Larry Summers Makes the Case for Higher Capital Requirements

September 20, 2016

How can we tell if the biggest banks have gotten safer? One way is to look at their balance sheets, especially the amount of debt they have. And there we can see that their leverage has decreased compared to where it was before the crisis.

But another way is to look at how the market trades the equity of those banks in the stock market. Are the stocks volatile, indicating risk? In a new Brookings paperHave Big Banks Gotten Safer?, Natasha Sarin and Lawrence H. Summers find that in big categories of traded activities, the biggest banks are where they were before the crisis. But rather than signaling the futility of financial reform, I think this paper makes an excellent case for the successes so far, and for pushing to further increase capital requirements.

One of the key pieces of evidence Sarin and Summers presents is the market beta. Beta is the relationship (covariance) between an individual stock and the overall stock market; a higher beta means that the stock will move more when the stock market moves. There are other reasons to care about market beta, especially for portfolios of stocks, but for our purposes it’s a useful measure of a stock’s risk that measures not just the stock by itself but also the stock’s relationship to the broader market.

Sarin and Summers find that for the biggest banks, “bank beta has actually increased in the aftermath of Great Recession.” And although “bank betas have been falling since the crisis, they

have yet to dip below the pre-crisis [2002-2007] levels.” The authors have other evidence, but much of the argument is related to this measure (and it’s the best one publicly available), so let’s dig deeper.

Beta A-Go-Go

To give you a visual of what they find, here’s a quick two-year monthly trailing estimate of beta against the S&P 500 for four of the biggest financial institutions: Citigroup, Goldman Sachs, Morgan Stanley, and Wells Fargo. There have been no fancy adjustments because this is a blog. The estimate trends hold for three years and five years, but I want to display it at two years to get it closer to our current time; as it’s trailing, any monthly measurement estimates the previous 24 months.


So the first thing to note is how low the beta was before the crisis. That shouldn’t surprise us: The market as a whole underestimates the risks of the largest financial institutions during the mid-2000s. Notice how low the risks were going into the financial crisis: The beta was below 0.7 for several of these stocks, an estimate I personally associate with stores that sell diapers and beer—things that are going to make a steady profit rain or shine. This couldn’t hold when it turned out finance was far riskier than anyone knew. Market perceptions of risk can drive up beta, and that’s an important driver here.

It’s even clearer when you look to the 1990s. Note that Sarin and Summers start their analysis in 2002, but by taking the beta data back to the ‘90s (Goldman only goes public in 1999) we can see that beta levels were even higher back then. So right there we know that there was something wrong in the pre-crisis period.

As you can see, Sarin and Summers are correct that beta increased in the years after the financial crisis to above pre-crisis levels. But notice that beta is trending downward since 2013 for these four big institutions, down to levels below where they were in the 1990s. Higher capital has a linear relationship with lower beta, and as the market is finally seeing what the new landscape will look like after Dodd-Frank and higher capital requirements, beta is coming down. Even with a greater awareness of risk, beta is back to trend lows. But it can and should go lower.

So we should raise capital further. Second, this shows Dodd-Frank is the right venue through which to push reforms; these risk measures can and will go lower with higher capital requirements. Note that beta is falling even after the implied TBTF subsidy has also fallen, which should increase beta and other market risk measures. To get both to go down is an accomplishment, and higher capital requirements would do this.

Sarin and Summers find that small and mid-sized banks haven’t seen a downturn in risk measures over the past several years, which gives us a sense that the graduated way that Dodd-Frank works is turning out well. As they find, “On many measures it appears large banks do better in the postcrisis period than their smaller counterparts, suggesting that at least to a certain extent, regulation aimed at lessening risk of large systemically important financial institutions is having the intended effect.”

The Brookings paper asks whether capital requirements could be “gamed” in such a way that they are binding on a balance sheet but not in real life, or at least not to the view of the market. To me, the important thing is to create well-diversified capital requirements that touch each part of the balance sheet.

You need risk-weighted requirements to ensure both the due diligence and that you are touching the assets side of the sheet. JPMorgan dropped in size to reduce its exposure to risk-weighted surcharges, showing that those charges aren’t gamed away. Because of the nature of the business, you also need debt regulations to reduce short-term borrowing and ensure that liquidity does not collapse. And you need leverage requirements as a baseline, especially in case an asset class collapses in a crisis. Raising all of these requirements in proportion would ensure that any one of them isn’t letting the banks off the hook.

Finally, we should remember that maximizing “franchise value” is not a goal of public policy. Sarin and Summers believe the underlying dynamic is a decrease in future profitability of banks relative to the pre-crisis era. They argue that this franchise value and related price-book ratio has declined, partially driven by new regulations and enforcement. They hint that new regulations might make the situation worse, because whatever stability is increased could be offset by reduced profitability, reducing franchise value.

But we don’t set up our financial regulations to ensure that banks make maximum profits. There are public policy goals of protection, accountability, and stability, and they will all have a push-and-pull effect on franchise value. Not investigating and seeking damages for bad actions such as LIBOR manipulation and Wells Fargo’s fake accounts would boost franchise value, but this would be a bad for protection. Committing to permanent bailouts would dramatically reduce beta but would be an affront to people’s demand for accountability. But having a transparent consumer lending market is conducive to more economic activity, while clear and accountable emergency lending powers help prevent a crisis. Those measures both boost banks’ business. The goal is a secure capital market, not any bank’s profits.

Indeed, this was an argument against the Consumer Financial Protection Bureau. While Dodd-Frank was passing, back in March 2010, Senator Shelby told a crowd of cheering bankers that “safety and soundness trumps everything, it trumps the consumer finance whatever.” Defrauding consumers is certainly one way to boost soundness and franchise value in the short-run; but beyond the unfairness, high levels of fraud eventually trigger large asset-level devaluations, which also hurts soundness, as it did in 2008.

There are standards that the public demands from the financial markets, be they increased scrutiny of derivatives or having interchange credit card charges treated as a public utility. These are either good or bad ideas on their merits, but to not do them in order to avoid the work of raising capital requirements is obviously a giant mistake. There isn’t any evidence that the costs of higher capital requirements would reduce franchise value more than they’d help stability, and there’s every reason to believe they’d make for a better market.