In his recent research, Wharton finance professor Michael Schwert uses a novel data set to examine the pricing of bank loans relative to capital market debt. His findings point to firms’ willingness to pay for the unique qualities of bank loans and raise questions about the nature of competition in the loan market. Schwert recently spoke with Knowledge at Wharton about his paper, “Does Borrowing from Banks Cost More Than Borrowing from the Market?”
An edited transcript of the conversation follows.
Knowledge at Wharton: This study looks at the interaction between public and private markets, but what’s new about this paper?
Michael Schwert: This paper studies the relative pricing of private and public debt, which surprisingly has not been studied directly in the academic literature before. There is a lot of research out there on firms’ decisions whether to borrow from a bank or to issue bonds, and there’s plenty of research out there on the pricing of bonds or the pricing of loans, but this is the first paper that has actually made a direct comparison between the pricing of the two types of debt.
Knowledge at Wharton: What are the key findings?
Schwert: The main challenge is to establish a clean comparison between the two types of debt…. If you were to compare two different types of firms, then you would have these unobservable risks that would be hard to compare between the two. Even if you compare the loans and the bonds from the same firm — because some firms use both — then you’re dealing with different levels of seniority. Loans tend to be senior to bonds, so they’re safer from a payoff standpoint for investors than bonds are, and they should have a lower cost of capital.
I use a sample of firms that use both types of debt, so I mitigate that first problem of comparing across firms that use different types, and I use a structural model of credit risk to adjust the seniority of the bond so I can establish a capital market comparison point for the pricing of the loan. What I find is that loans appear to be a very expensive source of financing for firms relative to public debt issued in the capital markets. Sixty percent of the average credit spread on a loan is credit risk that’s priced by the capital markets, and 40% is a premium that’s earned by the loan syndicate above the price of credit risk, as in the capital markets.
Knowledge at Wharton: What does this paper say about firms’ views on the relative value of a bank loan versus issuing bonds?
Schwert: One implication of this is that firms must place a high value on other services that the bank offers, such as flexibility in terms of pre-payment or renegotiation of the loan and information production that can spill over into the pricing of other assets. One function that banks serve is to screen and monitor borrowers, which provides a signal to bond market participants as well as equity market participants on the quality of the firm from privately observed signals that the bank has, but the market does not have.
Knowledge at Wharton: What questions do the findings raise about the nature of competition in this market?
Schwert: To interpret the finding that bank debt is more expensive than capital market debt, there are two possibilities. One is that the banking market is perfectly competitive, so the cost of providing these benefits that I just mentioned are passed through directly to borrowers. And that’s what this premium is reflecting.
The other possibility is that the banking market is imperfectly competitive, which means that banks are earning rents lending to firms, and they’re not competing aggressively with each other to undercut prices until they’re at the competitive level.
“Loans appear to be a very expensive source of financing for firms relative to public debt issued in the capital markets.”
I provide some evidence in the paper that the loan spread premium that I find is not correlated with variables such as the size of the loan or the size of the firm, which casts doubt on the perfectly competitive side of possibilities because there are some fixed components to the cost of providing these benefits that I mentioned before. If these costs were fixed, then we would expect the premium to be small for larger loans.
I find it to be similar for larger loans in terms of the credit spread. This premium is 100 times larger for a $10 billion loan than it is for a $100 million loan, which casts doubt on this perfectly competitive sort of explanation. But future work is necessary to establish whether there is imperfect competition in the syndicated loan market. That’s beyond the scope of this paper.
Knowledge at Wharton: What is the message here for firms?
Schwert: The message for firms is a nuanced one, and the reason for that is that if you look at the interest expense on each of these types of debt, it’s actually very similar for a healthy firm. The credit spreads on loans and bonds are very similar for firms that are rated BB or above. This is not inconsistent with what I just told you about loans being more expensive per unit of credit risk than bonds because the loans are senior to the bonds in these circumstances, so they’re actually safer and should earn a lower credit spread. At the beginning of the paper, I established a simple benchmark, which is that the loan spread should be one-third of the bond spread on average. So, the fact that the loan spread and the bond spread are similar does indicate that the loans are more expensive on a credit-risk basis. But at the same time, the fact that they’re similar means that for the shareholders of a firm or the manager making a decision on which source to borrow from, they actually have similar costs.
Knowledge at Wharton: What are some future lines of inquiry for this research?
Schwert: The next project I’m exploring is going to dig more into this question of whether the loan market is competitive or imperfectly competitive. I have a co-author at Columbia University, Olivier Darmouni, and a co-author at the Federal Reserve Board, Stephan Luck. We’re going to use regulatory data on all loans of more than $1 million that the top 30 banks have made — or all of the banks that are subject to stress testing under the Dodd-Frank Act.
The benefit of these data is that they will allow us to use local variation in competition because a $1 million loan, while it sounds large, is really a small business loan, so it would be made by a single bank that would be located in the same area as the firm.
Across the country, there’s a great deal of geographic variation where some areas are very competitive banking markets and some are much less competitive. We plan to look at whether credit market conditions are passed through differently when they deteriorate or improve to firms that are located in these different areas, which should shed some light on whether loan pricing is reflecting competitive or noncompetitive outcomes.