A new paper argues FinReg is slowing the climate transition. I’m not convinced

This article has been made free-to-view. Like what you read? Then consider buying us a coffee here👇

Tony and I don’t see eye-to-eye on everything when it comes to climate and the financial system.  Our recent back-and-forth on whether or not banks should act as the “climate police” could tell you that much.

One thing we do agree on, though, is that banks should be given every incentive to support the climate transition.  After all, as Tony has said, what’s a few bank failures in exchange for climate salvation?

Financial regulation (FinReg) has a powerful effect on banking activity.  It lays out the rules of the road for firms and can – for better or worse – shape lending and trading behaviors in ways that have profound implications for the financial system and real economy alike.  In an ideal world, FinReg would incentivize banks to support clean technology and renewable energy, while discouraging fossil fuel investments that don’t align with a 1.5°C warming pathway.

Is this happening in our world?  A new paper in Nature Climate Change dives into the nexus of financial accounting rules and loan portfolio management to find out if FinReg is harming, rather than helping, the climate transition.  

Spoiler alert: the authors think it’s the latter.

However, the way they’ve gone about trying to prove this is fanciful, to say the least.  The paper ends up being a fun exercise in data analysis with little relevance to the real world.

Let’s dig in, and I’ll explain.

The paper hinges on how banks calculate loan loss reserves (LLR), the sum total of loan income they have to set aside as ballast against expected credit losses (ECL).  Sizing LLRs is a finickity task, but generally involves running a bunch of historical loss data for a given portfolio into a risk model to project possible future impairments.  Under IFRS 9, the accounting rules used by European banks, ECLs are also shaped by firms’ forward-looking estimates of loan performance under various macroeconomic scenarios.

The authors’ contention is that the backward-looking component of the rules produces lower LLRs for high-carbon portfolios and higher LLRs for low-carbon ones, making “green” assets a drag on profitability.

This is because historically “carbon hogs” like oil and gas majors have been pretty safe bets for banks.  These companies have been around for decades, are consistently profitable, and have relatively low debt burdens.  In contrast, certain “green” companies have proven more fragile – some succeeding, others collapsing in ignominy. 

The authors argue that this historical bias lowballs the risk that an accelerating low-carbon transition poses to high-carbon companies.  For sure, in the past they’ve been safe bets, but in the near future, given the policy, technology, and market trajectory, they may turn into losers. 

Tony’s challenged this attitude before, and while I’m more open to the idea that the climate transition will make a mockery of past credit data, I am less sure how banks and regulators should respond.

What’s new in this paper, however, is the argument that this structural difference in LLRs acts as a drag on the climate transition, by discouraging banks to dump the “carbon hogs” on their balance sheets in favor of clean, green alternatives:

“Specifically, as indicated by the accounting rules, we assume that if a bank had to divest from high-carbon sectors and re-invest the proceeds in low-carbon sectors, the PCR [provision coverage ratio, calculated as the ratio of LLRs over outstanding loans] of such investments would need to increase to the higher level of the latter … This would in turn lead to a higher level of loan loss provisions and higher costs due to the structure of the accounting rules.”

They use data gathered from the European Banking Authority to size this differential, concluding that banks have an average PCR of 3.4% for low-carbon portfolios, compared to 1.8% for high-carbon loans.  The authors argue the PCR increase that would have to occur if banks suddenly transitioned their portfolios could cost potentially billions of euros:

“We estimate that for some banks, the transition could cost as much as 5 years of profits over the divestment horizon and, on an outstanding loan weighted average basis, 15% of the previous 5 years of profits due to a large increase in LLR.”

Understandably, a bank is not going to want to conduct a green overhaul of its lending book if it takes such a huge bite out of its profits!

But I have my quibbles with these estimates, and the authors’ conclusions, based as they are on a frankly absurd “divestment simulation” and a selective view of how LLRs fluctuate over time.

My main beef is with their choice to use an overnight, all-at-once transition on which to base their findings, one that is simply not credible.  The authors themselves admit “the divestment would probably be spread across multiple years” but justify containing their simulation to a single fiscal year on the grounds that “frontloading the entire impact allows us to better investigate the implicit incentive structure created by the regulation.”

No dice, fellas.  The loan market would freeze up if every bank in Europe tried to offload their dirty assets all at once.  There wouldn’t be enough buyers to absorb the portfolios, even within the burgeoning private credit industry, and the percentage that would be sold would go for distressed prices, acting as a brake on further sales. 

It’s also important to consider how regulators, who are first and foremost concerned about financial stability, would respond in this overnight transition scenario.  In the teeth of the COVID-19 crisis, European regulators took steps to prevent huge spikes in LLRs from destabilizing the banking system.  It’s not hard to imagine them extending similar relief measures to banks suddenly swamped with risky green loans.  

Indeed, among the COVID-19 measures was a European Central Bank recommendation that lenders’ “avoid excessively procyclical assumptions in their IFRS 9 models to determine their provisions.”  In other words, don’t assume the current chaos will translate into wildly higher credit losses down the line, and keep a leash on your LLRs.  This proved to be wise council, given the huge amount of government support for the real economy and financial sector kept many businesses from the abyss.

Furthermore, on a purely conceptual level the idea that an overnight divestment would lead to a credit risk increase holds no water.  Let’s step out of the data and into the imaginary world in which this system-wide lending shift is taking place.  Such a rapid transition would only make sense if political, economic, and financial actors were putting every effort into achieving net zero as fast as possible.  Given this, surely all manner of policy tools, state subsidies, and risk-sharing collaborations would exist to support the divestment of high-carbon assets and pile-in to green companies.

In such a world, backward-looking approaches to calculating LLRs may be discarded altogether.  But even if they persisted, and caused a jump in reserves, the pro-transition tailwind would ensure this would be a temporary phenomenon. Each year, banks would be able to lower their PCRs as data comes in showing green loans are safe and stable investments thanks to the supportive policy and market environment.  Remember, LLRs aren’t static. If credit downgrades and losses don’t materialize, banks can lower their LLRs and return loss provisions into income, plumping their profits. Again, this happened in the years after COVID-19, leading some banks to post record earnings.

A more realistic, multiyear divestment simulation would take this PCR evolution into account. Each new year of ECL forecasting would produce lower LLRs for green assets, and higher ones for fossil ones, making the mechanism of financial accounting supportive of the transition.

Look, I love a good piece of data analysis.  But when it comes to something as serious as claiming that FinReg is slowing the transition to net zero, you can’t defend your assertion with spreadsheets alone.  You have to consider the world as it is, and put together a convincing version of the hypothetical world in which your given scenario – in this case a rapid divestment – takes place. 

Otherwise you’re just having fun with numbers.