• affordable housing requirements imposed by the Community Reinvestment Act and/​or imposed on the government-created housing finance giants Fannie Mae and Freddie Mac, resulting in lending to uncreditworthy borrowers
• changes in law exempting various asset repurchase agreements (“repos”) from the normal freeze on all assets held by companies in bankruptcy, which gave the impression that repos were low-risk
• excessive leverage in financial institutions’ portfolios, allowing them to borrow and then lend and invest in housing finance
• elimination of the separation between commercial and investment banking
• “predatory lending” practices, which duped homebuyers into costly and risky mortgages
• the originate-to-distribute mortgage model, which obscured the quality of the underlying loans
• “exotic” non-amortizing mortgages, which also duped homebuyers
• badly designed compensation plans for financial executives, which incentivized risk-taking

There are numerous papers and books devoted to supporting and refuting each of those hypotheses. I’ve reviewed many of them in my “Working Papers” columns in this journal (see Summer 2011, Summer 2012, Fall 2012, Spring 2013).

Gary Gorton, professor of management and finance at Yale, says we should ignore all this chatter because it misses the essentials. In his new book Misunderstanding Financial Crises, he explains, “The cause of financial crises is the vulnerability of transactions media, the privately created debt of financial institutions.” By “transactions media” he refers to debt obligations that are often traded back and forth as de facto money. They are created by financial institutions to facilitate transactions, but the institutions cannot create riskless debt. For Gorton the history of financial crises is the repeated attempt by private institutions to create transactions media—in effect, creating private money—that trades at par even though it is backed by loans of uncertain value.

The History of Private Money Creation

Initially banks were unable to create transactions media that traded at par. During the Free Banking Era (1837–1863), bank notes for transactions were issued by around 1,500 different banks. The banks invested in state debt in an attempt to create riskless collateral for the notes, but the banknotes did not trade at par. A $10 bill issued by a bank in New Haven, for example, would be worth only $9.90 in New York City. “Today, we take it for granted that when we offer a ten-dollar bill in payment, it is accepted as being worth ten dollars,” Gorton writes. “Private banknotes attempted to achieve that and failed.”

The National Bank Acts of 1863 and 1864 taxed private bank notes out of existence and collateralized national banknotes with U.S. Treasuries. The Bank Acts were enacted to finance the Civil War rather than create an efficient medium of exchange, but contemporaries recognized (Gorton quotes from publications of the time) that money needs to be backed by collateral that is safe. According to Gorton, only government can provide riskless collateral.

But after the introduction of national banknotes, financial innovation still produced private money that was subject to panic: checks and demand deposits. People feared that during recessions checks would trade at a discount. Thus, during economic contractions, they demanded cash, which resulted in bank runs and jeopardized banks’ solvency.

This did not end until the creation of deposit insurance, first by states and then the federal government. In 1911 the U.S. Supreme Court found the mandatory Oklahoma deposit insurance system and fee to be constitutional, ruling that the purpose of the system was to allow checks to trade at par. Writes Gorton, “The advent of federal deposit insurance in the United States [in 1933–1934] was the start of a long period in which effective regulation eliminated the threat of systemic financial crises—the Quiet Period.”

For Gorton financial crises are normal in market economies. Only during the Quiet Period (1934–2007) did the public, politicians, and economists come to believe that quiet was normal. The Quiet Period lasted a long time, and memories of financial panics faded and crises were viewed as a thing of the past. Because modern macroeconomics developed during the Quiet Period, almost all economists tended to ignore the crucial role played by efficient financial intermediation in determining economic outcomes and came to believe that financial crises were a thing of the past. Thus when the panic of 2007–2008 occurred, professional economists were at a loss to understand what was happening.

Why did the Quiet Period end? For Gorton the Quiet Period was the product of the combination of deposit insurance, interest rate controls, and the lack of interest on commercial checking accounts that reduced competition and created economic rents for banks with charters. Banks didn’t innovate because they were earning rents from the status quo and were protected from competition that normally would erode the rents.

Innovation did take place outside the official banking system, however, and those developments increased payments to depositors and decreased interest rates for borrowers. Money market funds competed with traditional banks on the liability side, while junk bonds, commercial paper, and—later—securitized loans substituted for bank loans on the asset side. Borrowers interacted directly with investors through processes that bypassed banks.

The net result of this competition was higher costs, fewer deposits, and an erosion of charter rents for banks. Interest expenses were 5.48 percentage points below Treasury 10-year rates in 1979. By 1986 the same measure was only 1.32 percentage points below Treasury 10-year rates. Because of pressure from non-banking competitors, banks had to pay more on deposits, and they were allowed to because of deregulation during the savings-and-loan crisis of the 1980s.

