2. Taxonomy for translating physical and transition risk into sovereign risk

Although much of the scientific and media attention is understandably focussed on extreme weather events, sea-level rise and ecosystem collapse, there are several transmission channels through which climate change can affect public finances and sovereign risks (see, for instance, Figure 1 and Volz et al., Reference Volz, Beirne, Ambrosio Preudhomme, Fenton, Mazzacurati, Renzhi and Stampe2020). It is now increasingly recognised that the fiscal consequences of climate change and the policy responses to it are not limited merely to the direct physical damages.Footnote ¹ Indeed, some of the most significant near-term impacts on the global economy and public finances, especially in advanced economies, will derive from the climate transition as much as the physical risks from climate change itself. The ‘transition’ reflects the process of decarbonisation, for example the policies, consumer preferences, litigious actions and technological development that accompany the drive to reduce emissions. Chief among these may be an accelerated pace of human capital obsolescence (and the social disruption this would inevitably bring) as a labour force adapted to the fossil fuel economy struggles to find productive roles in a low carbon world. As a result, it is important to consider the potential risks and opportunities associated with the structural transformation of the economy.

Figure 1. Climate change to sovereign risk: a review of impact pathways. Source: Adapted from Volz et al. (Reference Volz, Beirne, Ambrosio Preudhomme, Fenton, Mazzacurati, Renzhi and Stampe2020).

3. Climate risks and sovereign creditworthiness

Despite growing evidence of the risk channels outlined in Section 2, mainstream estimates of sovereign risk—especially sovereign credit ratings—have so far been slow to adjust. Although credit ratings agencies have started to discuss climate-related risks, even acquiring climate data firms, there is little evidence that sovereign ratings have been affected. This is an important consideration. The efficient market hypothesis depends on the availability of reliable information about risk. If sovereign ratings cannot be shown to respond to the growing scientific and economic evidence of the consequences of climate change, markets and investors may struggle to correctly identify, price and manage risk throughout their portfolios.

Recent empirical analyses suggest that the scientific and economic evidence is sufficient to demonstrate an effect of climate on sovereign creditworthiness. By introducing an integrated assessment model into the European Stability Mechanism’s debt sustainability analysis optimisation model, Zenios (Reference Zenios2021) finds that adverse climate effects on debt dynamics arise as early as 2030. He argues that European institutions including the European Commission, European Central Bank, European Stability Mechanism, European Fiscal Board and member states’ finance ministries should incorporate climate change into sovereign risk assessments.

Cevik and Jalles (Reference Cevik and Jalles2020a) estimate the effect of climate vulnerability on credit ratings of 67 sovereigns between 1995 and 2017, finding that vulnerability negatively affects sovereign credit ratings. However, the result is not robust to different model specifications. More fundammentally, this backwards-looking analysis cannot be used to make inferences about the future affect of climate change on creditworthiness, which is clearly the more important question for managing the public finances over the medium- to long-run.

Klusak et al. (Reference Klusak, Agarwala, Burke, Kraemer and Mohaddes2021) is the first paper which offers a forward-look into the effects of climate on sovereign credit ratings. Combining climate-economic models with observed sovereign ratings and an artificial intelligence (AI) model, they construct the world’s first climate adjusted sovereign rating for 108 sovereigns under three climate scenarios. Results indicate that climate change will exert downward pressure on ratings as early as 2030 with an increasing magnitude throughout the century (see Figure 2). Preliminary results show that if global emissions follow their recent trend, the UK could see a 1-notch downgrade (on a 20-notch rating scale) by 2030, growing to more than three notches by the end of century under a high emissions scenario (RCP 8.5). If instead emissions profiles are consistent with limiting warming to 2C, the UK’s sovereign downgrade would be limited to approximately 1 notch by 2100.

