Bailouts and moral hazard, from Chrysler in 1979 to Silicon Valley Bank

There is an old story about a man who stops locking his car because he has insurance. He is not a bad man. He has simply noticed that the cost of carelessness has moved to someone else. Economists have a name for what he did, and it is one of the most useful phrases in public policy: moral hazard.

Moral hazard is the tendency to take more risk when someone else bears the cost of the risk going wrong. It is not a theory about wicked people. It is a prediction about ordinary ones. Change the incentives and you change the behavior, and nothing changes incentives faster than finding out the government will cover your losses.

My claim in this post is a historical one, and it comes in three parts. First, every major bailout since 1979 has been justified as a one-time emergency. Second, each one widened the circle of who could expect rescue. Third, the cumulative lesson to lenders and managers is that size and interconnection are a kind of insurance policy, paid for by taxpayers who were never asked whether they wished to sell it. I will take the case in order, give the defenders their best argument, and then say what a rule-based alternative would look like.

1979 and 1980: Chrysler, the template

The modern story starts with a carmaker. By the late 1970s Chrysler, the third-largest American automaker, was losing money and running out of cash. Management had made poor product decisions, the economy was weak, and a surge in fuel prices had hurt demand for the large cars it sold. The company asked Washington for help.

Congress passed the Chrysler Corporation Loan Guarantee Act of 1979, which President Carter signed in January 1980. It authorized the federal government to guarantee up to 1.5 billion dollars in loans to the company, on condition that Chrysler also raised money from other sources, including concessions from its workers, its lenders and its suppliers. Chrysler's chairman, Lee Iacocca, became the public face of the effort. The company later repaid the guaranteed loans with interest, in 1983. For decades, defenders have cited it as proof that a well-designed rescue can work.

I want to be fair about it. The Chrysler deal had real discipline built in. Stakeholders gave things up. The loans were guaranteed, not given. The government took warrants, a share in the upside. And Chrysler repaid. If every bailout looked like that, the argument against them would be much weaker.

But look at what the precedent taught. Before 1979 there was an unwritten assumption that a big industrial firm could fail, and the bankruptcy code was the venue. After 1979 there was a template for asking Congress to intervene, and a story, the repayment, that made intervention sound like a bargain. The success was real. The lesson was dangerous: the next supplicant would cite Chrysler and leave out the conditions.

1984: Continental Illinois and the phrase that stuck

In May 1984, Continental Illinois National Bank and Trust Company of Chicago, then one of the largest banks in the country, suffered a run. The bank had relied heavily on funds from large depositors and other institutions, and when confidence broke, those funds fled. Federal regulators responded with a rescue. The Federal Deposit Insurance Corporation put in billions of dollars and took control of the bank, and its regulators made it clear that depositors, including those whose accounts were far above the insured limit, would be protected.

That is where a phrase entered the language. During a congressional hearing in 1984, Representative Stewart McKinney of Connecticut observed that regulators had effectively decided that some banks were too big to fail. The phrase stuck, and so did the policy. If uninsured depositors in a very large bank get protected while those in smaller banks do not, then depositors have an incentive to move their money to the large ones. The size of a bank becomes a competitive advantage in the market for funding. That is moral hazard at industrial scale.

1998: Long-Term Capital Management

In 1998, a hedge fund called Long-Term Capital Management, run by a team that included Nobel-winning economists, lost most of its capital after the Russian government defaulted on its debt and markets moved in ways its models had said were nearly impossible. The fund was so heavily leveraged and so entangled with its trading partners that regulators feared its failure could spill into the broader financial system.

The Federal Reserve Bank of New York convened the fund's major creditors, and a group of private banks put up about 3.6 billion dollars to take it over and manage an orderly unwinding. No taxpayer money went into the rescue itself. Defenders can fairly say this case was a model of private-sector resolution under official encouragement.

Yet the Fed's role was the point. The message that the creditors and the market received was that a sufficiently large and connected failure would not be allowed to run its course. Whether or not a dollar of public money was spent, a public official had organized the rescue, and everyone noticed.

2008: TARP and the widening of the circle

Then came the crisis. In September 2008 the investment bank Lehman Brothers filed for bankruptcy and was not rescued, while the government stepped in to support other firms, including the insurer AIG and the mortgage giants Fannie Mae and Freddie Mac. Within weeks, Congress passed the Emergency Economic Stabilization Act, signed on October 3, 2008. It created the Troubled Asset Relief Program, TARP, with authority to spend up to 700 billion dollars. (Congress later reduced the authority to 475 billion dollars in the Dodd-Frank Act of 2010.)

