This article is part of our series on current academic research into a range of sustainable investment topics. The papers discussed were presented at the latest annual GRASFI conference.
Compelling academic papers
The Global Research Alliance for Sustainable Finance and Investment is a collaboration of universities committed to producing high-quality interdisciplinary research and teaching curricula on sustainable finance and investment. In this series, we highlight compelling papers presented at the latest GRASFI conference, with a comment from a ‘practitioner’ at BNP Paribas Asset Management (see below).
As the sustainable investor for a changing world, BNP Paribas Asset Management (BNPP AM) sponsors GRASFI’s efforts to bring academic rigour to the challenges of sustainable finance and investment. Through its sponsorship, BNPP AM is able to access leading academic research into sustainable finance and investment, helping to inform the broader debate. Our goal is to share these reflections with clients and the industry. Visit the GRASFI Conference website.
Investors worldwide are trying to understand the impact of climate change on their portfolios. The challenge is particularly acute for investors with extended investment horizons such as pension funds.
Three authors from Rotterdam School of Management, Erasmus University, have developed a novel framework to allow this type of investor to assess the impact of physical climate risks on long-horizon equity risk returns and make optimal portfolio choices.
In their paper Climate Change and Long-Horizon Portfolio Choice: Combining Insights from Theory and Empirics, they deploy Bayesian mathematics and the temperature long-run risk (LRR-T) model proposed by Bansal, Kiku, and Ochoa in 2019.
Their research includes three actionable insights:
The authors note that regulators are increasingly demanding that financial institutions quantify their exposure to climate risks. Moves include the IORP II directive from the European Union, which requires that workplace pension schemes across the bloc — which manage more than EUR 2.5 trillion — take long-term issues such as climate change into account and have regular risk assessments.
Yet they comment: “Investors struggle to fully grasp the potential impact of climate change on the value of their portfolio and are searching for approaches to quantify these risks.”
They say much of the research centres on the impact on current asset prices, while fewer look at more extended holding periods: “These long-term effects are difficult to measure due to three key problems: parameter uncertainty, data unavailability, and peso problems.”
The ‘peso problem’, first popularised by economist Milton Friedman, is the difficulty in assessing the value of assets that could be impacted by an unprecedented or infrequent, but high-impactevent such as a rapid devaluation of the Mexican peso.
As an illustration, they quote US economist Robert Stambaugh who said when we “expand the horizon to the next several decades, the possible effects of global warming range from negligible to catastrophic.”
To provide insight amid such uncertainty, the authors turn to Bayesian mathematics, a flexible and powerful branch of statistics that allows probabilities to be adjusted as new information emerges.
Historical data about the impact of climate change is limited, which is exactly why using Bayesian statistical tools can work well. They make it possible to include more information than just historical market data. Bayesian methods can also account for uncertainty around the estimated impact of climate change on financial markets.
Furthermore, information can be extracted from the LRR-T model by seeing how investor concerns over climate change impact current asset prices.
The authors say their work tackles a long-standing issue when assessing the impact of climate change on future returns: “We provide a concrete approach to deal with the widely shared concern that using historical data to study portfolio allocation has significant limitations in light of the uncertain long-term effects of climate change not present in the data.”
Existing historical data, they add, may not be informative because the frequency and economic impact of climate-induced disasters is likely to change.
However, investors do not have to accept the model as developed by the three authors. Instead, their model allows investors to decide for themselves the weighting of historical data versus that from models —and to see how that changes predicted returns.
“All investors are intimately aware of the prospectus disclaimer: ‘past performance is not a guarantee of future results.’ Never before has this statement been truer than today given the host of unprecedented secular changes acting on markets, including climate change. Embedding more forward-looking assessments into quantitative risk analysis is likely to be a central project for the next phase in the evolution of our industry. The authors of this report make an important contribution to this emerging discipline,” said Alex Bernhardt, Global Head of Sustainability Research at BNP Paribas Asset Management.
 Mathijs Cosemans, Xander Hut and Mathijs van Dijk; also see www.rsm.nl/