Figure 2. Climate-adjusted ratings to 2100 (RCP 8.5 vs. 2.6). Source: Klusak et al. (Reference Klusak, Agarwala, Burke, Kraemer and Mohaddes2021). The horizontal axis indicates current ratings by S&P and the thick black line represents exact matches between current and predicted ratings. The dotted lines are the best fit lines for climate-adjusted ratings under RCP 8.5 and 2.6, respectively, for 2030, 2050, 2070 and 2100 (Panels A and B). RCP 2.6 and RCP 8.5 are consistent with warming of less than 2C and 4.5–5C relative to pre-industrial levels, respectively.

4. Climate change and the cost of sovereign debt

Weak fiscal positions characterised by high debt, high deficits and low growth increase sovereign the cost of public borrowing (Barrell and Holland, Reference Barrell and Holland2009). Whilst early climate models indicated that the worst effect of climate change would accrue far in the future, sovereign debt markets may be expected to price these effects in earlier. Indeed, empirical evidence suggests that climate vulnerability has already raised the average cost of debt in climate-vulnerable developing countries. A first analysis of the nexus between climate vulnerability and the cost of capital by Kling et al. (Reference Kling, Lo, Murinde and Volz2018) showed that countries with higher exposure to climate vulnerability incur a risk premium on their sovereign debt.Footnote ⁴ Conducting a panel and principal component analysis with annual data for a sample of developing and advanced 46 countries over the period 1996–2016, Kling et al. (Reference Kling, Lo, Murinde and Volz2018) find that, controlling for macroeconomic factors, climate vulnerability has increased debt costs for a subgroup of 25 climate-vulnerable developing countries by 1.17 percentage points. In absolute terms, this translates into $40 billion in additional interest payments for 40 climate-vulnerable countries over the period 2007–2016 on government debt alone. Projections suggest additional interest payments attributable to climate vulnerability to increase to between $146 and 168 billion for the same group of countries over the next decade (Buhr et al., Reference Buhr, Volz, Donovan, Kling, Lo, Murinde and Pullin2018).

Subsequent studies have corroborated the positive effect of physical climate vulnerability on the cost of government debt. Using data for 98 advanced and developing countries over the period 1995–2017, Cevik and Jalles (Reference Cevik and Jalles2020b) replicate the methodology and measures employed by Kling et al. (Reference Kling, Lo, Murinde and Volz2018) and confirm their findings. Using refined measures of climate risk and resilience and a higher data frequency for 40 advanced and emerging economies, Beirne et al. (Reference Beirne, Renzhi and Volz2021a) also confirm the findings of a climate risk premium by Kling et al. (Reference Kling, Lo, Murinde and Volz2018). Moreover, Beirne et al. (Reference Beirne, Renzhi and Volz2021a) develop a set of panel structural VAR models that indicate the reaction of bond yields to climate risk shocks becomes permanent after around 18 quarters with high-risk economies experiencing the largest permanent effects on yields. Beirne et al. (Reference Beirne, Renzhi and Volz2021b) conduct an in-depth analysis of the link between climate change and sovereign risk in Southeast Asia, one of the regions most heavily affected by climate change. Both the country-specific and the panel estimates for six Southeast Asian countries with monthly data for the period 2002–2018 show that greater climate vulnerability appears to have a sizable positive effect on sovereign bond yields, while greater resilience to climate change has an offsetting effect, albeit to a lesser extent.

Looking at climate transition risk, Battiston and Monasterolo (Reference Battiston and Monasterolo2020) employ a financial pricing model with forward-looking simulations to estimate effects of climate risk on the sovereign bond yields of 34 countries. They find that economies that are highly exposed to carbon-intensive sectors incur higher yields on their sovereign bonds. Klusak et al. (Reference Klusak, Agarwala, Burke, Kraemer and Mohaddes2021) use an AI-based approach to estimate the additional cost of sovereign and corporate debt resulting from climate-driven downgrades under various warming scenarios. Under the low emissions scenario (RCP 2.6) the UK’s 0.92-notch downgrade by 2100 would amount to an increase in the annual interest payment on public debt of $2.0–3.0 billion and an increase of $250–440 million on corporate debt. Under the high emissions scenario (RCP 8.5), this rises to $7.54–11.31 billion and $0.94–1.65 billion in additional interest payments on sovereign and corporate debt respectively. Relative to 2020–2021 levels, the increase in annual debt servicing costs for UK central government debt (net of the Asset Purchase Facility) ranges from 9 to 13.6 per cent under RCP 2.6–33.8–50.7 per cent under RCP 8.5 (House of Commons, 2021). These preliminary estimates should be considered severe underestimates because the effect of downgrades is placed on the pre-pandemic level of debt. The figures would grow with the level of debt over the century.

