Wharton’s Itamar Drechsler discusses his research on bank deposits as a form of monetary policy.

Wharton finance professor Itamar Drechsler focuses his research on asset pricing, financial intermediation, macro finance and monetary policy. His latest studies,Banking on Deposits: Maturity Transformation without Interest Rate Risk” and “The Deposits Channel of Monetary Policy,” which were co-authored with Alexi Savov and Philipp Schnabl of NYU’s Stern School of Business, examine bank deposits as a channel of monetary policy. Drechsler spoke to Knowledge at Wharton about the novel approach taken in his work and why monetary policy continues to be such a challenging area for economists.

An edited transcript of the conversation follows.

Knowledge at Wharton: In one of your papers, you propose and test a new channel of monetary policy. Can you explain that?

Itamar Drechsler: The question of how exactly monetary policy works is one of the central questions of macroeconomics. It may be a little bit surprising, given that it has been studied for so long and used in practice, that we still don’t agree on what is the channel through which monetary policy operates, let alone all the channels. The channel we suggest is one that goes through the banking system. The idea is that when the central bank raises the short rate – in the case of the United States, the Federal Reserve raises the federal funds rate — then banks react to this in a way that isn’t completely competitive. It’s got a form of a monopolistic power to it. They don’t raise the interest rate that they pay on savings deposits to people by the same amount that the central bank raises the short rate that exists in the competitive market that banks lend to each other or firms lend to banks.

In fact, you can see clearly in the data that banks raise deposit rates by only about 35% to 40% of whatever the central bank raises its rate by. That means they get to charge depositors a big spread, so they make a lot of money off of this. But that by itself wouldn’t give you a channel. What gives you a channel is that while most people are insensitive to this spread — which is why the banks can charge it — some people do pull their money out of deposits. Because deposits are still a central source of funding for the banking system, as they pull some of the money out of deposits, they shrink the amount of funding available to banks, which in turn causes banks to have to contract the amount of lending and the amount of assets that they buy. That has very large effects in terms of quantities because there are about $10 trillion in deposits. Even if only 10% of deposits move out when the spread on them rises — when the price of holding them rises so much – that has a very big impact, even for the economy as a whole.

“There’s a standard workhorse model that people believe is in effect, but it’s also clear that this model has a lot of shortcomings.”

Knowledge at Wharton: How did you test this?

Drechsler: We have a number of tests in the paper. That’s probably one of the main strengths of the paper, hopefully, besides what I think is a novel idea.

If you look at aggregate data, meaning how much deposits are in the banking system, you can see that total deposits move inversely with the fed funds rate. But there are all kinds of different deposits. Some of them are more retail-like and, therefore, more affected by this than other ones. If you look only at savings deposits from retail, then you can see the effect very strongly. The problem with this is that it’s suggestive, but it doesn’t prove that it’s caused by the channel that we have in mind. One alternative cause you would think about is maybe when the Fed raises the interest rate, there’s simply less economic activity and the banks don’t want to borrow these deposits. In this case, deposits are reduced not because banks are trying to charge people a higher spread in order to make more monopoly-like profits. They simply demand less deposits.

That’s always a very difficult problem for monetary policy studies to resolve, the question of whether an aggregate effect is pushed by changes in demand or supply. The interesting way that we solve this problem is that we looked at branches within the same bank that are located in different areas, that are more or less competitive, judging by how many other branches of banks are located in their area. Areas where there are a lot of branches of banks are likely to be more competitive than areas where there are fewer ones.

What we showed is that in areas where bank branches are more concentrated — there’s less competition — banks raise their interest rates by less. In other words, they raise the spread between their interest rate and the competitive interest rate by more. We find that this is true even across different branches of the same bank.

Looking at branches of the same bank holds the bank’s overall demand for deposits constant, because the bank can move deposits from one branch to another. We show clearly that in areas that are less competitive, banks increase the interest spread they charge by more and then more deposits flow out of those branches, than at branches of that same bank located in more competitive areas. That shows us convincingly that it’s competitiveness that is driving the effect, not the bank’s overall demand for deposit funding. Ultimately, what we care about is putting all banks together to look at the aggregate impact of this channel. Nevertheless, to really tease out whether what is happening is what we’re suggesting, one has to do more micro analysis like this. That was a very clear, effective exercise because it’s really hard to think of any other reason why branches of the same bank would raise deposit rates by different amounts.

