2012 Financial Markets Conference
April 2012
An interview with Phil Dybvig, Boatmen's Bancshares Professor of Banking and Finance at the Olin School of Business at Washington University in St. Louis
Paula Tkac: Today I'm joined by Professor Philip Dybvig from Washington University in St. Louis to talk about maturity transformation. So, Phil, maturity transformation may be simply described, I think, as borrowing short and lending long. Is that really all there is to it, and is that the part that we care about from a regulatory perspective?
Phil Dybvig: Well, I usually think that there are two parts to maturity transformation. One part is that if you have a longer loan and that has a different interest rate because interest rates are different at different maturities, and you have a shorter loan, and you have one on one side of your balance sheet and one on the other side, then when interest rates move you have risk because the shorter loan's interest rate may be moving around when the longer loan's interest rate won't. That I think of as interest rate risk, or term structure risk, and that can usually be hedged by a variety of techniques. For example, using bond futures or using interest rate derivatives.
For the other type of maturity transformation, it's creation of liquidity or destruction of liquidity if it's consumption of liquidity, but usually in the context of banking we're thinking of creation of liquidity. So a bank will have demand deposits where people put their money in, and they can take it out any time they want to based on their own needs in their lives or because they are afraid the bank will fail. Yet, the loans that the bank makes are illiquid in that they cannot be terminated just at any time. If the bank has many withdrawals and people they lend money to are not necessarily going to be able to return the money immediately because you have a longer-term loan and shorter-term deposit.
Tkac: And so what kind of problems does that sort of maturity transformation or the liquidity creation cause for the banking system, or for the shadow banking system and the rest of the financial system?
Dybvig: Well, let me talk about the benefit first...
Tkac: OK.
Dybvig: ...Because the benefit is that that allows people to conduct the business that they want to. If they need money suddenly for a new car or a house, or just to buy dinner today, that money is immediately available, and they have that flexibility. But from the bank's perspective it could be a problem if people take out their money all at the same time. And the particular problem is that if one depositor looks around and thinks that all the other depositors are taking their money out, then that depositor will also have an incentive to take the money out because what the bank can liquidate on the loan side is not going to be enough to cover all the depositors. And so that's the way a "bank run" works. For the whole group it's not rational, they'd all prefer to have the money stay in the bank and just maintain their deposits, but individually it's rational if each person thinks that everybody else is taking their money out they are going to race...
Tkac: A bit of a "prisoner's dilemma"?
Dybvig: ...They're going to race to take their money out as well.
Tkac: So they end up all worse off. So what ramifications, then, does that have? It sounds like that bank could fail. Macroeconomically, how do we think about that in terms of the overall cost to economic activity?
Dybvig: So we want to regulate banks so that this doesn't happen very often. We want to regulate banks so that failures are rare events. If banks are regulated properly, maybe it's not a problem. But in the absence of something like deposit insurance or lending as a last resort by the Fed, then it's possible that bank runs can happen and people are legitimately concerned. If a bank has deposit insurance, then they should be able to point to the medallion from the FDIC or the other insuring agency and say, "Look, you don't have to take your money out now. You'll get your money back later if you don't take it out now. Please don't take your money out unless you really need it, it is just going to cause a problem." And everybody will rationally say, "Yeah, you're right. I don't really need my money today." And everything works well.
Tkac: Does this sort of run happen outside of the banking system? I mean, what I'm thinking of is the experience with, say, money market funds in the crisis of 2008. It sounds like a run was what we saw and a credit contraction, at least to certain short-term money markets, was the result. Is that the same kind of problem that you are talking about?
