A federal effort to test whether major banks can model their own climate risk has finally yielded an answer: They can’t — at least not reliably.
The problem was revealed when the Federal Reserve published the results May 9 of its first test of how well six of the world’s largest banks could measure their exposure to extreme weather and the clean energy transition.
“The banks are, in some ways, at an early stage of trying to understand and measure and manage those risks,” Fed Vice Chair Michael Barr said Wednesday during a House Financial Services Committee hearing.
Banks struggled to assess how they would be affected by hypothetical “climate scenarios” because climate-risk modeling is in its infancy and important data is unavailable, banks told the Fed. Climate risks are “highly uncertain and challenging to measure,” the Fed report says.
Experts say the results serve as the latest signal that financial firms and regulators must improve climate data and modeling so they can more accurately measure and address the financial threats of climate change. The six banks, whose total combined assets amount to more than $14 trillion, told the Fed they plan to take those steps.
The banks are JPMorgan Chase, Bank of America, Wells Fargo, Goldman Sachs, Morgan Stanley and Citigroup.
“The implication is lack of data, lack of expertise, the lack of analytical capability to do a thorough analysis,” said Richard Berner, a financial risk expert who formerly directed the Treasury Department’s Office of Financial Research.
The exercise showed “they need to develop their capabilities,” Berner added. “That’s a good thing, because awareness is the first step.”
The finding is significant. But it isn’t surprising.
Regulators overseas have encountered the same obstacles as they’ve required the banks they oversee to conduct similar analyses. The Bank of England said in 2022 that the climate-related exercises are still in their infancy “and there are several notable data gaps” — creating uncertainty around estimated losses.
Improving data, experts say, would help banks gauge and manage their financial vulnerability as the economy shifts from fossil-fuel-based energy and as intensifying natural disasters disrupt insurance and mortgage markets.
“What we don’t want to happen is for a fragile bank to go down and cause a systemic crisis, like what happened in 2008,” said Scott Kelly, senior vice president at climate risk analytics platform Risilience, who has worked with central banks including the Bank of England on climate scenarios.
Similar to a stress test
The Fed results come as financial regulators around the world try to gauge how climate-related risks could ripple through the global financial system.
Central banks in the United Kingdom, the European Union and Australia, have launched “climate-related scenario analyses,” which resemble traditional finance stress tests. The climate tests are newer, meant to be experimental and have no regulatory consequences for banks.
U.S. regulators are taking their own steps prompted in part by an executive order President Joe Biden issued in 2021. The order led a council of top financial regulators to recommend that all of its members, the Fed among them, undertake climate-related scenario analyses.
The Fed launched the exercise in 2023. The central bank asked the six banks to measure how the energy transition could affect their current commercial real estate and corporate loan portfolios under two scenarios. In one, governments have failed to adopt additional climate policies. In the other, the world has successfully limited global warming.
The Fed also had banks model how their residential and commercial real estate portfolios would fare amid hurricanes of varying strengths in the Northeast, as well as after a separate disaster chosen by each bank.
It’s highly unlikely any of the climate scenarios will actually occur, said Kelly of Risilience. But that’s not important. The Fed’s goal was to test banks against hypothetical climate shocks to ensure they are prepared for those that do occur, Kelly said.
‘An imperfect tool’ for climate risk
Banks ran into two problems — a lack of detailed information about their properties such as insurance coverage and the immature state of climate-risk modeling.
The Fed said “most participants” do not have information about how buildings they finance are constructed, such as whether a roof is made from asphalt or wood or if its windows can withstand powerful winds. The banks relied instead on assumptions around property characteristics and academic studies of historic events.
Another problem: All six banks said they do not know insurance coverage details including deductibles and the cost of replacing a destroyed property. That also forced them to rely on assumptions around what costs insurers would cover and what the banks would pay.
Modeling was another problem. The banks said they relied on climate modeling firms because they do not currently have sufficient in-house systems or expertise.
Climate-transition models are vastly oversimplified, and modeling physical risk involves using estimates and assumptions to fill in data gaps, said Clifford Rossi, a former chief risk officer for Citigroup’s consumer lending group.
“In the end you’re not sure what the reliability of that estimate really is,” said Rossi, who is now a professor of the practice at the University of Maryland’s business school. As a result, he added, “the models are nowhere near ready for prime time in making hard money decisions.”
Inconsistent and unreliable climate data is not isolated to the banking system.
It’s become such a pervasive problem among investors that the Securities and Exchange Commission in March finalized a controversial rule that would require public companies to provide investors and the public with information about their greenhouse gas emissions, climate goals and exposure to extreme weather.
“Scenario analysis is an imperfect tool to measure climate risks,” said Mekedas Belayneh, a policy advocate with Public Citizen’s climate program.”Modeling climate risk is very complex and requires data that isn’t out there right now.”
In addition to improving that modeling, Belayneh said, banks and regulators should consider “ways of thinking about risk that go beyond traditional financial modeling.”
The banks are well aware of the problems and plan to incorporate climate risk analysis into their risk management processes over time, as well as to invest in data, models and expertise to improve them, according to the Fed.
For now, a “high degree of uncertainty” inherent to climate modeling will affect “how the results of climate scenario analysis exercises could be used going forward,” the Fed report said.
Goldman Sachs and Morgan Stanley did not respond to a request for comment. Wells Fargo and JPMorgan declined to comment for this story.
Citigroup declined to comment beyond what the firm disclosed in its 2023 climate report, which noted that the firm has carried out its own analyses and is working to improve its internal data and modeling. As to the Fed’s exercise, Citigroup said the effort provided “useful insights” on potential climate vulnerabilities, which it plans to use to inform strategy around climate risk.
