Jonathan Harris and Barry O’Connell look at the link between climate risk and asset returns
The original post can be found here.
There is a strong relationship between climate risk and equity return
In recent years, climate risk has become an important topic for the actuarial profession. Regulators have examined the insurance industry's exposure to climate risk and the IFoA has established the Resource and Environment Board.
Arguably, discussion of the implications of a changing climate have focused on the liability side of insurance and on long-term implications. What can we say about the current implications for asset returns?
In this article, we highlight how the Actuaries Climate Index (ACI) has a strong relationship with returns on the equity value factor. This is an example of how the market is already pricing in climate risk. Thus, climate risk does not just have long-term implications; it is already having a significant impact on investment returns.
The Actuaries Climate Index
The ACI is defined as "an objective measure of changes in extreme weather and changes in sea level", and is a standardised average of six-monthly measures across the US and Canada that relate to the following climate features: temperature extremes (high/low), rainfall, drought, wind speed and sea level.
Figure 1 shows the ACI (grey) and its 12-month rolling average (red) over the whole sample period.
Connection between the ACI and equity factors
We focus our investigation of the links between the ACI and equities on the three most prominent factors for explaining patterns in equity returns: the market, value and size factors of Fama and French. The factors are for the US stock market, which is a natural fit to the ACI's coverage of the US and Canada. The market factor represents the return on a value-weighted portfolio covering the whole market. The value factor is defined as the return on a portfolio of value (eg low price-to-earnings) stocks minus the return on a portfolio of growth (e.g. high price-to-earnings) stocks. And the size factor is defined as the return on a portfolio of small stocks minus the return on a portfolio of big stocks.
To explore the link between the ACI and equity factors we use a technique called economic tracking (ET) regression. The principle behind this technique is that investor expectations of future cashflows and risks are contained in current stock prices. Changes in expectations of future climate risk levels may affect investors’ expectations of future returns, thereby changing current asset prices and returns. Thus, equity returns may contain information that can be used to improve forecasts of future climate risk. So an ET regression for the ACI looks at whether equity returns can predict the ACI 12 months ahead.
Following standard ET regression methods, we use a 12-month rolling average of ACI to smooth out short-term deviations and seasonality, and we look to forecast the change in ACI over the following 12 months. We include a number of variables that have been shown to forecast equity returns: the log price/earnings ratio, the default spread, the term spread and the risk-free rate (Campbell, Polk, and Vuolteenaho (2010), Balvers, Du, and Zhao (2017)). To take into account the pronounced trend in ACI over the sample period, we also include a trend term.
Table 1 reports the results of this regression both for the whole sample (January 1964 to August 2016), and over the two halves of the sample. This enables us to see if the results of the regression hold both in the past and in recent times.
The coefficient on the value factor is negative and significant at the 95% confidence level or higher for each period. The value factor coefficient is more negative in the more recent period, indicating an increasing sensitivity of the value factor to climate risk. Including the financial factors increases the R-squared from 7% to 9.1% for the whole sample. This shows the improvement that the financial factors bring to forecasting climate risk.
Figure 2 shows the actual versus forecast change in ACI over time. We can clearly see that including the equity and financial factors in our model increases the ability to forecast changes in the ACI.
The finding that the value factor is strongly linked to climate risk is relevant to investors that may seek to hedge their exposure to climate risk. It is also relevant to those who may seek to take on exposure to climate risk based on the belief that this risk will be rewarded in the long-term. It is also worth considering the potential for 'wrong-way risk' from an overall asset-liability management perspective: that is, if a firm's liabilities increase and the returns on their investment portfolio decrease at the same time as climate risk increases.
A scientific approach to integrating global issues into asset management
We have demonstrated how there is a strong link between the equity value factor and the ACI, thereby showing the relationship between climate risk and asset returns. Actuaries that are involved in the management of assets and investment risk need to be aware of these links, as highlighted by the paper ’Resource and Environment Issues: A Practical Guide for Pensions Actuaries’. Recent research has highlighted similar links between equity returns and other non-traditional variables, including those related to social inequality (eg Marfè (2017), Greenwald, Lettau, and Ludvigson (2014)). There has also recently been a push for the investment industry to become more aware of and aligned with the UN's Sustainable Development Goals and other environmental, social and governance (ESG) metrics.
We believe that combining statistical insights about the links between risk metrics and asset returns with the wealth of intuitive knowledge held by sustainability practitioners can generate precise insights into what data actuaries should be aware of, seek out and use.
Actuaries are skilled and experienced in financial risk modelling. Climate risk is highly uncertain in quantification, and to meet this challenge we must become ever more creative and broader in scope when applying our skills. The use of the ACI and ET regression in this article is just one example of new data and techniques that can be used. Perhaps this points the way towards how actuaries can incorporate long-horizon global trends and risks into their current investment models.
Jonathan Harris is co-founder and Head of Research of ET Index Research.
Barry O’Connell is actuarial advisor to ET Index Research and Associate Director of ILS Analytics at Twelve Capital