In the wake of the global financial crisis in 2008, Americans stopped buying new cars.

Not completely, of course: annual new car sales decreased by an order of magnitude, from about 16 million in 2005 to less than 10 million. But this drop in sales—coinciding with General Motors and Chrysler filing for bankruptcy and receiving bailouts from the U.S. government—was sharp enough to draw the attention of economists like Efraim Benmelech of the Kellogg School of Management.

Sales of durable goods like home appliances, personal electronics, and automobiles tend to decline during a recession for intuitively obvious reasons: “Many households lost money and equity so they couldn’t afford to buy new cars,” Benmelech says. But he wondered if mere belt-tightening by consumers was enough to explain why the market for new cars contracted so dramatically in the immediate aftermath of the economic crisis.

In research conducted with Ralf R. Meisenzahl and Rodney Ramcharan of the U.S. Federal Reserve Board, Benmelech identifies an additional cause of the contraction in car sales: a sudden freeze in the credit markets that normally allowed auto manufacturers to offer financing loans and leases to their customers through their network of dealers. When this “shadow banking system”—which included GMAC, the lending arm of General Motors—seized up, automakers found themselves in a peculiar bind.

“Some people actually did want or need to buy new cars,” Benmelech explains, “but they couldn’t because they didn’t have enough cash—and they couldn’t get a loan or a lease agreement either, because the financing institutions that would normally make these agreements were also constrained in the crisis.”

The federal government bailed out these financing institutions, as well. Though the bailouts were politically unpopular, the research suggests that they were necessary for restoring the health of the credit market—and the broader economy.

The Importance of “Captive Lessors”

Benmelech and his coauthors based their conclusions on an analysis of a detailed dataset from the automotive data company Polk. The dataset is composed of geographical and financial information from every new car sale in the U.S.—including the ZIP code where the sale took place and whether the sale was financed by a loan from a so-called “captive lessor” (the technical term for in-house lending agencies like GMAC).

“In this market, the major financial players are not banks— they’re these captive lessors,” Benmelech says. “They’re like mini-banks that belong to the automakers. We usually don’t hear much about them, but they provide about half of all the credit for new car sales in the U.S.”

Captive lessors lack a crucial backstop against financial stress that real banks can rely on in a crisis: demand deposits. “Banks can always raise funds this way,” says Benmelech. “If investors are pulling money out of mutual funds and the stock market, they have to park the money somewhere, and instead of putting it in their mattress, they deposit it in the bank.” According to Benmelech, the level of deposits may actually increase during economic recessions. But without this fail-safe to rely on, the captive lessors of struggling automakers found themselves increasingly constrained.

It did not help that the credit market that captive auto lessors relied on to make customer loans was hit especially hard by the effects of the financial crisis. That market is known as the “asset backed commercial paper” market, comprising very short-term bonds backed by bundles of car loans and lease agreements. “You had these nonbank financial institutions making fewer loans since they didn’t have the capacity to raise enough money,” Benmelech explains. “They were hit even worse than the average bank during the crisis.”

“The notion that you can have a permanently clog-free pipe—a crisis-free system—is unrealistic and wrong.”

Bailing Out the Lessors

Benmelech and his colleagues hypothesized that an inability to offer financing to potential customers was partially responsible for the deep drop in auto sales. And the Polk data on car sales, combined with credit-score information from Equifax, suggests that this was indeed the case.

“The ZIP codes that depended heavily on credit by captive lessors are exactly those that saw sharper declines in car sales, compared to areas where banks provided more credit,” says Benmelech. “The bottom line is that we can attribute about a third of the decline in new car sales in 2008 and 2009 to the financial difficulties of these captive lessors.”

In addition to bailing out GM and Chrysler to the tune of more than $63 billion, the U.S. government disbursed approximately $17 billion to their captive lessors—a move specifically intended to ease their unique “financial difficulties.” These bailouts were politically unpopular, but were they necessary? Benmelech believes they were.

“As much as the public would like to point fingers, our findings actually show why it was important to inject money into these captive lessors,” Benmelech says. “To put it bluntly, you can think of credit markets as plumbing. When credit doesn’t flow, it turns out we have a clogged pipe. You might get upset when your pipe gets clogged, but you’d like to know which pipe has the clog—otherwise you can’t clean it out. We identified the clogged pipe in the plumbing of credit markets for automobiles, and it’s the credit channel of captive lessors.”

Because of the unique importance of the automobile industry to the wider U.S. economy, the bailouts were “a smart move,” Benmelech adds. “After the purchase of a house, a car is the next biggest purchase a household usually makes. And it’s very hard to buy a new car off a credit card.” Still, auto sales in the United States didn’t rebound to precrisis levels until 2014.

Takeaways for the Next Credit Crunch

For CEOs of car manufacturers, the takeaway from Benmelech’s findings are clear: “Have a plan for when this market for credit goes away, because any crisis that impacts those markets is going to make things very tough on you.” But the broader implications, he says, speak to the inevitability of “clogged pipes” in our economy, and how agencies like the U.S. Treasury and the Federal Reserve can effectively respond to them.

“The notion that you can have a permanently clog-free pipe—a crisis-free system—is unrealistic and wrong,” Benmelech asserts. “We could prevent credit markets from clogging by never throwing anything into the sink, but we have to use that sink! What we need to do instead is be prepared to assess what we are willing to do when it does get clogged.” Giving billion-dollar bailouts to stricken automotive lending agencies may look like throwing good money after bad. “But we can tell you: this is what unclogged the pipe,” he says. “Without that money being provided to GMAC, things could have been even worse.”