For comparison, figures for G7 plus China show that climate-induced downgrades could increase the cost of sovereign debt by $18.18–27.27 billion ($113.93–170.89 billion) under the RCP 2.6 (RCP 8.5) scenario. The additional annual interest payments of corporations will reach $5.35–9.36 billion (20.75–36.31 billion) by 2100.

5. Measuring the impact of a transition to net zero on the UK economy and public finances: Using OBR and CCC projections

In this section, we demonstrate how the transmission pathways set out in Section 2 map onto the likely impacts of the UK’s transition to net zero on the national economy and the public finances, drawing insights from the UK OBR’s recent Fiscal Risks Report (OBR, 2021). The report uses the highly credible UK Climate Change Committee (CCC) figures to conclude that meeting net zero will leave GDP 1.4 per cent lower than the hypothetical counterfactual path in 2050 and public debt 21 per cent of GDP higher. It is important to note that the report makes no attempt to account for the positive impact on UK debt from forgone climate damages and risks, reflecting the fact that the UK is a marginal player when it comes to global emissions. The focus is therefore solely on transition costs from the UK’s efforts to reduce emissions and efforts to adapt to a changing climate and associated risks. Nevertheless, the headline analysis needs to be interpreted carefully and masks a much richer story.

Firstly, the likely costs and benefits are unevenly distributed through time. Higher net spending outlays early on are likely to generate higher net revenues down the line, thereby promoting rather than eroding public debt sustainability over the longer term.

Early fiscal outlays must be viewed as investment costs, which expand the public sector’s net worth by generating net revenues or public returns down the line. Indeed, the CCC numbers which the OBR uses show that by 2050, operating costs in most sectors are expected to be lower than under a net zero scenario relative to baseline, a trend that would be expected to continue through the second half of the century with the potential to boost the public finances. Starting with cars, and moving progressively sector by sector, the falls in operating costs become so large that in total they begin to offset the additional capex costs (see Figure 3) from around 2040 net ‘costs’ become negative.

Figure 3. (Colour online) Capital and investment costs and operating cost savings in the Balanced Net Zero Pathway. Source: UK Climate Change Committee (CCC), Sixth Carbon Budget (2021).

By truncating the cost benefit analysis at 2050, and assessing the integral only over the next 30 years, the OBR analysis includes the bulk of the (front-loaded) investment cost, but excludes decades and centuries of future net gains. To illustrate the point, consider a policymaker charged with assessing the case for education spending. The cost savings over this period from eliminating public spending on primary, secondary and tertiary education might well exceed the forgone net revenues from lower wages and social disruption in the first 30 years. Yet, a policymaker would be rash to scrap all education spending knowing that a stream of positive net returns will accrue for many decades into the future, just because they fell outside the analytically arbitrary cut-off date for the projection.

The second question surrounds the presumed impact of early action on productivity growth. The baseline numbers assume that additional investment to address climate impacts and transition requirements crowds out alternative investment and leaves UK productivity lower by mid-century. But this is far from a foregone conclusion. An accurate prediction of the impact of a low-carbon transition on GDP in 30 years’ time is, of course, unrealistic. But it is important to note that the numbers quoted by the OBR are hugely sensitive to changes in the economy’s dynamic productive potential. The OBR’s own calculations suggest a boost to annual productivity growth of less than 0.05 per cent would potentially wipe this out this loss.Footnote ⁵