Knowledge at Wharton: Why is it important for players in the area of monetary policy to understand this?

Drechsler: By changing the interest rate, the central bank appears to have a very big impact on the economy. The really interesting, or frightening, thing about monetary policy, depending on your point of view, is that we don’t really understand how it works. Some people probably feel they understand, but there’s no very clear consensus. There’s a standard workhorse model that people believe is in effect, but it’s also clear that this model has a lot of shortcomings. The channel of this model doesn’t really operate through the financial system. It mostly ignores the financial system and doesn’t predict that monetary policy has much direct effect on the holdings of financial institutions. In contrast, it seems clear that financial markets really believe that the central bank’s actions have an important effect on the whole economy, and on the financial system in particular. So, it seems very important to understand how monetary policy works. Why does it have effects? What do those effects depend on? Are changes taking place that will make this policy more or less effective in the future? How does it affect things like investment in real estate or securities or how much risk people take on? It’s understanding the basic mechanics of how the economy works.

“The kinds of spreads you see banks charging depositors are as large as the spreads you see anywhere else in financial markets — maybe bigger.”

Knowledge at Wharton: Why is it important for consumers? Are there implications for them?

Drechsler: There are household finance implications, but that’s not our main focus. The clear household finance implication is that, in our opinion, while banks provide very safe deposits for people because they’re backed by government insurance, at the same time this gives them the power to charge households very large spreads. The kinds of spreads you see banks charging depositors are as large as the spreads you see anywhere else in financial markets — maybe bigger. If the short-term interest rate was raised to, say, 4%, banks would only pay about 2% on deposits. There are $10 trillion of deposits, and you have 2% per year. That’s a tremendous amount of money. That’s $200 billion a year that depositors are paying in spreads.

Knowledge at Wharton: In another paper, you examine deposits through the lens of banks’ exposure to interest rate risk. Can you explain that?

Drechsler: That comes out of thinking about the paper I just described: If the short-term interest rate has such a strong effect on what banks do in their business, on how many deposits they have versus the ones that flow out and on the rates that they charge, then it comes back to a fundamental question of what is their interest rate risk?

The textbook model of banks that you read about or students get taught is that everybody knows that banks borrow short term, mostly from depositors, and they make longer-term loans. They make loans to households through mortgages. They make loans to firms that are not so short term. The standard view is that this is fundamental to banking. It’s called “maturity transformation,” and that by doing it banks earn the difference between the rate they charge for long-term loans and the rates they pay on short-term loans. But this introduces a basic risk, for example, one that came up in the savings and loan crisis in the 1980s and early 1990s. If rates rise too fast, then the rates banks have to pay on deposits will become higher than the rate they locked in on their long-term loans, and that this could put the banking system into trouble.

Some people believe that this is really fundamental to every banking crisis. What we showed is that, in fact, banks have very little interest rate risk. While it is true that banks do have long-term loans and short-term borrowing, deposits actually work like long-term financing because banks don’t have to change the rate they pay on them all that much when interest rates rise. If deposits were truly long term, then the rate would be completely fixed for a long time. In fact, deposits rates are not actually fixed, but in practice banks don’t have to change them that much. So in effect they end up working like long-term financing, and hence banks are protected from interest rate risk, despite the how their balance sheet looks.

That explains why, even though interest rates have changed tremendously over time, we haven’t seen episodes where interest rate risk per se induced a crisis in commercial banks. Interest rate fluctuations did induce the crisis in savings and loans, but that is the exception that explains the rule because it occurred right after savings and loans were forced to start paying more competitive rates for their funding. If banks didn’t have deposits to shield them from interest rate risk, we would have seen that kind of thing all the time.

Knowledge at Wharton: What does this say about the health of banks in the economy?

Drechsler: There are a lot of people that have posited that one of the main channels through which central bank policy works is by affecting this interest rate risk of banks. We’re saying that that’s not the case. We think that so long as the fundamentals of deposit banking don’t change, this situation will remain similar. But they very well might, now with electronic payments and different technological developments in how people do banking. It could, but holding that constant for a second, if the deposit franchise of banks doesn’t change, then interest rate risk is not the risk that they’re exposed to.

