Capital Rules Aren't the Only Reason Banking Is Boring

My cheerful assumption that funding models change, but the business of banking is eternal, seems to be wrong.

He's never bored though.

Photographer: Mark Wilson/Getty Images.

This morning I proposed a simple speculative model of how banking has been transformed. In this model, the business hasn't changed much, but higher capital requirements have made it less profitable for shareholders: You might make the same amount of money doing the same stuff, but you have to share it among more shareholders. This obviously wasn't the whole story -- banks are also making less money and doing different stuff -- but I suggested that it had some explanatory power.

So I figured I should check it against the facts. And ... nah, not that much explanatory power.

The first stylized fact that I discussed this morning is that former Treasury Secretary Tim Geithner and his colleagues and successors have "cut the profitability of banking roughly in half," as he put it in this morning's New York Times article. So is that true? And then some! Here is a chart of the profitability -- return on common equity -- of the six big usual-suspect U.S. banks (JPMorgan, Bank of America, Citigroup, Wells Fargo, Goldman Sachs, Morgan Stanley) for the last dozen years: 1


Loosely speaking, the average return on common equity for the (surviving!) big banks in the five years preceding the crisis was a bit more than 17 percent; now it's about 6 or 7 percent. So it's fallen by well more than half.

Is that because banks now have more than twice as much capital? Ehh, no. Here's a chart of common equity to total assets, both as reported under generally accepted accounting principles; that's not exactly a measure of regulatory capital, but it gestures in the same direction: 2


So long-term average capital is up from less than 7 percent (and stable-ish) to a little less than 9 percent (and rising); in 2014 it got to about 9.5 percent. So it's up, but only by about a third. Banks went from about 14 times levered to 10-11 times. 

The big culprit for declining profitability is just lower returns on assets, which averaged about 1.1 percent before the crisis and about 0.6 percent after, creeping up to a little less than 0.8 percent in 2014:


Loosely speaking, the fall in return on assets -- just profitability of actual business -- accounts for about two-thirds of the drop in overall profitability; the increase in capital accounts for about one-third. 3

Meanwhile total assets at the big six banks are up, though that has much to do with crisis-era mergers; they're drifting up very slowly now:


So go back to the Times story about how banking has changed. Here's the Times:

The capital rules have had the effect of encouraging banks to focus on parts of their operations in which they are potentially taking fewer risks -- like the divisions that manage money for pensions and investors -- and de-emphasizing the trading desks.

Here's Marianne Lake of JPMorgan, talking about capital:

We are trying to thread the needle, as you say, about making sure that we are as focused as we can be on maximizing the use of that scarce resource.

And Brady Dougan of Credit Suisse:

It’s become much more a game of driving the highest returns from the businesses that are most suited to the new environment.

All of these things are obviously true, but they don't show up so clearly in the big-picture numbers. The numbers don't tell a story of banks cutting back on balance sheets and focusing only on the highest-return businesses in a new capital-constrained environment. They tell a story of growing balance sheets, and of a shift toward businesses that make lower returns on assets, relative to the good/bad old days.

Some of that is probably because my charts look at un-risk-weighted accounting assets, and risk-based capital rules might be driving banks to focus on low-risk businesses that use less capital per dollar of assets. But there is a popular view that the (un-risk-weighted) leverage ratio is (and should be) the most binding constraint on bank capital, and that capital regulation in fact imposes punitive costs on low-risk cash holdings. You'd expect the overall effect of post-crisis capital rules 4 to be to push banks away from asset-intensive businesses (risky or not) and into fee-based businesses; the Times cites "divisions that manage money for pensions and investors" and Morgan Stanley's shift from a trading focus to an asset-management one.

What should we conclude? The optimistic view is that banks are focused on maximizing returns on capital, but they haven't quite figured it out yet. There will be a transition period while banks find "the highest returns from the businesses that are most suited to the new environment." Once they find the mix that's optimized for current regulation, my dumb simple model -- similarly profitable businesses, but lower returns (and risks) to shareholders because the businesses are supported by more capital -- should roughly work. And the graphs do hint at that: In recent quarters, returns on assets and capital ratios are both creeping up, while return on equity is staying relatively flat.

A variant on this view would be that banks are still adjusting on the cost side. By reducing headcount, for one thing, but there are other costs that might go away eventually. Post-crisis returns on assets would be measurably higher if Bank of America and JPMorgan and Citi had back the tens of billions they've spent on mortgage and foreign-exchange and other regulatory settlements. Surely one day the settlements will stop?

The pessimistic view would be that the activities that banks do are now just permanently less lucrative -- that the secret sauce of pre-crisis bank profits was, say, the sort of proprietary trading that has now been banned by the Volcker Rule and migrated out of banks. Or the secret sauce was fraudulently selling mortgage-backed securities, and now that option is no longer available either. Or perhaps costs are permanently higher: You need to pay more to compliance officers, and information-security technicians and long-term holding-company bondholders. Or the settlements won't stop.

I don't know! In any case, though, the changes to the business of banking do seem to be more significant than the changes to its funding model, and just doing the same basic stuff with a lot more capital probably isn't the future of banking.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

  1. Data is from the Bloomberg Return on Common Equity field, from the first fiscal quarter of 2003 through the third fiscal quarter of 2014. The average is weighted by total assets; the unweighted average is on average 75 basis points higher. High and Low are just the highest and lowest ROCEs among the six banks in each quarter; obviously different quarters have different top and bottom banks. The two five-year (umm, 19-quarter) averages are just averages (not weighted by assets over time -- i.e. each quarter counts the same) over Q1 2003 through Q3 2007, and Q1 2010 through Q3 2014; they are respectively 17.4 percent and 6.5 percent.

  2. And without the potential controversy of risk weighting. Anyway, the explanation for this chart is the same as above. Five-year averages are 6.7 and 8.9 percent.

  3. Like, the dumb math is:

    • Pre-crisis return on common equity was about 17.6 percent, on return on assets of around 1.15 percent and common equity capitalization of around 6.72 percent.
    • Return on common equity = (return on assets) x (assets / common equity). (That's not quite right, as Bloomberg's ROCE uses income net of preferred dividends, while ROA uses income including those dividends, but close enough.)
    • The effect of ROA going from 1.15 percent to 0.63 percent, just arithmetically, would reduce ROCE from 17.6 percent to 9.6 percent (7.9 percentage points).
    • The effect of the equity-to-assets ratio going from 6.72 percent to 8.86 percent, just arithmetically, would reduce ROCE from 17.6 percent to 13.3 percent (4.2 percentage points).
    • The effect of the two combined would reduce ROCE from 17.6 percent to 7.3 percent. (It's actually 6.5 percent, so this dumb arithmetic of averages over time doesn't entirely explain the change, though it comes close enough.)
  4. As opposed to the pre-crisis, "put all your money in AAA CDO-squareds!" kind of capital regulation

To contact the author on this story:
Matt Levine at

To contact the editor on this story:
Zara Kessler at

Before it's here, it's on the Bloomberg Terminal.