Such a productivity gain is not trivial, were it available. But it is important to recognise that this is not an absolute gain, it is a hypothetical gain assessed against a baseline scenario in which the UK charts a future based on high-carbon goods and services. However, such a scenario is highly vulnerable to negative output shocks associated with stranded and devalued assets. All told, there is evidence to suggest that the OBR are being overly conservative. Over time, investing in cleaner, more efficient and innovative production is likely to be more productive than relying on the technologies of the previous century (Way et al., Reference Way, Ives, Mealy and Farmer2021). This is even before account is taken of climate damages, which can mount much more quickly as productivity growth is eroded through endogenous impact of destroyed or devalued capital (Dietz and Stern, Reference Dietz and Stern2015).

The economics of innovation and productivity means the cost of meeting various decarbonisation and resource-management targets are not only hard to predict, but that they are path-dependent (Aghion et al., Reference Aghion, Hepburn, Teytelboym and Zenghelis2014). Initial conditions and decisions made today, in changing behaviour and generating innovation, matter greatly. Relatively small interventions can generate a virtuous cycle of innovation, investment and falling costs of green technologies as evidenced in the renewable energy and vehicle sectors. By contrast, lock-in of resource- and carbon-intensive infrastructure, behaviour and institutions increases the cost of attaining sustainable pathways.

The sheer scale of the low-carbon transition generates productivity enhancing network effects and economies of scale in production and discovery. These are already apparent having caught most commentators by surprise. The price of solar photovoltaic has dropped by 83 per cent since 2010, a period over which the price of wind turbines fell nearly 40 per cent (Liebreich, 2018). These technologies are increasingly producing cheaper power than most coal- and many gas-fired power plants (Figure 4).

Figure 4. (Colour online) Levelised costs of electricity (constant 2019 US dollars per kWh). Source: Committee on Climate Change, 2020.

The important lesson from Figure 4 is not just that renewable energy now competes with fossil fuels (net of infrastructure and subsidies), but rather how this transition came about and what lessons can be learned to stimulate further innovations. The pace and extent of cost reductions required a series of several interconnected, amplifying feedback mechanisms. These include:

• • Learning-by-doing. With deployment, lessons are learned on how to manufacture, distribute, install, run and maintain equipment more efficiently.

• • Economies of scale in production and distribution. Unit cost fall as production and distribution ramp up once large, fixed costs are incurred.

• • Network and coordination effects. This reflects the greater advantages of moving in tandem with others, such that the gains are higher the more economic agents are taking similar action.

• • Sector spill-overs. Not only have sustainable technologies been shown to have predictably higher cost-reducing learning rates, but they have also been shown to have positive productivity spill-overs into other sectors of the economy. Using data on 1 million patents and 3 million citations, Dechezleprêtre et al. (Reference Dechezlepretre, Martin and Mohnen2014) suggest that productivity enhancing spill-overs from low-carbon innovation are over 40 per cent greater than from conventional technologies (in the fossil fuel energy production and transportation sectors).

• • Social and institutional feedbacks. Acceptable standards of behaviour and social norms are liable to change (Licht, Reference Licht2008; Ostrom, Reference Ostrom2000; Posner, Reference Posner1997).

• • Evolution of consumer behaviour. Consumer tastes are key in the attribution of future value to goods and services and consumers routinely influence one another, leading to positive feedbacks and crowd effects and changing consumption patterns. The heterogeneity of real-world consumer behaviour is readily inferred from standard innovation diffusion theory where technology adoption typically follows s-shaped patterns starting with innovators and early adopters, through the majority to finally affect laggards (Mercure et al., Reference Mercure, Chester, Lynch, Jarvis and Jarvis2021)

• • Expectations. People’s changing expectations with shape the nature and pace of change. Perception of new technologies as superior leads to behaviour change which itself facilitates their successful adoption (Van der Meijden and Smulders, Reference Van der Meijden and Smulders2017). Expectations that the cost of a net zero transition will be inevitable and affordable can become self-fulfilling (Zenghelis, Reference Zenghelis2021a).

