This spring, as Professor Harry Mamaysky watched the pace of negative news stories about First Republic Bank pick up, he tracked the corresponding market reaction. When he later analyzed the data, he wasn’t surprised to see a delay in how long it took investors to fully absorb the poor reports.
Mamaysky, professor of professional practice in the Finance Division at Columbia Business School, says he spotted what he regards as a typical correlation but also often a gap between the time new information emerges and a related shift in stock and bond prices. “There are lags from when the news is released until market prices fully respond,” says Mamaysky, who has studied the effect of news on markets over time.
This phenomenon may affect investors personally, but it can also play out on a market-wide scale — prompting delayed reactions to major economic events. Mamaysky predicts that new technologies may help flag noteworthy information earlier, closing the lag somewhat. Both individuals and the overall market could benefit from a more rapid absorption of relevant information, he notes, routing out risk early to prevent systemic problems.
Mamaysky says as he analyzed the news surrounding the 2023 banking crisis and how it affected First Republic Bank this spring, it is only in retrospect, when he pieced together articles about the bank stretching back to January, that he saw they forewarned the troubles that finally led to the bank’s closure in May.
The first sign of trouble began January 13, when First Republic executives discussed on an earnings call the impact that rising interest rates would have on the amount the bank would need to pay depositors, he says. The following month, the bank disclosed in its annual Securities and Exchange Commission 10-K filing that the firm’s loan portfolio held a heavy concentration in interest-only mortgages, which typically have a greater exposure to rising interest rates. As investors watched the collapse of Silicon Valley Bank in March, they tried to minimize exposure to First Republic — arguably the bank’s closest peer — causing the stock to fall dramatically.
Later that month, Federal Reserve Chair Jerome Powell made a statement noting that credit conditions were likely to tighten for businesses and households due to turmoil developing in the banking sector. In early April, MarketWatch published an article looking at how First Republic Bank was one of the most heavily shorted among financial stocks. And finally, later in April, a tweet citing Bloomberg.com statistics commented on First Republic’s large number of fixed-only mortgages and their relationship to rising rates.
So, why didn’t investors pick up on these warning signs and flee First Republic’s stock at the first sign of trouble? Mamaysky explains that people are limited in the number of things they can focus on at any given time, so many investors will first focus on the big banks that are in the news. “For this one company, there was a lot of information available, but no one was really looking at First Republic,” he says.
Mamaysky points to studies showing that when company news is released, it typically takes between one and 10 days before stock markets fully reflect the news. And that lag is exaggerated when the news landscape is cluttered — if there is more than one noteworthy event in a week, it can take up to three months for the stock price to fully register the news, he says.
The bank was seized by regulators on May 1, but Mamaysky notes that while the problems in the industry were emerging as early as January, it took several months to unfold to the public. If all the issues had been flagged earlier, he says, maybe there wouldn’t have been a run on the bank and it would not have failed.
“Its failure was not a foregone conclusion,” Mamaysky says. “If everything was revealed earlier, maybe the bank would have survived.”
A Technological Solution
What could make a difference? Mamaysky and his colleagues believe the application of AI technology similar to ChatGPT could hasten analysis by screening news reports and regulatory filings. That was one of the topics Mamaysky explored at the 7th Annual News & Finance Conference, held at CBS in May. He co-organized the event with CBS’s Paul Tetlock, the A. Barton Hepburn Professor of Economics in the Faculty of Business in the Finance Division, and Paul Glasserman, the Jack R. Anderson Professor of Business in the Decision, Risk, and Operations Division. “Algorithms can process millions of texts and try to find these early-warning indicators that people just don’t have the capacity for,” Mamaysky says. “There’s just not enough time in the week to look through everything.”
Mamaysky anticipates that capitalizing on AI technology could help identify patterns or themes in overlooked corners of the market. As he explains, investors hang on every word when Microsoft’s CEO speaks at an industry event or announces a new product. “But when you have a smaller company like a regional bank, and an hour and half into an analyst call someone asks a question that is relevant, that doesn’t get attention right away,” he says.
Preventing a Crisis?
Using AI to look for patterns in our blindspots may not just benefit investors looking for profit. It could also improve the health of the overall market by heightening awareness of important developments in individual companies or sectors, Mamaysky says.
“We have access to so much information, but we don’t pay attention to a lot of it, even though it might be important,” he says. “Using AI, we can gain insights from this greater volume of information, to understand the things that we should be worried about.”
For example, he wonders if such tools were available during the financial crisis of 2008 and 2009, might they have highlighted the risks of banks’ exposure to subprime loans sooner? Such connections might have appeared as early as 2006 as precursors to the financial crisis, Mamaysky says, notings that the first news stories about subprime lending and securitization tendencies were published two years before the situation became a crisis. “People were unable to make the connections between what was happening with subprime mortgages and the broader financial system,” he says. “Two years down the road, the whole financial system stopped working.”
As Mamaysky points out, an earlier appreciation of the risks might have given regulators an opportunity to step in and slow down subprime lending, potentially averting the crisis. “They might have taken Lehman’s stock down to 50 percent, but maybe it would have put the fear of God in that management team two years earlier and they could have righted the ship and not gone bankrupt,” he says. “It would have taken a giant crisis and turned it into a mini crisis if we just had better information.”
Watch Professor Tano Santos, the Faculty Director of Value Investing and Advanced Value Investing programs at Columbia Business School, discuss the school’s approach to teaching value investing and finance: