Haubrich: Let’s start out with a broad question. How would you rate the U.S. response to the financial crisis?
Calomiris: It’s a great question. I would start by distinguishing between long-term response and short-term response. I’d say that the biggest short-term failure was not to require recapitalization of the investment banks between March and September of 2008. I would have liked to have seen recapitalization happen during calm markets when it was obviously necessary.
Also, with respect to the immediate response to the crisis, I would have designed TARP [the Troubled Asset Relief Program, signed into law in October 2008] differently. I don’t think TARP was well-structured. Congress got to write the form of capital assistance for banks as a footnote to TARP, and it was structured to make a profit on the transactions rather than to encourage the right kind of stabilization assistance.
In many ways, our response in the 1930s was better. I think there were lessons from history that we could have put to use in this crisis. And we suffered a lot, unnecessarily, with a liquidity crisis that wouldn’t have been so deep if we had used better ideas back in September and October 2008 about how we were going to support the financial system.
The Fed’s short-term actions were mostly appropriate—making liquidity available to the markets in various forms, particularly TALF [Term Asset-Backed Securities Loan Facility].
I think this was a successful and warranted program. It conforms to what I take to be the central message of a what a lender of last resort has to do, which is try to take some risk on its balance sheet during a crisis, but to do so in as senior a way as possible. I’m not saying the Fed got everything right.
I think that the Dodd-Frank bill, which is the main form of long-term response to the crisis, suffers from both sins of commission (bad ideas enacted in haste) and sins of omission (it doesn’t really deal with some of the key problems that underlay the crisis). The key problems that we should have learned about from a prudential regulatory standpoint were, number one, the subsidization of risk in housing through high leverage, effectively financed by the government, either explicitly or implicitly. We need to address that.
Second is the failure to accurately measure risk in the financial system on a forward-looking basis and to require capital accordingly. Going forward, we really need to address that problem, too. And third is the too-big-to-fail problem. I don’t thinkDodd-Frank adequately tackled that problem.
Haubrich: Part of the Dodd-Frank bill1 was setting up a variety of institutions, such as FSOC [Financial Stability Oversight Council], which are presumably going to provide stricter regulation for the systemically important or too-big-to-fail institutions. Do you think that will be an adequate response?
Calomiris: My view is that we should have an incentive scorecard for any regulatory idea that asks how it is going to affect the incentives of people in the marketplace directly. And how is it going to affect the incentives of regulators, supervisors, and politicians to live by the rules they write? The creation of a new bureaucracy is not really getting at that in any direct way. Maybe it will help, maybe it will hurt. But it doesn’t really get at the key problems.
I prefer other ideas that are under study and make a lot of sense. Contingent capital certificates and the restructuring of capital requirements are very promising. There are several people in the Federal Reserve System who are interested in that. If the oversight council is going to be a way to get good ideas like that formulated, then in conjunction with those new ideas it could be effective. So, it’s not a bad idea that we’re going to have more of a focal point on responsibility for coming up with good ideas in some coordinated way. Maybe it will help, maybe it won’t. But the ultimate test is going to be whether we see real mechanisms that matter for incentives coming out of those deliberations.
If there is a bona fide reason to promote affordable housing, it’s to create stakeholders in local communities. But you’re not a stakeholder if you have a 3 percent down payment on your home—you are a renter in disguise.
Haubrich: You mentioned one aspect that hasn’t been dealt with: the incentive for leverage in the housing market. What would you do about the government-sponsored enterprises Fannie Mae and Freddie Mac?
Calomiris: I think that they should be phased out. I think that all government assistance to housing that’s taking the form of lending programs that try to subsidize affordable housing through making lending easier, and especially through very high leverage and very low interest rates, are the wrong way to subsidize housing, for two reasons: First, it encourages systemic risk, just like what we’ve experienced. A little bit of a decline in housing prices causes huge distress throughout the financial system, precisely because of leverage.
Also, leverage encourages the wrong kind of risk that comes to the market because borrowers are not placing enough of their own resources at risk; people who are bad risks are willing to come to the mortgage market. So you get a bad incentive consequence in advance. Leverage also matters after the fact, by magnifying the losses in the financial system created by recession shocks or asset price declines.
We will need the political will to move away from the drug of leverage, which was attractive to politicians in the first place because it disguised thegovernment’s costs.
If leverage is not the way to promote affordable housing, then that means it is high time to phase out FHA [the Federal Housing Administration], Fannie Mae, and Freddie Mac. Not only is mortgage risk subsidization through high leverage risky, it also fails to achieve its goal. If there is a bona fide reason to promote affordable housing, it’s to create stakeholders in local communities. But you’re not a stakeholder if you have a 3 percent down payment on your home—you are a renter in disguise.
These extremely low down payments are a very recent trend, really the last two decades. It was all part of the desire to create invisibly—from a fiscal standpoint, that is, without showing the costs on the government’s balance sheet—these government supports, through high leverage and subsidized interest rates. But this approach doesn’t accomplish the housing objective, and it destabilizes the mortgage market and creates large costs to taxpayers.
Haubrich: OK, so what specifically would you do to ensure that homebuyers are stakeholders in their communities?
Calomiris: I propose a four-part plan: alongside phasing out Fannie Mae, Freddie Mac, and FHA as lending agencies, to create a downpayment assistance program modeled after the Australian program but on a meanstested basis. Every Australian qualifies for a $7,000 first-time homebuyer subsidy. This subsidy increases their down payment, reduces their leverage, and makes their home more affordable. It has a stabilizing effect on leverage ratios and creates a real stake for people in their homes and communities.
A second part of that plan is to phase in, over a 17-year period, movement from the minimum 3 percent down payment requirement to a 20 percent minimum. A third policy that might also make sense, again on a meanstested basis, is to provide some assistance for the cost of locking in longer term interest rates for low-income people because they’re potentially more susceptible to the debt-service fluctuation cost.
Finally, a fourth part of that plan might be to create home savings accounts that provide tax incentives for people to accumulate equity toward the down payment, again on a means-tested basis. As part of that, it would be interesting to think about also using some tax savings from payroll taxes, because most low-income people don’t pay income taxes; they only pay payroll taxes.
All of these proposed costs would be budgeted explicitly. Let’s have transparency so we show the government expenditure effects of these programs. Let’s create systemic stability, not instability. Let’s create homeowners, not renters in disguise. To me this would be a very rational approach to housing finance reform. We will need the political will to move away from the drug of leverage, which was attractive to the politicians in the first place because it disguised the government’s costs. Well, it’s a little hard to disguise them now that the costs of Fannie Mae and Freddie Mac’s losses are likely to top $300 billion. Maybe that means we’ll get the political will to be a little more honest.
Haubrich: Some people would say that Fannie and Freddie, at least in their early days, did a lot to promote homeownership—even for people with high down payments—by creating a mortgage securities market. Do you think the mortgage securities market could survive without them?
Calomiris: Yes. I agree with the statement that, as a historical fact, the creation of Fannie Mae was helpful for creating homogenization of a long-term mortgage market, and a mortgage-backed-securities market. That was a one-time contribution. Now that market exists. We have the advantages of the economies of standardization, the invention of the product. Fannie Mae is not needed as an inventor anymore; they did that a long time ago, and the private market can carry on just fine without them. Of course, the private market isn’t going to do 97-percent-leveraged mortgages at low interest rates! The reason the private market isn’t participating so much right now is because they’ve been crowded out by leverage-subsidized government lenders. If those leverage-subsidized government lenders disappear, the private market knows exactly how to come in and take their place once we allow reasonable down payments in the market.
Haubrich: Let me switch gears and talk a little bit about the subject of the conference we’re hosting. Can countercyclical capital regulation work?
Calomiris: I think it can. First let me define it. We know that capital requirements can constrain bank lending under some circumstances. It’s not just government capital requirements but also market capital requirements. Remember, government capital requirements are only about 30 years old in the United States, and we were the first country that I’m aware of that passed them. Capital requirements mean requiring a certain ratio of your assets to be in the form of equity capital.
It’s a fairly young idea. Before the government required it, the market required it. Now both the government and the market require it. The one that’s the binding constraint—the one that has the effect in determining capital requirements—is whichever is the higher of the two.
People are worried about two problems. First, during recessions, banks may lose capital as a result of loan losses. For example, during a recession, to preserve their capital-to-asset ratio, they may have to cut their risky assets, meaning their loans. The other thing people worry about is that going into booms, that constant capital ratio maybe isn’t high enough to discourage excessive lending. These are the two arguments people make for dynamic capital requirements; that is, maybe we want capital requirements to be higher during booms as a way to discourage excessive lending. And maybe we want capital requirements to fall during recessions as a way to avoid really severe credit crunches. I think there is at least some evidence for the tendency for excessive lending in some booms. I don’t think it’s a general problem, but it does happen occasionally.
The key question is whether we can measure it. Can we measure when lending markets are excessive and when lending markets are going through recessions? We’d like to vary capital requirements—to relax them during recessions and to increase them during booms, especially booms where we’re very worried about excessive lending. Is this something that can be measured? I would say yes. It can’t be measured perfectly but it can be measured fairly well.
A paper by Claudio Borio and Mathias Drehmann at the BIS [Bank for International Settlements] showed that very severe recessions tend to happen when lending growth is very high just prior to the recession, during the boom, and when asset prices are growing very fast. So one could impose higher capital requirements based on a dual threshold. When loan growth gets very high and asset prices are growing very quickly, you’d like to impose higher capital requirements, and doing so will help you achieve a better soft landing, cooling down the loan growth and helping the economy.
Haubrich: Now capital requirements are changing. Basel III2 is coming up with a set of recommendations. Could you give us your take on those proposals?
Calomiris: They’re talking about phasing in an increase in capital requirements for Tier 1 capital. It’s a maximum leverage requirement rather than just a risk-based capital requirement. These are good ideas. But are they enough? No.
A little increase in capital doesn’t solve the problem. The key problem was the failure to measure risk on an ongoing, forward-looking basis, fairly accurately, and to require capital accordingly. If you just bump up capital a little bit, what’s to prevent risk from going up even more? If financial institutions want to create risk and want to hide it from their regulators, under the existing Basel system it’s almost trivial to do it.
The Basel system still, unbelievably, says that the way we measure risk is asking banks what the risk is and asking ratings agencies what the risk is. That is not dealing with the incentive problems that got us into the mis-measurement of risk in the first place, so we have to think more creatively about mechanisms that can solve this problem.
The Basel system still, unbelievably, says that the way we measure risk is asking banks what the risk is and asking ratings agencies what the risk is.
One paper that was published in 2003 by Don Morgan and Adam Ashcraft in the Journal of Financial Services Research shows that interest rate spreads on loans are very good predictors of loan default risk. In the common parlance we would say, “Duh!” Because after all, that’s the point of spreads. They’re supposed to compensate banks for risk. The authors showed in 2003 that we could’ve used interest rate spreads as good forward-looking measures of risk. We didn’t use them! If we had used them in the subprime crisis, we would’ve budgeted a lot more capital against subprime risks and we would’ve been better off. So ideas coming out of Basel to tweak capital requirements in a way that’s not related to measuring risk are doomed to fail.
What we need to do is take risk measurement seriously in a way that deals also with incentive problems. Notice that my proposal to gear default risk measures to loan spreads is incentive robust.
Why? Because no bank is going to cut its interest rate to save a little bit on its capital requirements. That means that interest rates will provide robust measures of risk, and we can therefore reliably use those interest-rate spreads to measure risk. That’s just one of several ideas that illustrate the idea of incentive-robust regulatory reforms. These reforms would sometimes force banks to maintain more capital—but capital commensurate with risk—using measures of risk that we could rely on.
Haubrich: Some people argue that high enough capital requirements would reduce the incentive that banks have for taking risk. It sounds like you don’t agree with that?
Calomiris: The problem is that financial firms are designed to be able to arbitrarily increase risk. Financial contracts can reshape and re-cut risk in many different ways to create a little risk or a lot of risk, depending on what they want to achieve. You can’t say that a loan has X amount of risk, because banks can construct loans that might appear very similar that have very different kinds of risk. We have to have a flexible means of measuring risk.
What we really have to ask is what was missing in regulation. What was missing was what we’ve been talking about: accurate measurement of risk on a forward-looking basis.
If we just say we’re going to increase capital a little bit, banks could very easily just make sure that the composition of risks of the loans in their portfolio are commensurately higher, therefore achieving nothing in terms of stabilizing a system.
That’s the story of what is sometimes called regulatory arbitrage, that is, the private sector undoing the effect of any regulation. If the regulations are dumb, the private sector will always smartly undo them. The regulations have to be smart enough so that they’re robust to incentives.
Haubrich: To follow up on that, to what extent do you think the problem behind the crisis was, say, regulatory arbitrage, and to what extent was it deregulation?