Figure 5 demonstrates how these economic phenomena interact to drive innovation and transition. Note that initially, low levels of installed capacity are associated with high costs. This means markets alone were highly unlikely to deliver the necessary transition. Early-on, government leadership (subsidies and signals) was needed to drive initial investments. Learning-by-doing and economies of scale drove efficiency gains as installed capacity grew. Notably, coal has not become significantly cheaper over the same time period. Instead, social norms and consumer behaviour have driven demand for renewable energy, and in many parts of the world, for the complementary purchases such as electric vehicles. Strong policy signals on the phase-out of internal combustion engines have further shaped expectations, and manufacturers are responding by increasing the range of EVs available on the market.

Figure 5. (Colour online) Renewable electricity costs fall as installed capacity rises. Source: Our World in Data (2021).

Whilst it remains to be seen whether the story of renewable energy is an exception—the timescales, cost profiles and innovation pathways in other sectors may differ—it is clear that an early start, corresponding to the OBR net zero scenario, is therefore likely to cut the costs of transition and improve economic performance compared with delay and indecision. It may also be pivotal to shaping the long-term sustainability of the public finances.

The potential net benefits from early action would be expected to be higher if the UK moves in tandem with the other countries in imposing ambitious policies to generate innovation at scale. In such an environment, the net gains from low carbon investments would be far higher and the estimated increase in borrowing costs and corresponding downgrades in sovereign ratings reported in Sections 3 and 4 would also be far smaller, testifying to the importance of global coordination. Although Klusak et al. (Reference Klusak, Agarwala, Burke, Kraemer and Mohaddes2021) identify the possibility of climate-driven downgrades even under RCP 2.6, their model makes no adjustment for early investments in adaptation. As the treatment of climate change in sovereign ratings methods becomes more sophisticated, one would expect investments in adaptive capacity to reduce the downward pressure of climate on ratings.

6. Competitive position and macroeconomic context

The transition will also affect any advanced economy’s competitive position in changing world markets. Most European countries are small to medium sized open economies. Productivity and growth over long time periods will be driven by comparative as well as absolute advantages in new markets and these are not set in stone but are shaped by policy. This goes beyond comparing absolute investment and operating costs.

Falling behind on policies and investments may induce significant displacement of high productivity (and high productivity growth) activities to other countries, if the world shifts to low carbon resource efficient markets (e.g. as a result of global policies hostile to high carbon activities and undercutting technologies outcompeting high carbon incumbents).

Hidalgo et al. (Reference Hidalgo, Klinger, Barabási and Hausmann2007) and Mealy and Teytelboym (Reference Mealy and Teytelboym2021) use network analysis to demonstrate that it is easier for countries to become competitive in new green products that require similar production capabilities and know-how to existing sectors. This helps demonstrate how green transitions are highly path-dependent: countries which successfully invest early in green capabilities have greater success in diversifying into future green product markets. This also has potential implications for industrial policy seeking to develop comparative advantages in fast growing new sectors. The potential fiscal risks should be clear.

Other studies focus on ‘revealed technological advantage’ to determine where opportunities for sustainable growth and recovery might reside. Martin et al. (Reference Martin, Unsworth, Valero and Verhoeven2020) compare broad categories of technologies and find that the UK is relatively specialised in wind and ocean energy, as well as biotechnology and pharmaceuticals. They argue that the returns to public investments in these technologies are likely to be disproportionately high, improving the UK’s longstanding poor productivity performance and competitiveness while also addressing large-scale disparities across and within regions (see also Stern and Valero, Reference Stern and Valero2021).

