Economic modelling of climate risk

On July 24, UK insurance group Aviva declared that it would pursue a tougher line on coal-intensive companies it invests in, potentially selling out of such holdings. It joins other insurers taking action on carbon risk, including the French group AXA, which in May said it would sell €0.5b coal investments this year, measure the carbon-footprint of its portfolio, and triple its clean-energy investments to €3b by 2020. 

Aviva executives pointed out that coal companies are at great risk of “existential threat” because, barring an extraordinary surge in carbon capture and storage, their businesses have little place in a world that tries to avoid dangerous climate change. The occasion for comments on climate change and investment was the launch of a paper on “Climate Value at Risk”[i] by the Economist Intelligence Unit (EIU), commissioned by Aviva. The EIU paper represents the latest addition to a small but rapidly-growing body of work that seeks to identify and measure the implications of climate change for the financial sector.

The paper uses the traditional approach to modelling broad economic impacts of climate change: the Integrated Assessment Model (IAM), and specifically,  William Nordhaus’s DICE model[ii]. Similar IAM-based approaches were employed in the reviews led by Lord Stern[iii] and Ross Garnaut[iv]. Those reports focused on the change to GDP from both the effects of climate change itself and policy action to reduce the impact of climate change. The EIU paper instead seeks to model the impact of climate change on total “manageable” assets, to determine the value at risk from climate change by 2100.

For asset managers today, the EIU estimates the value at risk from climate change (VaR) as $4.2tn by 2100 — equivalent to the entire GDP of Japan. Under extreme scenarios of 5C and 6C temperature rises from climate change, the losses are $7.2tn and $13.8tnb respectively. However the paper acknowledges that this figure is likely to understate the risk. The EIU uses a discount rate of initially 5.5% p.a. falling to 4% p.a. by the end of the century, due to slowing growth. This represents the average private sector discount rate for asset managers. However, for government purposes, the EIU paper uses much lower discount rates of 3.8%p.a. initially, falling to 2% p.a. That leads to a VaR of $13.9tn under an average scenario, rising to $43tn under a 6C scenario — equivalent to 30% of the value of all current assets. 

Anyone with a passing interest in climate science will probably know that if we reach 6C of warming, damage to financial assets might well be the least of our problems (or our children’s and grandchildren’s problems). As the report notes, there is a lot of uncertainty in the likely impact of global warming if temperature rises exceed 3C. It’s not unreasonable to wonder whether, at 6C, human civilisation as we know it would continue at anything like its current form. How, then, can anyone predict with any confidence what this might mean for the economy or for financial assets?

This is a problem identified by several economists such as Harvard’s Martin Weitzman, who identifies five broad problems with Integrated Assessment Models and the benefit-cost analysis (BCA) approach upon which they rely[v]:

The EIU paper acknowledges throughout that it is essentially modelling the unmodellable. IAMs, it says, “can be and are fairly criticised” for a range of shortcomings, and “are also generally considered to be conservative in their estimates of damage”. The authors say that while they’ve taken great care to acknowledge and incorporate scientific uncertainty, “we acknowledge that the modelling faces limits and that the results should be interpreted as a guide to the likely magnitude of impacts”. 

So why do it this way at all? The alternatives, to put it bluntly, are limited: the EIU paper’s authors note that there are no other modelling tools yet developed that “quantify the economic cost of climate change within a consistent framework, as is required for this project”.

Some economists have argued that there’s a better way than cost-benefit analysis. MIT economics and finance professor Robert Pindyck in 2013 noted[vi]:

“Probably the best we can do at this point is come up with plausible estimates for probabilities and possible impacts of catastrophic outcomes…[and] think of a GHG abatement policy as a form of insurance: society would be paying for a guarantee that a low-probability catastrophe will not occur (or is less likely).”

Pindyck was clearly thinking of policymakers; whereas the EIU paper was commissioned by Aviva, and appears to have asset owners and managers in mind for its audience, rather than just policymakers.

It’s quite reasonable, then, that some of the most striking parts of the paper have little to do with the modelling exercises. Perhaps the most interesting are the author’s claim that it is virtually impossible for large asset managers/owners to manage their way out of climate risk:

“The impacts of climate change, at least for modest degrees of warming, can be expected to concentrate in sectors of the economy sensitive to weather conditions, for instance agriculture, energy, forestry and water. However, these sectors are connected with the rest of the economy through supply and demand linkages, and shifts in the prices of goods and services because of climate change will affect overall spending patterns and household incomes.

Allied to the fact that at higher degrees of warming the impacts of climate change are expected to become increasingly economy-wide, this means that climate change poses a systemic risk, coming through weaker growth and lower asset returns affecting the entire portfolio. As a result, asset managers may struggle to avoid climate risks by moving out of vulnerable asset classes and regions. This is because, at least under lower-probability and higher-impact outcomes, our findings suggest that climate change will primarily have a macroeconomic impact that affects the entire portfolio of assets. The interconnected nature of the problem is likely to reduce returns, even on those investments not actually harmed by physical damage. Given this result, asset managers will face significant challenges diversifying out of assets affected by climate change.”

If it is indeed the case that climate change represents a systemic risk — and in an area rich with uncertainty, it makes intuitive sense — what are individual financial institutions to do, whether they are investment managers, insurance firms, banks or any other agent with long-term obligations? The Climate Institute’s recent paper[vii] on financial systems risk recommended financial regulators consider how climate change might affect financial stability. The EIU paper suggests that won’t be achievable without broader climate policy to reduce these systemic risks.  

CPD: Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.