Martin Schneider

Professor of Economics, Stanford University

Paper: "Banks' Risk Exposures" (with Juliane Begenau and Monika Piazzesi)

Andolfatto

What is the question that you're pursuing in this body of work?

The paper's main takeaways,
according to Schneider:

  • The old, traditional view of the bank is that it borrows short term and lends long term.
  • One might think the modern bank works differently—the modern bank that's now a lot into security markets—but, in fact, it seems from our findings that it's even more that way: borrow short, lend long.

Schneider

Broadly, this is about how the value of banks responds to market conditions. What regulators typically require banks to provide is accounting information from which it is difficult to get an idea of what will happen when, say, interest rates change to the value of the bank's portfolio. What the paper tries to do is to take this accounting information and transform it into a measure where we can easily study the changes in value when market conditions move.

Andolfatto

When you're transforming this information from the accounting statements of the bank, are you applying some sort of theory to it? Or is this just purely an empirical exercise?

Schneider

I would say it's an empirical exercise with minimal structure. What we use, you can think of it as like translating the statements from one language, the one of accounting into the one of economics.

Andolfatto

So you're serving as a translator. You're translating accounting information into a language that regulators and economists can understand. For what purpose?

Schneider

So I want to get an answer to the question. Suppose there is a change in interest rates. For example, this is quite relevant now where we have very low interest rates, and many people expect perhaps an increase in the near future. So one wonders the uncertainty about that, how is that reflected in the uncertainty that banks face?

Andolfatto

So, generally, surprise changes in the interest rate movements have effects on asset prices and on the portfolios that banks have, and therefore, these surprise movements in interest rates may affect the soundness of the banks' financial positions and increase their risk exposure and risk in some manner?

Schneider

Yes.

Andolfatto

And so what do you find exactly?

Schneider

The main part of the paper is kind of the methodology for how to do this translation. A lot of work went into especially looking at positions that are hard to translate, in particular derivatives books which are large for major banks, but for which the accounting information is not very informative.

Andolfatto

What is a derivatives book exactly?

Schneider

A derivative in general is a security where the payoffs are dependent on the value of some other security. The most important example is an interest rate swap where you and I make a contract where we commit to make interest rate payments to each other where one side, say I, make always the same interest rate payment, whereas you pay a floating rate that we fix to be some market rate.

For example, let's say I'm a bank who has a lot of long-term assets, and I'm financing them with short-term debt. Then my profit margin will depend on what the short-term interest rates are. Suppose I wanted to get rid of this risk. Then I might want to make this contract with you where you give me fixed payments in exchange for the floating.

Andolfatto

So the bank might want to buy insurance against some bad event happening and these derivatives contracts are insurance contracts?

Schneider

There are two ways that derivatives as any financial instrument can be used. One is to somehow provide insurance or allocate risk differently. And the other is, of course, suppose you and I disagree on what the direction of interest rate is in the future. Then we might take just different sides of a bet, and we might do that because we each think we're smarter than the other. So then one interesting question for looking at this huge volume in derivatives that we see now: Is this basically a big insurance scheme, or is it a big casino where people just trade on market conditions?

Andolfatto

Is that a question that your paper is addressing?

Schneider

We think that the tools that we provide here can eventually be used to look at all the players that would be needed to answer this question. Right now, we provide results only for the largest banks. So the largest banks, they have their traditional business as well as important business making the market in derivatives. And we find that the derivatives position that they have is not an insurance against fluctuations in the other business, at least not most of the time. Now, that does not mean that, overall, the economy as a whole is not an insurance scheme because, of course, the banks might be the people who insure others. But further research will tell us.

Andolfatto

Suppose we discovered, in fact, that the vast bulk of this activity was strictly for speculative purposes and had very little to do with insurance. Does that necessarily have some bearing on the design of regulation?

Schneider

So it depends. There are models in which when people disagree they can contract with each other, so let them do it. As long as markets work well, that is fine unless we somehow think that people shouldn't be following their beliefs. However, there is also a class of models that has gained a lot of attention now in the financial crisis where when you have market frictions and then you have people differ in beliefs, then you get inefficient outcomes. And so for those cases, if we were to find that a lot of this derivatives activity comes from heterogeneity in views, people betting, then that might mean that regulation should be increased in order to alleviate the problems.

Andolfatto

I guess another way of putting it is: If this speculative behavior was isolated to this group of banks or whatever, then fine. Who cares really? It's just people playing in a casino. But banks, of course, play a major role in the payment system of the economy. Many firms and people are linked through the banking system. They receive their wages through banks. Firms make payments that way. And so to the extent that something goes wrong in the banking sector, this might have ramifications for how activity elsewhere in the economy is conducted. But that I guess is also true whatever banks are doing. The issue is how connected they are to the rest of the economy and whether or not the system takes proper account of this interconnectedness, if you like?

Schneider

A related aspect for regulation is that we've seen recently banks getting help in, say, bailouts that occurred. And so one hypothesis that one might have is that if you have a set of players in the economy who get bailouts, then those might be willing to take more risks and perhaps insure others. So that is another pattern how sort of policy might be connected with the findings that we have here.

Andolfatto

So policy implications from your work: Do any kind of pop out right immediately? Are there some caveats that regulators or bank supervisors, some sort of lessons they can draw on the work that you have done so far? Or is this really something that is left for the future?

Schneider

So what I hope where this will go is that right now the way the regulatory framework works is that people look at positions one at a time, and they kind of assign risk weights one position at a time. Our approach combines all the positions and represents them as portfolios.

Andolfatto

By positions you mean positions of individuals' banks?

Schneider

Securities, loans, derivatives, yes. And we can also do it across institutions if we like. And so I think that expressing things in this language, which is the way that economists think about how a risk is allocated, will help get a better assessment of what the overall risk is than if we just combine these accounting measures.