Can big profits be a red flag in banking? Yes, a new study finds.

Sophie Park/Bloomberg

Bankers have long measured their success through their return on shareholders' equity, but a new academic paper finds that the metric has a troubling downside: U.S. banks with higher profits are more likely to suffer stock-price declines during crises.

The result held up during this year's banking crisis, which was sparked by the crash of the once-high flying Silicon Valley Bank, where a high ROE had masked massive risks. High-ROE banks have also suffered more in past crises, the paper found, adding to the skepticism among academics about the much-used measure.

"I know bankers love to focus on ROE, but I'm not sure that should be the goal," said Amiyatosh Purnanandam, a University of Michigan finance professor who co-authored the study with economists Ben S. Meiselman and Stefan Nagel.

Return on equity measures a bank's profitability over a certain time period. The constricted time horizon prioritizes short-term performance over durable gains to shareholders' investments, Purnanandam argues. Though high-ROE banks see their stock prices rise during good times, their equity values suffer more when a crisis hits, the paper found. 

That trend held up for U.S. banks during both the 2008 financial crisis and the savings and loan crisis, which wiped out hundreds of lenders in the 1980s and 1990s. The paper's authors argue that regulators should look more closely at ROE since it's a strong predictor of whether an individual bank may come under stress. 

Regulators have spent decades trying to prevent the issues that caused the previous crisis, the paper argued, calling the increasingly complex models they have developed a "perpetual game of cat-and-mouse." 

Taking the simpler approach of looking at profitability can offer big clues about the institutions that are taking larger risks, the paper's authors argue.

"While the underlying risk that triggered the crisis differs from one crisis to another — e.g., interest rate risk, mortgage losses or exposure to sovereign debt — the link between profitability and systematic tail risk is always strong," Purnanandam and his co-authors wrote.

They added that the "predictive power of ROE is pervasive."

The paper is the latest analysis to examine the shortcomings of ROE. In 2017, a pair of French finance professors examined 273 large banks in 28 countries and found that those with higher pre-crisis ROEs performed worse during crises.

"While ROE is used as a chief performance measure in banks, our results indicate that in reality ROE constitutes a good proxy for the risk exposures of a bank and its vulnerability to crises," wrote co-authors Christophe Moussu and Arthur Petit-Romec.

High returns may point to a solid business model, Moussu said in an interview, but it's hard to disentangle the "bad ROE" from the "good ROE," since the relatively opaque nature of bank operations makes it hard to see in real time how much risk banks are taking.

The paper by U.S.-based academics looked at banks' stock performance in the middle of crises, finding that higher-ROE banks were hit harder on days that were tough for all bank stocks and the stock market more broadly.

This year's mini-crisis in the banking sector provides some evidence for the paper's thesis. Western Alliance Bancorp, which ranked #1 on American Banker's list of top-performing large banks this year because of its industry-leading return on equity, was among the banks whose stock prices were hit hardest after SVB's failure.

The Phoenix-based bank stayed alive after an aggressive effort by its management to convince depositors that the company was sound. Still, its stock price remains down 23% this year. 

Two other banks in American Banker's top 10 — Dallas-based Comerica and Salt Lake City-based Zions Bancorporation — are down 39% and 30% this year, respectively.

Comerica and Western Alliance declined to comment for this story, while Zions did not respond to requests for comment.

The second-highest bank in American Banker's rankings the prior year was SVB, which reported an eye-popping return on average equity of nearly 18%. Ultimately, the Santa Clara, California-based bank faced a quick demise when depositors woke up to its shaky financial position — a large portfolio of flopped bond investments combined with an eye-popping amount of uninsured deposits.

More evidence for the recently published paper's thesis can be found in the oil bust of the 1980s, when Texas banks were "the darlings of the industry," said William Isaac, who was the Federal Deposit Insurance Corp.'s chairman at the time. Isaac later became the chairman of Cincinnati-based Fifth Third Bancorp and now leads the consulting firm Secura/Isaac Group.

During the 1980s, nine of the 10 largest banks in Texas failed because their "high return on equity was illusory" and masked major weaknesses, Isaac said. While measuring a bank's return on equity is important, Isaac said, investors can't "look at it in isolation."

Jeff Davis, a longtime bank analyst who now works with banks at the valuation and advisory firm Mercer Capital, agreed that a more comprehensive look at bank strategies, funding makeup and asset quality is necessary.

"It's a mosaic," he said. "It's not a single data point." 

Still, he said it's critical for bank management to look at what they achieve for those who buy the company's stock.

"You're running it for the shareholders. They own it," Davis said. "And the number-one metric is return on equity, since it's their equity."