Banks not only faced more competition, but corporations also had more cash to invest. From 1980 to 2006, the ratio of cash to assets of corporations more than doubled from 10.5 percent to 23.2 percent. Commercial checking accounts at traditional banks did not pay interest. Since the early 1980s, corporate treasurers gradually invested all this cash in asset-backed securities, repos, and money market funds, which earned interest income. This “shadow” banking system was collateralized with assets thought to be safe and liquid. But these financial instruments were not insured and thus were vulnerable to panic when investors grew uncertain about the wisdom of trading them for cash at par. Frightened investors could rush to redeem the instruments, causing investment banks and other financial institutions to become insolvent. In essence, the shadow banking system was susceptible to bank runs just as the traditional banking system once had been.

The Panic of 2007–2008

The deterioration of house prices and defaults in subprime mortgages were not enough by themselves to cause a systemic crisis. A 2011 Yale doctoral dissertation by Sunyoung Park examined $1.9 trillion in AAA-rated subprime bonds issued between 2004 and 2007. The realized principal losses as of February 2011 were only 17 basis points (0.17 percent). The Financial Crisis Inquiry Commission, created by Congress to probe the causes of the financial crisis, noted that only 4 percent of subprime mortgages and 10 percent of “Alt‑A” mortgages (mortgages to borrowers with good credit scores, but that involve more aggressive underwriting than traditional conforming or “jumbo” loans) had been “materially impaired,” meaning that losses were imminent or already had occurred by 2009. So the shock itself wasn’t very large. How, then, did we get a crisis?

Gorton argues that the asset-backed securities repo market at its peak had $10 trillion in assets, about the same size as the traditional commercial banking sector. Normally $100 million “deposited” in a repo agreement would be collateralized by $100 million in bonds backed by pools of mortgage, car and student loans, or credit card receivables. Once investors got nervous about the uninsured nature of repo in August 2007, they demanded more in face value of collateral than they deposited. These so-called “haircuts” increased the most on mortgage-related assets, but they also increased on car loans, student loans, and credit card receivables, which had none of the characteristics of the suspect mortgage instruments at the heart of the crisis.

When investors demanded such “haircuts,” financial institutions were forced to sell assets (the bundles of securitized loans). The simultaneous selling of similar assets led to severe discounting from face value. By September 2009, Gorton reports that repo transactions had fallen by 50 percent from their pre-Lehman bankruptcy levels. The runs on repos spread to commercial paper and prime brokerage markets as well. This was “a breakdown of the central nervous system of the economy—financial firms.” The privately created debt outside the deposit insurance system did not trade at par.

The Common Characteristics of Financial Panics

What unites the panics across nearly two centuries? Gorton argues that in financial crises depositors seek to exit money that is risky (bank deposits) and obtain money that is safe and remains acceptable as a means of payment at par valuation (cash). The particular forms of bank debt vary over time from private bank notes in the 1840s, to checking accounts in the 1870s, to overnight asset repurchase agreements in the 2000s. The demands for cash are on such a scale that they cannot be met.

During crises, according to Gorton, financial institutions have only three options:

• Suspend the contractual right of convertibility of deposits to cash.
• Sell assets to raise cash to satisfy the demands of depositors (but the simultaneous sale of similar assets by financial institutions raises much less cash than face value, and the result is the liquidation of the banking system).
• Receive support from the government or from central bank purchase of assets.

Historically, before deposit insurance, financial crises were always handled through suspension of normal banking rules about access to bank deposits. The suspension of the rules has always troubled those who care about the rule of law. In 1837 John Quincy Adams commented on the suspension of convertibility by saying, “We are now told that all the banks in the United States have suspended specie payments—and what is the suspension of specie payments but setting the laws of property at defiance?” Adams was asking whether suspension was fraud. Shouldn’t the banks then lose their charters?

The New York state legislature legalized suspension of specie payment (the termination of the right to exchange banknotes for gold) for one year in 1837. In 1846 New York enacted a constitutional amendment that supposedly prohibited the legislature from repeating that policy. But during the 1857 panic, convertibility was suspended again. And the New York Supreme Court ruled in 1857 in Livingston v. Bank of New York that suspension of convertibility during crises would not trigger liquidation, even though the state constitutional amendment was designed to prevent time-inconsistent behavior. In 1896 Yale professor William Graham Sumner described the Livingston decision as a coup d’état—but he believed that the time-inconsistent behavior of the law was probably a better alternative than liquidation of the banking system.