Calomiris: I’ll start with the second part of the question. Deregulation had nothing to do, in my view, with the crisis. The main deregulation that happened in the United States in the 1980s and the 1990s really dealt with two important issues. One of them was whether banks should be allowed to branch throughout the United States. In 1994, national Â deregulation of branching was a culmination of state and regional deregulation of branching that was occurring throughout the ’80s and early ’90s. That was a stabilizing policy. It has been shown time and time again for a whole host of reasons to be a very good idea as a matter of economic policy.
The other main deregulation was for banks and investment banks. It allowed banks to engage in the underwriting of corporate securities, which they previously had been limited in doing. This crisis had nothing to do with the underwriting of corporate securities. Furthermore, subprime mortgage activities, which were important in creating the crisis, were something banks could have engaged in prior to the two deregulations I’ve talked about, and in fact did. Deregulation allowed commercial banks to engage in traditional investment-banking and corporate-underwriting practices, and allowed commercial banks to branch. These were stabilizing, in fact, during the crisis. Why? Because of greater diversification of income.
Furthermore, remember, the way we dealt with this crisis was by allowing investment banks to be purchased by commercial banks. That was possible because of deregulation, and it helped to stabilize the system in reaction to the crisis. When you hear the political rhetoric about deregulation, I think what people really mean is that we had a failed prudential regulatory system. Banks and investment banks were both under prudential regulation under the Basel standards. The investment banks were being regulated, under Basel, by the SEC [Securities and Exchange Commission]. And what we can say is that it wasn’t a very good regulatory system. We did not maintain capital commensurate with risk very effectively.
But starting in 2002, the investment banks were all regulated, long before this crisis hit. If anything, prudential regulation was increasing over time during the phase when the crisis took hold. What we really have to ask is what was missing in regulation. What was missing was what we’ve been talking about: accurate measurement of risk on a forward-looking basis. The other problem that arose between March 2008 and September 2008 was that once we bailed out Bear Stearns, others expected bailouts instead of recapitalizing as they needed to. They didn’t want to recapitalize in a way that would dilute their stock values. Why not take a bet, hope that if things go badly they’ll get bailed out? If things go well, they won’t have to dilute by issuing new equity. That too-big-to-fail problem needs to be addressed, too. We’ve done little to address those two basic problems.
Haubrich: You mentioned both the problem with too big to fail and the advantages to banks diversifying across the country with a lot of branches. Where do you come down on the advantages or disadvantages of the increasing concentration of the U.S. banking system? That is to say, does it matter that we have a handful of global banks that account for so much volume?
Calomiris: That is a difficult topic, because it has at least three dimensions to it that I can think of. One of them is the question of whether there are economies of scale in some technical, economic sense in banking. The simple answer to that question is yes, but there aren’t economies of scale to the simple deposit-taking and lending functions of banks beyond a fairly small level. It does make sense—as Canada has done and just about every other country and finally us, too—to have nationwide lenders that are diverse from the standpoint of deposit-taking and lending. But that doesn’t mean that you only have three or four of them. You’re not going to get big enough economies of scale coming out of deposit-taking and lending to justify the size of the largest banks that are operating in the U.S. right now.
That brings us to the second issue. The largest banks are really operating a global platform where they’re not just in the business of deposit-taking and lending. They’re in the business of global relationship management for global nonfinancial companies. Those global nonfinancial companies need to perform transactions in scores of countries and in hundreds of different products. The platform that’s necessary to physically deliver that has to be global and it has to have huge overhead costs associated with it; and therefore, to be in all those places and to absorb all that overhead, global banks have to be huge.
The biggest confusion in the literature has been confounding the benefits of larger scale for banks involved in simple deposit-taking and lending, which are small, with the benefits of scale from global relationship management, which are substantial. We know there aren’t large economies of scale to traditional deposit-taking, and all of our statistical studies are about that. But they sometimes are misunderstood as having something to say about global universal banks. And they don’t have anything to say about them. The answer to the economic question—do U.S. banks need to be extremely large?—is that to be global relationship managers, they do. Would we be better off if we had 10 instead of three such huge banks? I think the answer is, obviously, yes. Why do we want to have 10 instead of three? Because, as we’ve just learned in the crisis, we want to be able, if one of them goes bad, to find something to do with it. If we’re down to only three, and one of them goes under, we’re down to a duopoly in the banking system, at least from the standpoint of global banks, and that’s not a good idea. The hard part is how we get there.