The impact of adopting policies to transition the economy to net zero, and associated risks, is a function of the broader global macroeconomic context. Recent evidence on productivity strongly suggests the world needs markedly higher investment (Zenghelis, Reference Zenghelis2021b). Weak private confidence has meant ex ante desired investment lagged desired saving for more than a decade, driving real risk-free interests rates to below zero.Footnote ⁶

Negative rates in turn push the efficacy of monetary policy to its limits, fuelling asset price rises that disproportionately benefited the rich. A fundamental policy error following the crash of 2008 was introducing austerity programmes to balance budgets prematurely. Reduced public borrowing to invest exacerbated the problem of excess net saving (Auerbach and Gorodnichenko, Reference Auerbach and Gorodnichenko2017). Research suggests European countries suffered a decade of lower growth than the United States, which adopted a far more relaxed fiscal stance (Jordà and Taylor, Reference Jordà and Taylor2016).

7. Conclusion and policy imperatives

Climate change poses new risks which will affect the financial system, fiscal sustainability and sovereign debt markets. At the same time, attempts to meet agreed climate targets will require an unprecedented structural transformation of the global economy over the next two or three decades. The changing landscape of risks warrants new risk management and hedging strategies to contain climate risk and minimise the impact of asset stranding and asset devaluation. Yet, conditional on action being taken early, the opportunities from managing a net zero transition would substantially outweigh the costs.

Public investment in complimentary assets alongside supportive policies and institutions are prerequisites to ‘crowd in’ sustainable capacity and bolster the structural resilience of the economy. This historic growth opportunity has been reflected in large stimulus and recovery packages adopted in the Europe, the United States and other advanced economies. Each has placed clean investment at its heart.

Yet the call for public investment to boost productive capacity, following an unprecedented government response to the global pandemic, understandably has raised concerns about fiscal space. Public debt relative to GDP is already close to historic highs in many European countries. Combined with record broad money supply growth it has also rekindled fears of the return of inflation (Haldane, Reference Haldane2021).

These justifiable fears notwithstanding, the risks from higher public debt in the short term, are far lower than the likely benefits (Stern and Zenghelis, Reference Stern and Zenghelis2021). There are increasing opportunities associated with a public drive to steer a zero-carbon economy which can crowd in investment and expand capacity. The IMF Fiscal Monitor for October 2020 argued that an additional £1 in public borrowing to invest in ‘job-rich, highly productive and greener activities’ would generate an additional £2.7 of additional output (IMF, 2020; Gaspar et al., Reference Gaspar, Mauro, Pattillo and Espinoza2020). This marks scope for strong crowding-in. The markets agree: they continue to lend to governments at real interest rates which remain at near-record lows. The most promising way to bring down public debt in the medium term is to borrow to invest now (Chudik et al., Reference Chudik, Mohaddes, Pesaran and Raissi2017).

A fundamental feature of transition and climate change is that the infrastructure, skills and ideas of the last century can quickly become a liability. Resource- and carbon-intensive physical, human and knowledge assets risk becoming devalued or obsolete. Early indications suggest that this may already be affecting some asset prices, even as fossil fuels continue to play a key role in the economy. However, due to the difficulty of translating climate risks to financial risks, markets have been slow to respond. It is for this reason that the Financial Stability Board’s Task Force for Climate Related Disclosure is calling for mandatory disclosure and stress testing of assets (TCFD, 2017).

It is increasingly recognised that growth is not only compatible with sustainability, but also requires it (Zenghelis, Reference Zenghelis2021b). Recent studies have highlighted the urgent need to strengthen the quality of natural assets (New Climate Economy, 2018). They also point to numerous near-term opportunities associated with tackling particulate pollution, congestion, ill-health, biodiversity loss, inefficiency and waste and building more attractive, liveable cities. This is especially true after the COVID pandemic, which marks a period of dislocation and change affording a potential opportunity to lock in to pathways which create durable and resilient jobs (Stern et al., Reference Stern, Unsworth, Valero, Zenghelis, Rydge and Robins2020).

A recent survey of 231 economic experts from G20 countries, including central bank and finance ministry officials, found that major fiscal recovery packages that focussed on clean R&D spending, clean energy infrastructure, connectivity infrastructure, building upgrades, energy efficiency and investment in green spaces were considered best placed to boost long term growth (Hepburn et al., Reference Hepburn, O’Callaghan, Stern, Stiglitz and Zenghelis2020).