TARP was designed first to buy toxic assets, then quickly redesigned to inject capital directly into banks. Large banks received public investments, some of them regardless of whether they had asked for it. The program also reached beyond banks, to insurers, to the automakers and to a mortgage modification effort.

In 2009 the government extended the rescue to the auto industry. General Motors and Chrysler received federal loans and entered bankruptcy proceedings, Chrysler on April 30, 2009 and General Motors on June 1, 2009, emerging from restructurings in which the government became a large shareholder in GM. Here again there was a fair argument: the restructurings went through the bankruptcy process, creditors took losses and the companies shed obligations. But the bondholders and unions in some cases fared better than the ordinary rules of priority would have suggested, which is a hint that the political process, not just the law, shaped the outcome.

The government and the Congressional Budget Office reported that much of the TARP money was ultimately repaid, with the bank programs returning more than they cost, while the auto and some housing programs lost money. Read those reports directly, because the totals depend on how you count. But note how the defense works: the bailout is described by its financial result, as though the only question is whether the Treasury got its money back. The deeper question is what the rescue did to the incentives of the next generation of managers and lenders.

2023: Silicon Valley Bank

Now we reach the latest case, and it is the sharpest, because it was done in a weekend without any vote of Congress.

Silicon Valley Bank, a California bank with a large concentration of deposits from technology companies and venture-backed firms, had invested heavily in long-term securities whose value fell when interest rates rose. When depositors, most of whom held balances far above the 250,000 dollar insurance limit, rushed to withdraw their money, the bank failed. Regulators closed it on March 10, 2023, and shut down Signature Bank two days later.

On March 12, the Treasury, the Federal Reserve and the FDIC announced that they would use a "systemic risk exception" to protect all depositors at both banks, including those above the insured limit. The Fed also set up a new lending facility, the Bank Term Funding Program, to support other banks.

I want you to see what happened, in logical terms. The statutory insurance limit was 250,000 dollars. Everything above that was, by law, uninsured. Those uninsured depositors included sophisticated companies that were in a position to watch the bank's balance sheet and withdraw their money if they judged it unsafe. The market discipline the law intended, that large, informed depositors would monitor their banks, was set aside overnight. After March 2023 any large depositor can reasonably ask why he should bother to monitor a bank at all, since the 250,000 dollar limit proved in practice to be a suggestion.

The best case for bailouts

The defenders have a real argument, and I want to state it as they would. Here it is.

A financial system is not a normal industry. Banks borrow short and lend long, and their stability rests on confidence. When confidence breaks, a run on one institution can become a run on all of them, and the damage reaches people who did nothing wrong: workers whose employers cannot make payroll, families who cannot reach their savings, businesses that cannot get credit. The Great Depression was, in significant part, a series of banking panics that Washington failed to stop. A government that learned that lesson, and acts quickly to stop a panic, has saved the real economy from a disaster. You cannot let the whole city burn to teach one arsonist a lesson. The cost of a bailout, they say, is small compared with the cost of a collapse, and the right time to debate moral hazard is after the fire is out.

There is wisdom in that. A panic is real, and a government that did nothing in the face of one would be negligent. Walter Bagehot, the nineteenth-century editor of The Economist, argued in Lombard Street (1873) that a central bank should lend freely in a crisis. I accept the premise.

The answer, in numbered points

1. The defenders confuse stopping a panic with rescuing the owners. Bagehot's rule, as it is usually summarized, was to lend freely against good collateral at a penalty rate. That protects the system. It does not protect the shareholders and managers of the failed firm. A rescue that wipes out the equity, replaces the management and charges a stiff price for the help would stop the panic and preserve the incentive to be careful. What we have too often done is the reverse: protect the creditors, sometimes the managers, and call it stability.

1. The "fire is out" argument never ends. There is always another emergency in which the timing is wrong for a debate. Chrysler in 1979, Continental in 1984, the crisis of 2008, the weekend of March 2023, each time the answer was that there was no time to think about incentives. If the debate is never held, the rules are made by the crises themselves. The pattern is the opposite of what a constitutional republic expects: decisions of enormous consequence, made by a handful of officials, under time pressure, with the public excluded.