In contrast, they have been exposed to credit risk, such as in the mortgage crisis. Some people think about it as maybe liquidity risk, but it’s certainly not interest rate risk. We think much of the work that’s been done on that is sort of misguided. Also, there’s a lot of discussion of forcing banks to only make short-term loans, which is called narrow banking. There were a lot of propositions like this after the financial crisis because they were concerned that long-term loans and short-term financing are what caused the crisis. We’re saying that’s not it. It was probably the credit risk that was associated with mortgages.  Making banks be narrow in that sense actually could be counterproductive because their deposits function like long-term funding. Making them hold short-term things would actually create a mismatch where there was none to begin with.

Knowledge at Wharton: Drawing on your previous answer, what do you think will be the impact of introducing all of these new banking products into this industry?

Drechsler: The straightforward answer would be that they lose their ability to charge these high spreads on deposits, which would fundamentally impact their ability to make long-term loans for the reasons I just described and potentially could change the banking system tremendously. Of course, they don’t like that and are trying to sort of co-opt this system.

But we’ve seen developments before that suggested that this could happen. For example, in the 1970s, money market mutual funds were developed to be able to borrow from people and pay them higher deposit rates. While that’s a large industry, it still has not eliminated banks’ deposit franchise. Over this time, we’ve seen very healthy growth in banks’ deposit franchise. So, both of these parts of the system are big. It didn’t get rid of the banks.

Then maybe 15 to 20 years ago, we had internet banking. For example, I have an E*Trade account. By getting rid of branches, they also could have paid people rates that are completely competitive. Instead, they act much like regular banks, maybe slightly more competitively. The demise of the deposit franchise has been predicted before and hasn’t come to fruition, so I hesitate to predict what would happen this time. There’s just something about the bank branches and also large banks that people think are safer, which will make it quite difficult to change the situation. Whether it’s desirable or not is a separate question.

Knowledge at Wharton: Do you think that’s the case with automation and introducing machine learning and AI into the mix? Because I think a lot of banks would almost prefer that you don’t talk to a human or go to a branch.

“The demise of the deposit franchise has been predicted before and hasn’t come to fruition, so I hesitate to predict what would happen this time.”

Drechsler: All these things are operational things that could help them to make their system more efficient. For the first time in a long time, they’ve stopped increasing the number of branches. But the question is whether people find value and pay for the deposit franchise of banks, whether it’s branches or the other services the bank provides, the name of the bank, the marketing of it, the government guarantee. I have to say that I don’t know what it is among all these things. I can just see the effects that are pretty clear.

In some sense, all the tools are out there for people to move their deposits in a way that will keep things competitive. But I think for most people, it’s just not worth it. There’s a lot of paperwork that needs to be done because you’re supposed to be protected from potential issues with your savings accounts. It’s the most fundamental thing people have. And that is a double-edged sword. It keeps you safe and allows for the government guarantee. On the other hand, it means that you don’t just move your deposits every day to whichever place offers the highest rates. The result is that the market settles into a kind of equilibrium where nobody pays particularly high rates at all.

Knowledge at Wharton: What are some other future lines for your research?

Drechsler: I also do research in completely different things. I’m very interested in understanding differences in the kind of returns people make across different stocks. This is a very active part of financial research that people call anomalies, which is where you can predictably make higher average returns on some stocks — for example, value stocks relative to growth stocks, or stocks that have had recent high returns, or stocks that have a lot of distress risks seem to have particularly bad returns, even though people know they’re distressed.

My line of research there is trying to understand a portion of the market that’s been hard to study, which is one where people borrow these shares in order to short-sell them, meaning to bet against them. It’s been hard to get data on how hard it is to borrow these things. You typically have to pay a rate for them. For most stocks, it’s very low, but there’s a nice segment of the stock population where you actually have to pay high rates.

Since we’ve gotten this data, we’ve seen that these are the ones that generate particularly bad returns. The theory is that those are the ones that sophisticated investors are paying money to borrow in order to short, while apparently unsophisticated people insist on continuing to hold these things, despite indications that other more sophisticated people feel that their price is too high.

We’ve done some work on this, and the results there are very striking because we have access to this data that before was kind of opaque and wasn’t available to people. The first line of research I describe is very focused on retail people. This one is very market-centric. It’s about the interaction of sophisticated investors with the more simplistic, naive investors.