Physical shocks and economic resilience: a central bank perspective − speech by James Talbot

Physical shocks increasingly matter for central banks

Governor, distinguished guests, I am sorry that I cannot be there with you in person at such a prestigious event. It is very much an honour and a pleasure, and my thanks to Dr. Fabrizio López-Gallo Dey for inviting me to speak today.

The physical impacts of climate change are becoming larger and more proximate, and their short-term economic impacts are also increasing. Annual global damages from weather-related events have more than doubled since the early 2000s, reaching over $200bn in 2025^{footnote [1]}.

In light of that, the economic effects of climate change are increasingly on the radar of central bankers. When extreme weather events affect production, output and inflation, that matters for our primary goal of maintaining price stability.

Let me take two examples that are close to home:

Increasingly high summer temperatures reduce water levels on key transport arteries for industrial trade, like the River Rhine in Germany. When water levels fall, ships reduce the amount of cargo they carry, disrupting the movement of goods and increasing transport costs.

Such exceptionally low water levels have become a recurring feature in recent years. The economic effects have also been material: recent research estimates that 30 days of low water on the Rhine reduces German industrial production by around 1%^{footnote [2]}.

In October 2024, extreme rainfall caused devastating flash floods in Valencia. Aside from the tragic loss of human life, economic losses amounted to over 20% of regional GDP. Spanish Q4 GDP growth was reduced by between 0.1 and 0.2 percentage points^{footnote [3]}.

The economic impact would have been larger were it not for the high level of insurance protection for affected households and businesses. Around 250,000 insurance claims were filed following the floods and Spain's Insurance Compensation Consortium had paid more than €4bn in compensation by late 2025, making this the largest event the scheme has ever managed^{footnote [4]}.

Even countries like the UK, which may be less susceptible to large physical shocks, can be affected by the impact of events abroad.

Over the past four years, as chair of the NGFS Workstream on Monetary Policy, I have worked together with more than sixty central banks to set out a framework to assess the impact of climate change on the economy, to measure its effects and to determine what monetary policymakers should do about it.

I’ll talk today about how the physical impacts of climate change can affect the economy in the near term; how we diagnose their magnitude, persistence and propagation; whether their effects belong in the central forecast or in the risks around it; and how that diagnosis informs monetary policy judgements. We set out each of these steps in an NGFS guide for monetary policymakers that we published earlier this year^{footnote [5]}.

I also want to look beyond the near-term policy horizon. In particular, how repeated losses, adaptation, fiscal space and insurance protection gaps shape the capacity of economies and financial systems to absorb future shocks.

I will argue that these two horizons are connected: The resilience an economy has built determines the scale and persistence of the macroeconomic impact of a shock over the monetary-policy horizon. And, in turn, repeated shocks can erode that resilience over the longer term, eventually exacerbating the effects of climate shocks when they occur.

How physical shocks transmit through the economy

Let me start with how physical shocks can affect the economy.

Physical hazards can operate through several channels at once: They can damage capital. They can reduce labour supply and lower productivity. They can disrupt transport and supply chains. They can weaken demand and confidence among economic agents, tighten credit conditions and raise uncertainty.

Importantly for monetary policymakers, those channels can interact and may also move output and inflation in different directions, making the policy response more challenging.

But different physical hazards can transmit through the economy in different ways. To give some examples^{footnote [6]}:

• Heatwaves transmit as a shock to productive capacity and labour supply. They push down on GDP while increasing inflation volatility, making them relevant for monetary policy even though direct physical damage is not as visible as for a flood or storm.
• Droughts reduce productive capacity in agriculture pushing down on production and up on prices. They can raise inflation both through the effect on domestic production and through more expensive imports.
• Floods and cyclones damage the capital stock, reducing demand and investment, which can in turn push down on output and inflation. But floods can also be inflationary where they hit food supply, logistics or essential infrastructure.

So while there is no one-size-fits-all pattern for the effects of a physical shock, there are some common themes:

• Shocks are often transmitted through a disruption to production;
• That usually means that output and inflation often move in opposite directions; and
• Even when shocks occur overseas, these can be transmitted through changes to exports and commodity prices, particularly for open economies. Indeed, a 2022 study for the euro area showed that for every 1% unanticipated increase in globally traded food prices, overall inflation rose by 0.1%^{footnote [7]}.

