Syndicated loans happen when large corporations put together funding for new projects, and a very large sum (think “a billion and a half dollars,” Petersen says) is split up among several large lending banks. The transaction details are public and transparent, the borrowers have the usual incentive to keep their interest rates minimal, and the asset being traded—debt—is abstract and theoretically frictionless.
All of which means that the market for syndicated loans should be as close to perfectly competitive as possible.
Or so Petersen thought. Then his colleague Justin Murfin of Yale University showed Petersen data indicating that for the past 25 years, the interest rates for syndicated loans tended to spike up in the winter and drop in the summer.
Petersen “very patiently explained to him that there’s no way this could be true,” he recalls.
He and Murfin speculated that perhaps firms that borrow in January are riskier or that the projects they finance are riskier. Yet the pattern survived no matter what they controlled for in the data. So after a year trying to disprove this seasonal distortion, with no success, “I said, ‘Something really interesting must be going on here,’” Petersen recalls.
Their subsequent research shows that imperfect competition is causing the market for syndicated corporate loans to self-organize into “expensive seasons” in January and February, and “cheap seasons” in May and June. The results have implications for bank regulation and competition. And they provide empirical validation of economic intuitions that people tend to negotiate “the cost of doing business” in the same fundamental ways—whether they are hiring plumbers or assembling nine-figure loans.
Friction in the Syndicated Loan System
Some markets are unavoidably seasonal.
Think, for example, about commodity markets for corn or wheat, where the asset is physical. A buyer who knows she will need more corn in the winter, when it is expensive, can try to stockpile it in the summer when it is cheap—but she may incur untenable costs of storage, not to mention the fact that the corn might rot in the meantime. Likewise, a seller with a surplus of corn in the fall faces similar costs.
Unlike physical commodities, the market for corporate loans would seem to be free from inconvenient constraints of space and time. “The people in these markets are incredibly sophisticated—if it were cheaper to borrow in May than in January, they’d all be doing it strategically and this seasonal effect would be arbitraged away,” explains Petersen.
So why don’t they?
Petersen and Murfin found that in fact, they do—but only up to a point. Their research is based on 25 years’ worth of syndicated loan activity from the Thomson Reuters Dealscan database.
“Everybody that can get out of borrowing in January and February, and move to May or June, does,” Petersen says.
Whoever is left falls into one of three borrowing categories, he explains: “either you didn’t plan, you can’t wait, or you have poor credit.”
The first and second categories do not refer to asleep-at-the-wheel CFOs. Instead, they capture what Petersen calls “unexpected” or “non-deferrable” projects—a common occurrence for any business.
“If you have an unexpected opportunity to buy a firm in January, and you either do it now or you never do it, you’ll probably end up borrowing in this expensive season, even though you don’t really want to,” Petersen explains.
It is the same scenario as a homeowner who has a pipe burst on Sunday afternoon: it is more expensive to hire a plumber on the weekend, but the need is unexpected and non-deferrable—so you pay the premium.
However, while we know why Sunday is an expensive day to hire a plumber, it is not clear why January is the expensive season for these loans and May is the cheap one, Petersen says. But it does make sense that they are six months apart because that makes it more costly for borrowers to wait for the cheap season to arrive.
The final category of corporate borrowers—those with poor credit—has a subtler but equally powerful incentive to act when loans are expensive. Money and debt may be just abstractions, but they do have a “cost of storage” just like corn or wheat: bank interest.
A firm with poor credit may be able to plan ahead and foresee in May that a project will need financing in January. It could set up a cheap syndicated loan to finance the project, then keep it in a bank until needed. “But if your firm has a junk credit quality, and you set up the syndicated loan at 10% [in May] to put in the bank at 4.9%, you’re losing huge amounts of money while it sits there,” Petersen explains.
Essentially, firms with poor credit are like farmers with too much corn: they could stockpile what they need, but it would just rot. Instead, they have to pay the premium for loans in the expensive season, because storing the money costs even more.
Fishing the Same Pond
But what about lenders? Why isn’t competition between them smoothing out the seasonal swings? Petersen and Murfin discovered that lenders are competing—just not as much as they might.
The reason is built into the structure of the syndicated loan market. Because of the loans’ size and complexity, they are serviced by a relatively small number of large banks that both compete and cooperate with each other.
Take a firm that wants to borrow $1.5 billion, Petersen explains: “The bank says, ‘We don’t have a billion and a half we can lend you—we wouldn’t be diversified. So we’re going to call a bunch of other banks and slice the loan up into smaller pieces.’” One bank leads this process, and the rest—called the syndicate—help fund the loan.
Firms that must borrow in January and February have to take what they can get regarding interest rates. “From a bank’s perspective, those are awesome customers,” Petersen says. “Everybody wants to lend to them.”
Normally, competition between the banks for these customers would drive interest rates down. However, the fact that the syndicated loan market requires banks to regularly cooperate dampens this natural competition.
Petersen uses the analogy of a group of acquaintances gathering around the same pond every summer to fish. Each one wants to catch a lot of fish, but also knows that any extreme acts of competition—say, dragging a net across the whole pond in the middle of the night—would ensure he’d never be able to fish there again because his fellow anglers would kick him out.
“So instead of aggressively competing with each other individually, they all might just decide on their own to compete less aggressively in hopes of everyone catching more fish,” Petersen says.
He cautions that no outright collusion is necessary for this pattern to emerge.
All syndicated loan activity is logged in a public database, “so if anybody were to ‘cheat,’ everybody would know immediately,” he says. Instead, “each of them knows that, ‘I can just do what’s good for me, play nice, and kind of hope everybody else does, too.’ And most of the time it works.”
The catch is that if enough banks enter the market, “eventually somebody will undercut everyone else,” Petersen adds. “That’s why this game works better with a small number of players that keep playing together, versus a large number of players who don’t. We’re all sitting there at the pond fishing, but out of our peripheral vision, we are watching everybody like a hawk. It’s a unique feature of this market that makes this kind of cooperative behavior—where we’re watching each other but not actually talking to each other—more likely to occur.”
A Cure Worse than the Disease?
Petersen and Murfin refer to this seasonality as an “anomaly which less-than-perfectly competitive lenders are more than happy to preserve.”
So if the only thing preventing near-perfect competition is the fact that the market is dominated by too few banks, why not break up the banks?
One reason, Petersen says, is that breaking up big banks into smaller ones, or introducing other banks into the market would also make the loans themselves much more complex and costly to put together.
“If you need ten banks to syndicate a loan, that’s one thing. But if you have to coordinate between a hundred banks, then that becomes arduous,” Petersen explains. “The syndicated loans would just become smaller, which means less access to capital. A CFO who has to pay extra to get a loan in January may be unhappy about it. But a CFO who hears, ‘I’m sorry, we just can’t loan you the money, so you can’t sell computers’ is going to be much worse off.”
Furthermore, Petersen points out that while banks benefit from being able to charge firms more for loans in the winter, the transaction is still a net positive for both parties.
“Why would a firm need to syndicate a giant loan in January? Probably because someone just dropped a golden egg in their lap that they didn’t expect, and they need to move quickly to seize the opportunity,” Petersen says. “Now, I don’t know if they’re ‘winning’ in this situation. But they’re definitely not losing.”