Research showing that climate models grossly underestimate the macroeconomic risk posed by climate change was addressed by a panel of experts on Wednesday.
Hosted by Green Central Banking and the Climate Safe Lending Network, and moderated by Boston University professor Madison Condon, the event brought together global specialists to discuss the disconnect between climate finance models and climate science.
The panel – which included Ann Pettifor, political economist and author; Steve Keen, distinguished research fellow at University College London; and Sahil Shah, co-founder of risk consultancy Tipping Frontier – considered the paradigm shift needed in economic modelling to effectively assess and manage climate risks to the financial system.
Sanjay Joshi, responsible investment consultant for Hymans Robertson and co-author of The Emperor’s New Climate Scenarios, opened the panel by stressing that a key function of climate scenarios is to capture tail risks. Yet he suggested they are often failing to do so in practice.
Keen then elaborated on his research critiquing the “neo-classical” economic approaches to climate risk that commonly underpin the scenarios used by central banks. These, he said, “need to be thrown out” as they are “completely devoid of realism”.
This criticism was echoed by other panellists. According to Shah, models need to be recalibrated to translate top-down global approaches into bottom-up catastrophe models which run particular tail risk scenarios looking at direct and indirect impact pathways.
Keen’s investigation of 39 neo-classical economics research papers on the cost of carbon found that none would pass peer review by climate scientists. Rather than including global circulation models, they relied on faulty internal estimates and failed to include climatic changes aside from temperature rises.
Pettifor expressed concern that the underestimation of climate risk has resulted in a lack of political will from central banks to address the “grave”, “very present” and “looming crisis”.
She added that central banks have eschewed agency in scenario development and are overly reliant on the private sector. Yet it is central banks, she argued, that are best equipped to do so and ultimately responsible for managing “the financial system and in particular the monetary system in the interest of the public sector”.
She pointed to the response to Covid-19 to demonstrate that, in the face of crises, central banks can support government action without compromising their independence. Until central banks adopt a mindset commensurate to the scale of the crisis described by climate scientists, Pettifor said it is unlikely sufficient action will be taken to contain systemic risks.
Scenarios should build local resilience instead of hedging risk
According to Joshi, climate science has shown that the “likelihood of mass financial collapse is so strong” that “there is no way to hedge this”. There is “simply a fiduciary duty to reduce emissions”, as “the systemic harms that come about from the externalities from one part of the portfolio” affect banks’ entire balance sheets.
The panellists agreed that a paradigm shift is needed to move away from designing scenarios with the goal of hedging risk towards research into building localised resilience, as well as appropriately targeted adaptation and mitigation finance.
The panel also considered various ways to make climate modelling fit for purpose.
As quantitative modelling relies on backward-looking data that is ill-equipped to fully capture climate risks and tipping points, Joshi explained, his research proposes complementing them with “rich narrative scenarios”.
Such qualitative scenarios begin with a geographically located climate shock, and consider how the “real world” climate event can impact interconnected risk drivers, cascade through the financial system, and what actions could be taken.
Another promising option discussed was catastrophe modelling, which Shah defined as the distribution of a particular natural hazard’s likelihood and intensity. Catastrophe models have a three-dimensional view of climate events, where risk is a function of not only the hazard but also geographical exposures and vulnerabilities.
Shah explained that widening access to vulnerability data is necessary to improve catastrophe modelling and enable a steady flow of adaptation finance to developing economies.
However, there are inherent trade-offs to consider, said Shah, as making vulnerability data on critical infrastructure open source may pose national security issues and affect asset values. He proposed the use of satellite image proxies and data trusts as possible workarounds.
Finally, the panellists agreed that interdisciplinary research teams, including social and climate scientists as well as other relevant experts, are needed to take a holistic view of macroeconomic impacts. Shah recommended starting with a broad qualitative approach to identify “the right subject matter experts to fill in the various different blanks” for specific scenarios.
This page was last updated November 27, 2023
Share this article