This latter channel is also particularly important for the UK since we import around 40% of our food. In recent work at the Bank of England, we have looked at the impacts of extreme weather events on commodity prices and UK inflation^{footnote [8]}. The estimates from the model suggest that global weather-related disruption has played a non-trivial role in UK food and headline inflation over the past two decades

This makes climate-related shocks relevant to our regular inflation forecasting. Indeed, many central banks are now assessing the impact of food, energy, supply-chain, and weather-related shocks. That is because these shocks are affecting inflation and output forecasts, or the risks around them, in an increasingly meaningful way.

The NGFS note on the economic and financial impacts of extreme weather events, produced for the 2026 French G7 Presidency, uses 31 case studies to show that these effects repeatedly appear in output, employment and prices across a geographically diverse range of countries^{footnote [9]}. Supply-chain and trade linkages can also propagate these shocks beyond the country or region directly affected.

With that in mind, let me turn next to consider how central banks should diagnose the shock, assess its impact on their forecasts and the risks around them and therefore their monetary policy judgements.

A framework for assessing the impacts of climate shocks

Here I will draw on the recent NGFS guide on monetary policy strategy that I mentioned earlier^{footnote [10]}. The report provides a five-step approach to assess these impacts: identify, assess the propagation, assess whether a trade-off arises, set policy, communicate and monitor.

Step one is to identify the nature of the shock itself: How large is it? How long is it likely to last? Which regions and sectors are affected? Is it localised or widespread? And is it a one-off event or one in a sequence of shocks?

Those distinctions matter.

A short-lived disruption affecting a relatively narrow part of the economy during the low season of economic activity poses a very different challenge from a shock affecting food, energy, or critical infrastructure, or services such as tourism during peak season. And a sequence of individually temporary shocks may have quite different macroeconomic consequences from an isolated event.

Having identified the shock, step two is to assess how it propagates through the economy.

Here, the familiar distinction between direct, indirect and second-round effects is particularly useful.

Direct – or first round – effects are the immediate relative-price effects from the shock: for example, a drought raising food prices, a heatwave reducing labour productivity, or a flood disrupting production and logistics.

Monetary policy cannot repair infrastructure, restore a harvest, or offset the initial physical damage. If the shock is short-lived and expectations remain anchored, the standard monetary policy response is usually to look through the first-round effects, because policy operates with a lag and tightening in response to a temporary supply shock can impose unnecessary output costs.

But the initial shock can spread more widely, having indirect effects, as firms pass higher input costs across supply chains. And there can be second-round effects through wages and inflation expectations. Workers may seek compensation for higher living costs; businesses may adjust their price-setting behaviour; and households and firms may revise their inflation expectations.

Food and energy prices can be particularly important here. Households observe them frequently and a well-established literature shows that they have an outsized effect on expectations about future inflation. Research conducted at the Bank of England reinforces that result for the UK, finding that more than 60% of households report that food prices heavily influence inflation perceptions. Food prices also matter significantly more for household inflation-expectations dynamics than other components, with effects at short, medium and long horizons.

Frequency of shocks matters too. An isolated rise in prices will feel different to a series of repeated upward moves. Indeed, a sequence of individually temporary shocks which arrive before the effects of previous disturbances have dissipated can look and feel like more persistent inflationary pressure. Repeated disruption can also erode fiscal buffers and weaken private-sector resilience.

Central banks should therefore focus closely on frequency, persistence and pass-through to overall inflation and inflation expectations.

From diagnosis to the monetary-policy response

Step three is next: does the identified shock create a trade-off between inflation and output?

As I noted earlier, hot and dry events provide relatively clear examples. By reducing agricultural production, a drought can weaken output while putting upward pressure on food prices. A heatwave pushing down on productivity can increase inflation volatility at the same time as hitting output. But if shocks primarily damage capital and reduce output, such as floods and storms in some circumstances, they can push down on inflation where demand effects dominate.

If you permit me a brief digression at this point, the situation becomes no easier when considering the economic impact of the policies enacted by governments to ensure that the necessary climate transition happens. These policies can also drive large and persistent relative price movements, shifts in investment and reallocation of production across sectors of the economy. Modelling the impact of these shocks is therefore no less important. Here, the NGFS and IMF published joint work earlier this year on how countries’ Nationally Determined Contributions to reducing emissions under the Paris Agreement will affect inflation and output across a range of economies^{footnote [11]}.

Step four is to set policy and communicate the decision.

Here, the assessment depends importantly on the wider economic environment at the point the shock arrives. If inflation is already elevated and expectations are less firmly anchored, the risk that an initial price increase develops into broader inflationary pressure may be greater, and the case for responding becomes stronger. If the economy has spare capacity and demand is weak, the case for looking through is stronger.

