other financial authorities have said they plan to run analyses along similar lines in the
future. 131

Some authorities have also voiced support for considering macroprudential policies to
mitigate the climate related risks to financial stability. 132 Due to the inherent uncertainty of
climate risks and its systemic nature, consideration might be given to the possible use of
macroprudential instruments, which may be used to build resilience.

6. Next steps
The FSB will conduct further work to assess the availability of data through which climate-related
risks to financial stability could be monitored, as well as any data gaps.

This further work will be completed by October 2021.

128
Ibid.
129
Ibid.
130
Ibid.
131
See M. Carney (2020), The Road to Glasgow.
132
F. Villeroy de Galhau (2018), Green finance – A new frontier for the 21st century, Bank of International Settlements; P.R. Lane
(2019), Climate Change and the Irish Financial System, Economic Letter Vol.2019 No.1.

32