Climate risk test asks banks to look too far down the road

Bankers have shown this year they could handle a broadly destructive health pandemic that blindsided the world. But they’re also preparing for another headwind: climate change.

As key financial intermediaries, banks have an important role to play in managing a transition away from carbon, but there is one idea that does not appear ready for prime time: stress testing for climate change.

Fortunately, U.S. banks are fully engaged on assessing and disclosing climate risks. And perhaps more importantly seeking to develop markets to assist in a transition away from carbon-intensive business.

For example, the Partnership for Carbon Accounting Financials recently announced a global standard to allow banks, asset managers and investors to use a common method of reporting greenhouse gas emissions tied to lending and investment portfolios.

Financial innovation is also underway as 2019 had a record issuance of green bonds exceeding $250 billion, and development of green asset-backed-securities, hedging products and reforms to carbon-offset markets. In terms of risk management, banks are actively developing their climate scenario analysis capabilities to understand the risks they face across their businesses.

But some regulators have suggested going further in an attempt to quantify those risks for capital purposes using government-run climate stress testing. Such testing comes with assumptions about how the climate will evolve, how governmental policy will change and how both will affect bank borrowers. Those are very difficult projections to make, and can become highly speculative over the longer term.

For perspective, consider in 2008 when the United States was a major importer of oil and natural gas, it was universally projected to face higher prices for the former and a shortage of the latter. Ten years later, the U.S. was the world’s largest producer of both oil and natural gas, the largest natural gas exporter in the world and among the three largest oil exporters.

Remarkably, United Kingdom and European climate stress tests envision a 30-year projection with embedded assumptions about how global energy markets will change over that period. For perspective (and humility), consider that the Federal Reserve’s financial stress tests have a time horizon of two to three years, even in an area with much richer history and data to support projections.

The even greater challenge, however, is to predict how banks will change their businesses over that same timeline. The average weighted maturity of a commercial-and-industrial loan is three years, so a bank’s portfolio would turn over 10 times in a 30-year stress horizon. Of course, to the extent that climate change dims the prospects of a given industry or company, the bank would take one of those ten opportunities to reduce its exposure and lend to a firm whose prospects have brightened.

Here’s an example. Assume a U.S. bank has a revolving line of credit to Ford Motor Co. Most would predict that the climate change risk of that loan is relatively low, as the bank can decide to terminate it if Ford’s gas-powered vehicles lose market share over the years. There is also the distinct possibility that Ford will increase its production of electric vehicles. And of course, the bank could also hedge the risk by lending to Tesla as well, or buying credit insurance on Ford.

So, what is a bank to assume for the next 30 years? The U.K. test would require a bank to model cash flows and collateral values over the next 30 years, reflecting “judgments about how companies would be positioned in light of both their underlying risks and opportunities, including an assessment of their current mitigation and adaptation plans.”

But what could that possibly mean? And how relevant is such an analysis when a bank has the option simply to exit the credit at multiple intervals along the way? And that is just the supply of credit.

Presumably, carbon-intensive businesses that begin to perform poorly will shrink, reducing their demand for credit, and thereby banks’ exposure to any subsequent default. Climate stress tests do not appear to take this factor — basically, half the equation — into account.

Certainly, it is important for bankers and their regulators to measure and manage climate-related financial risk in a way that provides an accurate picture of what is at stake. But trying to capture climate change effects decades in advance — without considering the extraordinary adaptability of the financial system and economy — and incorporating those results into the regulatory capital framework is no easier than predicting how pandemics or machine learning will affect banks by 2050.

Regulators should instead ensure that climate risks are well understood and appropriately disclosed. They should encourage innovative efforts by financial companies to develop markets that can smoothly speed up a transition to a greener economy.

Beyond finance, policymakers managing a transition to a green economy have far more effective tools available to them, such as direct regulation or through market mechanisms. Stress testing for banks, however, seems a highly inefficient vehicle and one that risks degrading the integrity of financial regulation.