This is also where a central bank faces an important judgement: if the effects of a physical shock are material and sufficiently well understood, they may belong in the central forecast.

But where there is substantial uncertainty around the scale of a shock or the way it will propagate, scenarios can allow policymakers to explore and communicate risks. Scenarios can complement the central forecast by helping us explore precisely the kinds of uncertainty I have described today: whether a shock proves temporary or persistent; whether supply or demand effects dominate; and whether initial price increases generate material second-round effects.

The fifth and final step is to monitor and assess whether the economy is evolving in line with expectations and adjust policy in response to new data developments. Since there is also evidence that climate shocks can affect credit supply, collateral values or financial intermediation, central banks should also monitor the operation of the monetary transmission mechanism.

You will notice, of course, that the process I describe here is not that different to how central banks assess and respond to other economic shocks. That is because the impact of climate change is now one of the set of economic shocks that monetary policymakers consider.

From immediate shocks to the resilience threshold

Since climate shocks are likely to increase in both magnitude and frequency, another important question is whether economies and financial systems can continue to absorb rising losses from these shocks over the longer term.

The capacity of households, businesses, governments and financial systems to absorb losses is a key determinant of the recovery profile from a climate shock. Fiscal space, insurance penetration, climate adaptation, infrastructure quality, financial-market depth and institutional capacity all help to determine a country’s resilience threshold. Where those buffers are stronger, shocks are more likely to be absorbed. Where they are weaker, the effects can become more persistent and can spill into credit, financial stability and public-finance channels.

Less developed economies typically have lower resilience thresholds, meaning similar physical hazards can generate larger and more persistent macroeconomic damage. A previous study by Raddatz (2009) estimated that a climate-related disaster resulted in a cumulative per capita output loss of 0.25%, 0.5% and 1% for high, middle and low-income countries respectively^{footnote [12]}.

But resilience is also not fixed. Repeated shocks can gradually erode the very buffers that help economies absorb losses in the first place, including fiscal space, private-sector balance sheets and insurance capacity. Without effective mitigation and adaptation, economies may therefore become less resilient over time, with the same physical hazard generating increasingly persistent effects on output, inflation, credit and public finances.

There is also evidence that these interactions can be non-linear. The NGFS’s 2023 work on compound risks examines precisely what can happen when shocks combine, either across time or across locations. It shows that the overall macroeconomic impact of compounding shocks can be materially greater than the sum of the parts^{footnote [13]}.

This in turn feeds back to the short-term policy response. Resilience directly affects the scale and persistence of the macroeconomic consequences of a shock, and therefore how difficult it is for monetary policy to look through it. But if losses are absorbed quickly through insurance mechanisms or other buffers, the inflation-output trade-off may be shorter-lived.

There is therefore an important interaction between resilience and monetary policy. The more resilient an economy is, the easier it is to absorb physical shocks. As resilience erodes, the same shock can become progressively harder to look through.

Insurance protection gaps and why they matter for central banks

The provision of insurance helps economies to cushion the effect of physical shocks by pooling and transferring losses. High insurance penetration and rapid payouts can reduce household distress, firm insolvency and credit losses as well as supporting a faster recovery in output.

But as physical risks increase, rising and repeated losses could place pressure on the availability and affordability of insurance in areas exposed to these shocks. If insurance coverage falls, losses will instead fall on households, businesses, banks and governments.

Two recent events provide a sobering comparison. In Spain, public-private insurance arrangements helped absorb a significant share of losses from the 2024 floods whereas in Pakistan, following the 2022 floods, only around 1% of total damages were insured, leaving households and small firms far more exposed to income and wealth losses, generating larger macroeconomic impacts.

From a microprudential perspective, withdrawal of insurance provision may enhance the safety and soundness of insurance firms, because rising insured damages can significantly dent profitability and potentially threaten firm resilience.

As the insurance supervisor, our role is not to direct underwriting decisions by individual firms. But as a macroprudential authority, we will need to take a broader view of the implications for the financial system as a whole. Indeed, the Bank of England’s Financial Policy Committee highlighted this issue at its September 2026 meeting^{footnote [14]}.

Where insurance coverage falls, uninsured risks will need to be redistributed:

• For households, this means higher repair costs when physical risks strike, lower property values and potential difficulties obtaining a mortgage.
• For businesses, it can mean bearing losses from the destruction of physical goods, capital and infrastructure.
• For banks, it can mean weaker collateral values and higher credit losses.
• For governments, it can mean more frequent pressure to act as insurer of last resort.