Dybvig: It is, and I think that there were lots of things in the crisis that looked like runs. Ned Prescott at the Richmond Fed actually has a very interesting article he wrote on that. In the case of money market funds, we think of money market funds as normally being very liquid on both sides so that they hold very short-term assets that are easy to sell in the market and they also have depositors who can take their money out. And in any normal type of time you say, "Well, they are not doing very much liquidity transformation, it's not really a problem for them to have runs." But during the crisis some of those assets that are normally liquid in good times became illiquid, and part of the reason is that they had to have assets that were quite safe or else they would not meet the regulatory requirement. But they still skated on the edge a little bit and, in order to compete with the other money market funds, would invest in things that in normal times are very safe, but in times of crisis looked more like the bank assets. And so what happened is that when the money market funds started losing money because the... especially the bank commercial paper that they held, lost value then depositors started running on the funds and it looked just like a bank run.
Tkac: Like a bank run. So I know that in the aftermath of the crisis with all of the things that we saw, as you said looked like runs, a lot of folks have gone back to your paper with Doug Diamond about bank runs and trying to understand how these problems can be solved—so how we can have it happen less often. You mentioned there and you just did mention deposit insurance and the ability of a depository institution to go to the discount window. Are those solutions sufficient? It seems like in the crisis what we found is that, well, money market funds don't have deposit insurance, but that those particular solutions, at least, didn't stem the broader problem of liquidity transformation and the risks that it might have.
Dybvig: Well, there are several things going on. If it were just the liquidity transformation, then deposit insurance or the discount window or possibly even borrowing and lending amongst banks could help the problem. The difficulty is that there is another problem, which is that the bank assets are risky, and so there's a legitimate source of risk. Any firm can fail because its assets are risky, and this is something that's an important subject of regulations. Because the bank deposits have downside risk that's borne either by the depositors in the absence of deposit insurance or by the deposit insurance fund when there's deposit insurance in place, the people who run the bank realize they're not going to bear the downside risk. So they look at profits on the upside and perhaps will want to take too much risk. Something needs to be done to ameliorate this incentive. The same incentive exists in other types of firms and the private markets have things for minimizing that risk. Banks, in order to provide liquidity, have to be quite levered, and so the problem is especially bad for banks, but the solution is likely deposit insurance or like the capital requirements and the regulation of the banks activities look a lot like restrictive covenants that are in place in other kinds of firms for debt in the private sector.
Tkac: So following up on that discussion of solutions, several regulatory reforms since the crisis seem to have come at the risks of liquidity transformation in different ways. So we've seen reforms in money market funds to, sort of, shorten the mismatch between their liabilities and their asset holdings. We've seen liquidity coverage ratios being suggested under Basel. So I'm curious how you think about those new regulatory reforms and how they get at the problems that you're concerned with?
Dybvig: So I think the problems do reduce the liquidity risk and the exposure of the banks to this risk, and exposure more importantly of the deposit insurance funds and the governments to these risks. The danger is, and the important question is, whether this is also limiting the ability of banks to do the maturity transformation that was its contribution to society in the first place, the contribution of the banking sector. So this is actually an area where we don't know as much as we'd like to at this point. Part of it's because it's difficult to have a theoretical framework outside of various simple models for analyzing liquidity and understanding the scope of all the different costs and benefits of liquidity. That's something that's important for the empirical work. I think there has been some empirical work, but doing empirical work is very difficult now because we don't have the theoretical foundations, the framework for thinking about things.
Tkac: So it sounds like you're saying we don't really have right now, within, say, the economic profession or more broadly a way to define what even the optimal desirable amount of liquidity transformation, liquidity creation is. We know that there are costs in terms of economic activity foregone if we don't do enough. We know there are risks if, we perhaps, do too much, but we don't know where that Goldilocks sweet spot is of just right.
Dybvig: Yeah, I think that's fair. I don't think it's that far off for us to get reasonable estimates. And I've seen people who've come up with back-of-the-envelope estimates and unfortunately, they differ a lot as to whether it will cost a lot or cost a little to have less liquidity than what we have now. But I think that that's a really important feature. If the value of the liquidity creation is small, then we should have very narrow banks that are extremely safe and not worry too much about regulating the assets. If the liquidity creation is important, as there's at least some circumstantial evidence from the crisis, then the job of the regulators is much more difficult, but we need to understand that trade-off.
Tkac: Well, thank you.
Dybvig: Thank you very much.