Wharton finance professor William Diamond, who is new to the faculty this year, is diving deep into the function of banks within the financial system. In his latest research paper, Diamond explains why banks tend to play it safe by investing in assets with the lowest risk, how such behavior contributed to the subprime boom during the financial crisis, and the implications for regulators. The paper is titled, “Safety Transformation and the Structure of the Financial System.”
Knowledge at Wharton: What are some of your areas of expertise or focus?
William Diamond: The broad area I’m interested in is financial intermediation, which is studying banks or other financial institutions and the role they play in the economy. I guess what got me interested in this was the 2008 crisis, which happened when I was getting involved in economics and trying to understand crises and financial regulations related to them, but then more broadly how [financial intermediaries] relate to the rest of the economy, too.
Knowledge at Wharton: You had a recent paper about the role of intermediaries in publicly traded securities markets. Can you talk about that paper and some of the key takeaways?
Diamond: If we look at older models of what something like a bank would do, people often emphasize the expertise they have or the information they have when making loans. If a bank makes a loan to a company, they’re going to spend some time monitoring that company and learning how creditworthy it is. For a lot of the banking system, this makes a lot of sense. But if you look right now — particularly at the modern U.S. banking system — there are highly developed countries with deep financial markets. You see a lot of publicly traded assets on these balance sheets — mortgage-backed securities, government debt, things that are not so difficult to sell. The models we have of why banks exist in the first place are somewhat in tension with this idea. One of the goals I had in this paper was to understand how it could be the case that I can log on to my brokerage account and buy exactly the same assets that these banks are holding, but they play some important role in the economy.
Knowledge at Wharton: What are the practical implications for this research?
Diamond: My model has some implications for how to think about a bank’s risk management policy. In this paper, the real goal of a bank is they create safe assets. When you put your money in the bank, you want to know it’s there. If you’re going to make a transaction and swipe your card, you want to be sure that money is there. If you had something like a risky stock portfolio and went to pay at the store with this, maybe the stock market crashed in the meantime. So, the goal of a bank is to have these very, very safe deposits.
What a bank needs to do then is invest in assets that ensure these deposits actually are safe. This is the theory I have of how a bank chooses its portfolio. They invest in mortgages. They make loans to companies. Almost everything on the asset side is debt of various sorts, whereas they could very well play the stock market. Why don’t we see banks investing in speculative tech startups? The real purpose of this is to make sure that these deposits are very safe in the first place. If banks create lots of safe deposits, that’s a really cheap way of funding themselves. You put your money in the bank not because it’s a really high rate of return, but because you need it to make transactions.
From the banker’s perspective, if they can satisfy your needs for a stable, safe medium of exchange, they can get really attractive rates of funding. They create something very safe, and in order to do that they have to invest in the lowest-risk asset portfolio they possibly can. This [paper] would ask a banker, how do you think about the cost of investing in different assets?
Knowledge at Wharton: The paper points to how a growing demand for safe assets may have contributed to the subprime boom. Are there implications for regulators who want to make sure that doesn’t happen again?
Diamond: The main practical implication of the paper is more on the regulatory side. First I build a model of why do banks exist, assuming the private sector is doing their job correctly, and then use this as a framework for thinking about various government policies toward regulating banks’ risk-taking or quantitative easing, which was something the Fed did recently. It’s a way of thinking about the whole financial system, both banks’ balance sheets and how much the nonfinancial sector borrows.
In the context of the subprime boom, effectively what a lot of people in macroeconomics have argued is that the rise of China, which is a country that is growing very fast but has a weak financial system, led [to an increase in] the global demand for safe assets in general. If you look at U.S. government debt, huge portions of that were bought by the Chinese Central Bank throughout the 2000s as they were growing because they couldn’t create something as safe as U.S. government debt themselves with their poor financial system.
“If banks create lots of safe deposits, that’s a really cheap way of funding themselves.”
From our perspective in the U.S., what this means is that safe borrowing rates were pushed down. You can see that interest rates were falling in this period. In the context of my model, this means that the banks’ business model of creating safe assets has just now become a whole lot more lucrative. If you can borrow very cheaply by creating something very safe like a deposit, you’re going to do more of that than when the price of safety was lower. The Chinese bid up the price of safety, and that’s an increase in demand. To shift along the supply curve, we now have a larger and riskier financial sector providing these safe assets. This connects to the subprime boom because a bank’s goal is to create something safe, and to do that they choose a portfolio that is as low-risk as possible.
If they’re going to expand their portfolio and they’re investing in something low-risk, mechanically they must be buying riskier assets than they did before. Before, they would make loans to only very creditworthy households. Once they’ve exhausted that market, you have to move into the subprime sector or just give everybody larger loans, and collectively people will be riskier then, too.
Knowledge at Wharton: What are you studying next?
Diamond: In this research agenda, I think the next step is two different things. One is understanding where this demand for money comes from in the first place. If people need something safe to perform transactions, what exactly is the problem that a safe asset solves when people go to the store and buy something?
I’m working on one paper where first I have a model of transaction restrictions that money solves. Then I’m trying to attach to it something related to my original paper, which is how does a bank then create assets that meet this demand for money, and put that together in a deeper framework. This is needed if you want to think about assets that are more or less like money, and how safe an asset needs to be in order to be useful for making an exchange. If you put your money in a savings account, that’s effectively riskless. The government backs that up. If you put it in a money market fund, that’s not backed by the government, but you can still write checks on it. There’s some intermediate degree of riskiness where things can still be somewhat like money.
“You put your money in the bank not because it’s a really high rate of return, but because you need it to make transactions.”
A second related thing I’m working on is thinking about how other financial institutions can be very similar. A bank invests in a portfolio of very low-risk assets, like mortgages, in order to create a lot of deposits. But also you can think about insurance companies doing something quite similar. A life insurance company will promise that some time when a person dies in the next 30, 50 years or so, a fixed quantity of money will be paid out. You can think of that as a very long-dated safe asset, where the creditworthiness of this insurance company means that you trust at some time in the far future the money is going to be there when your family is in need. A life insurance company has a very similar risk management problem to a bank in that they need to hold an asset portfolio that makes those promises credible.
I’m working on a second paper that tries to compare what sort of assets should a bank hold, what sort of assets should an insurance company hold, and can we compare what we see in the data about what their investments look like and try to square that with a theory of the fundamental problems that they solved in the first place.
