The 2022 Nobel Prize in Economic Sciences has been awarded to former Fed Chair Ben Bernanke of the Brookings Institution, Douglas Diamond at Chicago and Philip Dybvig at Washington University in St. Louis “for their research on banks and financial crises”. Anil Kashyap explains the rationale for this recognition.
This year’s Nobel Prize in Economic Sciences should be thought of as recognizing the research that informs our thinking on three questions about banks: i) Among the many things that banks do, what are the essential features? ii) Why are banks the institutions that perform those functions? iii) What happens when things go wrong in the banking sector?
The answers lie in three papers cited by the Nobel committee that all were published in the early 1980s: two theoretical contributions by Diamond and Dybvig (1983) and Diamond (1984); and an empirical contribution by Bernanke (1983).
This research shows that the essence of banks is to take deposits that convey services (for example, checking accounts) from savers and use the funds to make loans to households and businesses. In other words, both bank assets (the loans) and liabilities (the deposits) are valuable and you can’t understand banking without recognizing both functions.
This may sound trivial, as deep insights often do, but it is hardly obvious. For instance, after periods of financial instability, one often hears commentators ask why do we let the banks gamble with savers’ money? If we just forced them to be much, much safer (say, by investing only in government securities), they would not cause so much trouble. The laureates’ work explains why the lending and deposit-taking naturally co-exist.
The theory starts from Diamond and Dybvig’s research. They ask why any firm would usefully offer demandable deposits? The alternative is that individuals could themselves just buy safe short-term securities, which could be bought and sold as needed. They point out that in most cases people will not have an immediate need for making payments, so it is expensive (in an opportunity cost sense) to have everyone self-insure and save using only the most liquid savings vehicle.
If we recognize that most savers won’t need their savings at any given time, they could invest the savings in higher returning assets. Of course, if they do need to sell such assets, that will be expensive since the reason that they earn higher returns is because they are illiquid.
The resolution to this tension is to create an organization (a bank) that pools the needs of many people and counts on the idea that not everyone will need their savings at the same time. The bank essentially becomes an insurance vehicle, where it mostly invests in the illiquid assets and sells some of those assets when necessary to meet the withdrawal needs of the people who truly have an urgent requirement for funds. This allows the bank to pay a higher return on the checking accounts in exchange for paying a little less on the longer-term savings accounts. Because none of us know when we might exactly need to make a payment, this is an attractive service.
Unfortunately, the arrangement comes with one obvious defect. The insurance breaks down if everyone decides that they want their money back. The bank has no way of separating who really needs to withdraw from who might be doing so because of panic. I may go to the bank because I see a line there, not because I need my money. As Diamond often says, fear of fear itself is a problem. There are various ways to solve that, deposit insurance and liquidity regulation being obvious candidates.
That raises the next question of why then make the investments in loans? This is the question that Diamond (1984) addresses. He points out one problem that anyone making a loan has to deal with is what to do about the risk that a borrower cannot fully repay the loan. The sensible thing to do in that case is renegotiate and get whatever is available.
Unfortunately, if the borrower knows that is an option, then borrowers will always claim that they cannot repay and seek to get a reduced amount. The best way to deal with that is for the lender to monitor (audit) the borrower to learn about repayment prospects. For any individual saver, this monitoring would be very expensive: most won’t have the expertise to do it. Furthermore, the interest each saver would need to charge to cover this cost would be very high.
This suggests that pooling funds can again be attractive. Suppose many savers turn their money over to a specialist, the bank, and that specialist will make many loans and monitor the borrowers. This seems attractive, but runs into the problem of “who monitors the monitor”?
The savers who have turned their money over to the financial institution to invest now face the same problem that was present if they each wanted to make a loan individually to a final borrower. The bank can say it can’t repay the deposits because the borrower defaulted. How do they know if the financial institution is telling the truth about what it learned when it monitored the borrower?
Thus, it may appear that the agents will each have to monitor the financial institution, which means that they need the bank to pay them enough interest to cover this cost. It seems pooling and delegating is not helpful.
The central insight in Diamond (1984) is that the “who monitors the monitor” problem can be overcome by having the financial institution make many loans. If the financial institution has a well-diversified set of loans to many different borrowers, then it is unlikely that all the loans will fail at once. It will not be credible for the financial institution to lie and argue that its borrowers have not paid, and they will not have to monitor to know this.
In that case, the financial institution will repay its depositors by using the proceeds from the successful investments. By spreading the risk across many loans, the cost of borrowing in the economy falls (relative to the direct lending case) and the duplication of monitoring by the depositors is avoided.
Thus, between these two papers, we see that it makes sense for banks both to take deposits and to make loans. This raises the final question of whether this arrangement creates problems for the economy. Bernanke (1983) is one of many papers showing that when banks fail, the overall economy suffers. What is the source of the spillovers?
The seminal work of Friedman and Schwartz (1963) suggested that the problems stemmed from the losses for savers. In studying the Great Depression, they argued that when banks collapsed, the savers lost access to their deposits. Without those deposits, people and businesses could not make payments and without payments, the economy could not function.
Of course, this is not the only explanation. When bank deposits fall, so do loans. Hence, we can’t tell whether it is the lending side or the deposit side implosion that is critical. (In reality, it is probably some of both).
Bernanke (1983) argues that there was a clear role for worrying about the lending contraction on its own. He noted that both theoretical and empirical work suggested that it was not just the collapse of liabilities that mattered, but also the loss of loans that had an independent effect.
For instance, during the Great Depression, large firms that had non-bank sources of credit fared better than smaller firms and farmers, who were especially dependent on banks. He also pointed to contemporaneous survey evidence showing that firms reported pressure to repay loans and the unavailability of new bank credit as a source of stress in the economy.
Bernanke’s work triggered a wave of subsequent work, much of it by him, aimed at establishing the importance of shocks to bank lending (and credit supply more generally) as a major source of instability for the macroeconomy. As the prize citation noted, this attention to preserving lending capacity and credit availability was a central consideration that motivated economic policy during the global financial crisis of 2008.
Likewise, much of the regulatory response to the global financial crisis has been informed by Diamond and Diamond and Dybvig’s research. For example, the idea that demandable claims that are offered by money market funds (and other non-banks) are prone to runs has been widely embraced as risk that needs to be addressed.
I hope this brief tour gives a sense of the significance of this work. It has had a profound effect on the direction of research and our understanding of the role of banks in the economy. Perhaps equally important, it has significantly altered economic policy and the practical way in which the financial system is regulated. For both reasons, this research is well deserving of the Nobel Prize.
Anil Kashyap
October 2022
References
Bernanke, Ben S (1983), “Non-monetary Effects of the Financial Crisis in the Propagation of the Great Depression”, American Economic Review 73(3): 257-76.
Diamond, Douglas W (1984), “Financial Intermediation and Delegated Monitoring”, Review of Economic Studies 51(3): 393-414.
Diamond, Douglas W, and Philip H Dybvig (1983), “Bank Runs, Deposit Insurance and Liquidity”, Journal of Political Economy 91(3): 401-19.
Friedman, Milton, and Anna J Schwartz (1963), A Monetary History of the United States, 1867-1960, Princeton University Press.