Perhaps it would be desirable to break up the big banks. I’d like to see this done, if it is done, in a fairly cooperative way, to find a way to get the large banks to simply spin off some of their business to create new banks. I don’t have an easy solution for how to do that. But I think that would be a desirable outcome if it could be done.
I think we can make a lot of progress on the too-big-to-fail issue by statutorily limiting the percentage of unsecured creditors’ principal that can be bailed out by the FDIC [Federal Deposit Insurance Corporation]. Currently, the FDIC can bail out 100 percent of any principal of any debt on planet Earth, if we go through a few procedural hoops that aren’t very likely to prevent unnecessary bailouts. But if we limited the percentage of debt principal that could be bailed out to no more than 90 percent or 85 percent or maybe, to get more aggressive, 80 percent, then this statutory limitation would put large creditors on notice that even if the bank gets bailed out, you’re not going to be spared some losses. That starts creating market discipline within these banks.
Haubrich: We’ve talked a lot about banks and bank regulation. You’re, if not unique, at least very special among prominent banking scholars in that you’ve also spent some time actually running a commercial bank [Greater Atlantic Financial Corporation]. What sort of perspective does that give you, from both sides of the aisle?
Calomiris: I learned a lot of things from that experience. One thing I learned by being head of the credit committee was that finance is actually not wrong! That is, a lot of the things we worry about as finance theory people, problems having to do with incentives of borrowers, are very real, and bankers have been incorporating them into the way that they engage with borrowers from time immemorial. In particular, a fancy idea in finance called asset substitution comes into play: If a borrower gets into trouble, the last thing you should do is give them more time with no strings attached, because the borrower is going to have incentive to take on more risk. That is part of the culture of banking and very much ingrained in the way that loan contracts are written and enforced, with covenants and collateral requirements and short-term maturities. So I think one thing I learned is that finance theory is right about bank loan contracts.
Haubrich: We haven’t talked much about monetary policy. Do you think monetary policy contributed in a material way to the financial crisis?3
Calomiris: Yes. From 2002 to 2005, the real federal funds rate was negative. And the only other four-year period in post-war history where that was true was 1975 through 1978, the high inflation period. If you looked at it from the standpoint of the Taylor rule as a function of the unemployment rate and the inflation rate, the Fed stopped following the Taylor rule during the period 2002 to 2005.4 The Fed kept the fed funds rate about 2 percentage points on average below what that rule would have implied. So the Fed surprised the market from the standpoint of the Taylor rule, ran very loose monetary policy with a negative real fed funds rate, and there is pretty convincing statistical evidence that this contributed to the underpricing of risk leading up to the crisis.
That said, I can tell you from a historical perspective that monetary policy mistakes like that—and I regard it as a mistake—do not cause these kinds of crises. They cause asset-price problems, but to get into a banking crisis you need the large losses relative to bank capital, and you can’t blame that on monetary policy. Contrast, for example, with the dot-com boom and bust, where the actual losses were greater than the subprime boom and bust. Yet it didn’t have any effect on the whole financial system. Why? Because it wasn’t a leveraged loss. Housing leverage and banking leverage in the recent crisis translated into an overpricing of some assets and into the demise of the financial system.
Yes, the Fed contributed to the overpricing of real estate assets and other assets, but that doesn’t translate into a financial crisis. You need the other distortions on the microeconomic side, particularly the housing finance distortions, to really understand the depth of the crisis.
Haubrich: What books do you have on your night table?
Calomiris: I have one book called The Lords of Finance, which I’ve just purchased, and it looks like a fun book. I’ve glanced through it and it goes through the actions of the central bankers in the 1920s and the 1930s worldwide, and I’ve heard it’s a very entertaining book.
Haubrich: Thank you very much.
- Dodd-Frank refers to the financial-market, regulatory-reform legislation, sponsored by Sen. Christopher Dodd and Rep. Barney Frank, which was approved by Congress in July 2010. | return |
- The Basel Committee on Banking Supervision has been meeting since 1974 to improve and coordinate the quality of bank supervision worldwide. The most recent effort to strengthen capital requirements was dubbed Basel III. | return |
- For Federal Reserve Chairman Ben Bernanke’s take on this question, see www.federalreserve.gov/newsevents/speech/bernanke20100103a.htm | return |
- The Taylor rule is the monetary policy rule for targeting the federal funds rate that many believe the Fed seemed to be following prior to 2002. | return |