1. The expectation of rescue changes behavior before the crisis. This is the heart of moral hazard. A bank that believes it will be rescued can borrow more cheaply, hold less capital and take bigger risks, because its creditors share the belief. The cost of the implicit guarantee shows up in the pre-crisis behavior and the pre-crisis profits, as well as the post-crisis bill. The savings and loan disaster of the 1980s, fueled by deposit insurance without strict supervision, is the classic illustration in banking.

1. Ad hoc rescue is a poor way to make a rule. A rule says in advance who is covered, by whom and on what terms. An ad hoc rescue says it afterwards, to the people in the room. That favors the well-connected and the large, who have the ear of the Treasury. It violates a principle conservatives defend in every other setting: the same law for everyone, announced beforehand. (I made a parallel point about picking favorites in Solyndra to Amtrak, why government picks losers when it picks winners.)

1. The bill is not only financial. The lasting cost is the erosion of the belief that a person who takes a risk should bear it. That belief is a moral foundation of a free economy. If the prudent small-town bank watches the reckless big one get rescued, the lesson is not to be prudent. And ordinary citizens, who see the system protect the well-connected and not them, learn a lesson about fairness that cannot be good for self-government.

What a rule-based alternative would look like

The honest alternative is not "let everything burn." It is to commit in advance to rules that make failure survivable and bailouts unnecessary. Here are the elements, many of which have serious advocates on both sides.

• More capital, held by the banks themselves. The simplest protection against failure is a thicker cushion of owners' money before taxpayers' money is touched. Higher equity requirements for large institutions are the most direct tool.
• A credible resolution process. A bank must be able to fail, with shareholders and some creditors taking losses, without a panic. The Dodd-Frank Act's orderly liquidation authority was meant to do this, though critics doubt it would be used as intended. A sharper bankruptcy approach for financial firms is one proposal.
• A clear, narrow lender of last resort. Follow Bagehot. Lend against good collateral at a penalty rate, publish the terms and apply them to all.
• Deposit insurance limits that mean what they say. If the limit is 250,000 dollars, then that is the limit. If Congress wants to raise it, let Congress do so by statute, with a premium that the banks pay and a vote that everyone can see.
• No bailouts by weekend. The decision to extend a guarantee beyond the law should need explicit congressional approval within a short window, so that the people's representatives, not a small group of regulators, own the choice.

None of this is simple, and none of it removes risk. But a rule has a virtue that a rescue lacks: it can be debated before the crisis, by people who are not frightened. Compare that with The 2008 housing crash began with Washington's push for easy mortgages, which shows how the earlier policy choices fed the crisis that later required the rescue, and with The Fed's dual mandate and the 1977 law that gave it two masters, which explains the other institution at the center of these decisions.

Judging the founding standard

The Founders had a view of money and debt. They wanted a government that paid its debts honestly, as Hamilton argued in his 1790 report on public credit (see Hamilton's 1790 Report on Public Credit and what honest borrowing requires), and they distrusted schemes that let a few profit at the expense of the many. Nothing in the Constitution says the federal government may guarantee the debts of private companies. That does not settle every question, but it frames the right one. When Washington backs a private firm's losses, it is exercising a power that has to be justified, not assumed.

And there is a plainer, human point. A free society depends on people being responsible for their choices, and on people being able to count on the rules staying the same. When the rule becomes "the big ones are protected and the small ones are not," trust drains away from the system and the institutions. A conservative defends the system by insisting that the rules apply to everyone, including the very large.

The case in brief

1. Moral hazard is the predictable rise in risk-taking when someone else bears the cost, and every bailout since the Chrysler Loan Guarantee Act of 1979 has extended that expectation.
2. The path runs from Chrysler in 1979 and Continental Illinois in 1984 to Long-Term Capital Management in 1998, TARP's 700 billion dollars in October 2008, the 2009 auto rescue and the guarantee of all deposits at Silicon Valley Bank in March 2023.
3. The defenders are right that panics are real, but a rescue that protects the system should not protect the owners and managers who caused the failure.
4. A rule-based alternative would combine higher capital, a credible resolution process, a narrow lender of last resort, deposit limits that mean what they say and congressional approval for any extra guarantee.
5. The principle is the one conservatives defend everywhere: the same rules for everyone, announced beforehand, with those who take the risk bearing the loss.

What you can do: read your own bank's deposit insurance terms on the FDIC's website, then ask your member of Congress in writing whether the deposit insurance limit should change only by an open vote of Congress.