In 2007–2008 this same logical conundrum repeated itself. Ben Bernanke used language in the Federal Reserve Act to offer lines of credit to financial institutions that were not banks and he offered financial support to the commercial paper and asset-backed securities markets. Critics charged this violated the rule of law just like John Quincy Adams did 170 years earlier.

In the absence of deposit insurance, bailouts and contract suspension are the only means of preventing liquidation of the banking system during financial crises. In the current crisis, bailouts were chosen rather than contract suspension, but the popular reaction, as well as the reaction of Bernanke himself, was similar to John Quincy Adams’ reaction to contract suspension. (From Bernanke’s March 15, 2009 60 Minutes interview: “I slammed the phone more than a few times on discussing AIG. I understand why the American people are angry. It’s absolutely unfair that taxpayer dollars are going to prop up a company that made these terrible bets—that was operating out of the sight of regulators, but which we have no choice but to stabilize, or else risk enormous impact, not just in the financial system, but on the whole U.S. economy.”) Both bailouts and contract suspension would seem to violate the rule of law and the public’s general sensibilities about how markets should work.

Narrow Banking, Free Banking, or Deposit Insurance?

Analyses of the financial crisis by authors other than Gorton focus on the details of the current private money debacle and recommend changes in those details: e.g., eliminating housing affordability goals; increasing capital requirements; pushing homebuyers toward traditional, rather than exotic, mortgage instruments; re-separating commercial and investment banking (adopting the “Volcker Rule”); changing the bankruptcy status of repos. Gorton argues that private-sector innovations in debt contracts arise spontaneously as memories of previous financial crises fade, and so cracking down on the last wave of private debt innovation is akin to generals fighting the last war. Privately created debt is inherently susceptible to economic and informational shocks because it cannot be riskless, he argues. None of the changes enacted in the 2010 Dodd-Frank Act and the recommendations of academics (with the exception of bankruptcy changes) address the fundamental problem as Gorton sees it: the inherent instability of privately created debt. So what regime changes could tackle this problem?

Narrow banking | Some libertarian analysts agree with Gorton’s characterization of the inherent instability of privately created debt. They conclude that the costs of banking exceed its benefits and propose that conventional banking (the private creation of debt that trades at par) be banned and replaced with “narrow banking,” where the payment system is fully backed by truly risk-free investments—cash and treasuries. Narrow banking is an attempt to carefully demarcate the difference between the transactions media of the payments system (checking accounts, passbook savings, and other demand deposits like repo) and all other investment, which would be at-risk.

An important problem with narrow banking is the time-inconsistency of government policy as described by Gorton. Defining the demarcation line between transactions media that are fully backed with safe collateral and at-risk investments seems to be subject to change. Fannie Mae and Freddie Mac were not legally backed by the government and their liabilities were not covered by deposit insurance, nor were money market funds, nor overnight repos. But when the financial system came under severe stress, the government changed the policy and backed that debt. Likewise, if future stress were to hit at-risk, privately created transactions media, placing the broader economy in jeopardy, the White House and Congress would surely ride to the rescue again. This is what Gorton describes as the paradox of financial crises: the necessity of time-inconsistent behavior.

A second problem with narrow banking is that it would require the legal suppression of what most people now call banking, i.e., financial intermediation in which short-term deposits support longer-term investment through privately created debt. Banks arise naturally in a free society, as Gorton’s history documents, and the narrow banking regime would require the use of the power of the state to suppress the transformation of demand deposits into investment.

Free banking | Another libertarian proposal is a return to “free banking.” Unlike narrow banking, free banking allows fractional reserves. But unlike the pre–Civil War era of free banking described by Gorton, in which branch banks were largely forbidden, branching would also be allowed. This would strengthen the banks: some (maybe much) of the discounting of private banknotes during the pre–Civil War era was the result of the geographically undiversified nature of the loan portfolios of those banks rather than the state debt used as collateral.

In addition to geographic diversification, truly free banks would include convertibility suspension clauses as part of their initial deposit contract, a feature of free Scottish banks in the 1700s as described by George Selgin. That way, currency convertibility suspensions would not violate the rule of law. The problem is that in a world in which currency convertibility suspensions were possible, depositors would certainly have incentive to withdraw deposits before the suspension was announced. And given fractional banking, those who withdraw sooner are more likely to succeed than those who wait, which puts banks at risk of runs. So the essential game-theoretic logic of financial crises is thus not altered even if contractual suspension replaces the time-inconsistent paradox described by Gorton.