For the UK, Bank of England staff have modelled how a widening flood-insurance protection gap could spill over to the UK mortgage market^{footnote [15]}.

This analysis suggests that the share of UK mortgagors without insurance could rise from around 5% today to around 7-10% in 2050, or up to 16% following a severe flood event. The aggregate effects on house prices and mortgage impairments in that modelling are relatively small compared with the severe macroeconomic stresses against which the UK banking system is tested. But that does not mean the issue is trivial: the consequences for affected households, exposed lenders and local economies could still be significant.

There is an important lesson here: the answer cannot simply be more of the same insurance. Insurance capacity cannot sustainably compensate for underlying physical risks that continue to rise. Insurers will likely increasingly need to include risk prevention in product design, underwriting, and pricing practices.

Indeed, while a successful transition to net zero can limit the future increase in physical hazards, we will also need to consider the role of adaptation in a world where climate shocks are more frequent.

Investment in ensuring resilience to climate shocks can support insurability. This may require collaboration between businesses, households and governments. Updates to planning rules and building regulations could further mitigate the impact of increased physical hazards.

Actions to improve climate-related risk identification, measurement and modelling capabilities across the insurance sector, as envisaged under the recent update to the Bank of England’s supervisory expectations, should strengthen firms' ability to assess and price climate-related risks^{footnote [16]}.

But the financial sector also has a long tradition of coming up with innovative solutions to difficult problems. In this regard, there may be scope for greater collaboration between insurers, reinsurers and banks to share data, develop products and risk-transfer mechanisms that support resilience and help narrow protection gaps.

International cooperation

Let me finish with a word on international cooperation. Climate change is an issue which transcends borders and cannot be fixed by one country acting alone. In recent years, great strides have been made in multilateral fora like the NGFS to assess and model the impacts of climate change on the financial system. The NGFS’ flagship climate scenarios are an important tool in underpinning climate risk assessment in the financial sector. The work of the International Sustainability Standards Board to ensure consistent and better climate disclosures is helping financial firms to better manage climate-related risks and to make better decisions on where to allocate capital.

The 2025 South African G20 Presidency focused on the issue of natural catastrophe protection gaps, drawing on joint work from the World Bank and the International Association of Insurance Supervisors^{footnote [17]}. That work helped to raise our collective understanding of these issues as well as suggesting a range of approaches to tackle them, tailored to country circumstances.

In 2026, the French G7 Presidency took this work forward by strengthening the evidence on the macro-financial consequences of extreme weather, while examining how to tackle protection gaps^{footnote [18]}. This issue is also on our radar as the UK develops priorities for its own G20 Presidency in 2027. In my view, there is value in better understanding how insurance protection gaps affect the economic impact of extreme-weather events and in exploring what can help sustain insurability and resilience as physical risks increase.

That could include looking at what constrains insurability, which approaches appear capable of operating at scale, and the conditions needed for them to work sustainably.

Insurance cannot close protection gaps on its own. But better data, investment in adaptation and more resilient infrastructure can help sustain affordable cover. Timely mitigation also remains essential to limit the future increase in physical risk.

That brings me back to the two connected horizons. In the near term, central banks will need to understand how physical and transition-related shocks affect inflation, output and expectations in order to deliver on their price stability mandate.

Further ahead, the availability of insurance will determine both the economic impact of climate shocks and the capacity of economies to absorb the resulting losses. It will also shape the prevailing environment in which central banks will pursue their price and financial stability goals.

As macroprudential policymakers, we understand that the availability of insurance matters to the resilience of the financial system as a whole. So as protection gaps widen, addressing them will require concerted action across governments, regulators, international institutions and the private sector. Multilateral fora can play an important role in supporting that dialogue and sharing lessons on what works.

As climate change increasingly affects our economies and financial systems, central banks will need to understand these effects, just like any other shock. This will enable us to make sure that the financial system remains stable and that price stability is maintained. And in turn that can provide a sustainable platform for governments to deliver the climate transition.

I would like to thank Danae Kyriakopoulou for assistance in writing this speech. I would also like to thank Lukasz Krebel and Fergus Booker for their input and helpful comments, as well as India Rimmer, Hannah Copeland, Boromeus Wanengkirtyo, Dario Bonciani, Riccardo Masolo, Silvia Sarpietro, Nikoleta Anesti, Vania Esady, Matthew Naylor, Will Banks, Kemal Erçevik, Alex Curtis and Jemima Hall for their underlying research.