Selgin positively describes the role that discounting of the notes could play in preventing random runs on financial institutions. That is, those banks with loan portfolios thought to be troubled by the market would have notes that traded at greater discount than those that did not. But, for Gorton, once privately created transactions media begin to trade at less than par, and the discounting is volatile rather than constant, then they no longer function as effective transactions media. If energy and effort must be exerted to ascertain their price relative to par, then little difference exists between using privately created currency with time-varying discounts relative to par and stocks and bonds as transactions media. Such transactions media would be efficient but filled with transaction costs.

Deposit insurance | The response favored by Gorton would extend to all demand deposits the current safety net of deposit insurance and access to the Fed. He also would limit the repo exception to bankruptcy’s “automatic stay” rule, which halts most creditors’ efforts to collect on debts. Limiting the repo exception would reduce the risk of runs on troubled financial firms’ assets. (This is the one exception to my claim that Gorton is uninterested in the details of the recent crisis.) This proposal would bring the shadow banking system into the regulated system, extending deposit insurance to special banks that invest in asset-backed securities. Money market funds would be covered indirectly through their purchase of debt of these special repo banks. The bankruptcy privileges (the exemption of repos from the automatic stay provisions of bankruptcy) would be eliminated for repos outside the approved venues. As a consequence, Gorton believes all repos outside the approved venues would disappear.

But even he recognizes that the logic of his historical inquiry suggests the impossibility of any particular solution:

To design a bank regulatory environment that addresses the vulnerability of bank debt and fosters economic growth is possible in principle. But because of the paradox of financial crises, it might not be possible in practice. …This suggests that the idea that any one policy, such as deposit insurance, would forever solve the problem of crises is naive.

A second essential component of Gorton’s solution is central bank discretion:

Because of the paradox of financial crises, central bankers must be independent so that they can take unpopular actions to keep the banking system from being liquidated…. During noncrisis times most economists think that the central bank should focus on fighting inflation based on rules rather than discretion. But in crisis times it is the opposite.

The classic rule of Walter Bagehot provides some guidance: during a crisis a central bank should lend freely to solvent firms, against good collateral and at high rates. But, Gorton writes, “It is hard to see how the answers to these questions could be pre-specified as rules. And if they could, these rules would likely not be followed in the next crisis.”

So what would stop financial institutions from taking on higher-than-appropriate risk if they know Washington stands ready with bailout money? Gorton argues that such concerns did not enter into the recent crisis:

At least at the outset of the crisis, the run was not contaminated by expectations about the government’s crisis responses. It is hard to understand how policies of too-big-to-fail and moral-hazard-related incentives emanating from government regulation could have affected the dealer banks. The government did not know of the existence of the shadow banking system. Further, in light of the testimony of dealer bank CEOs before the Financial Crisis Inquiry Commission, it is doubtful if even the dealer banks understood the changes to the financial system.

Rene Stulz and his colleagues provide empirical support for the Gorton perspective. They attempted to determine why the holdings of highly rated securitization tranches differed so much across banks before the financial crisis. The median was 0.2 percent and the mean 1.4 percent. Citibank had the largest amount at 10.7 percent. Observers commonly argue that investing in these assets was a form of excessive risk-taking. Stulz et al. argue against the logic of bad outcomes as conclusive evidence: “ex post adverse outcomes are not evidence of risk management failures.… [I]t does not logically follow from the poor performance of highly-rated tranches that risk management failed.” Instead Stulz and his colleagues examine whether the variation in risk management practices before the crisis was related to holdings of highly rated securitized tranches controlling for other bank characteristics. Using an index that measures the centrality and independence of risk management to a bank, Stulz found no relationship with larger holdings of highly rated tranches.

Some argue too-big-to-fail explains why large banks held these assets. Large banks (assets greater than $50 billion) did hold more than small banks, but among large banks (36 banks have assets greater than $30 billion as of the end of 2012) holdings of highly rated tranches as a percent of assets did not increase with bank size.

Even though Gorton may be correct that the moral hazard created by the paradox of financial crises did not enter into the current crisis, it may affect behavior going forward. Empirical evidence suggests that after the start of the recent crisis, large financial institutions paid lower interest for deposits because the market perceived them to be too big to fail. According to the New York Times, in the last quarter of 2006, before the crisis, banks in all size categories paid 3.6 percent to 3.65 percent on deposits—a very narrow range. In contrast, in the fourth quarter of 2009, institutions with more than $100 billion of assets paid an average of 0.77 percent annual interest on deposits, while institutions with less than $10 billion in assets paid an average of 1.73 percent. The 10 largest banks hold about $3.2 trillion of America’s $7.7 trillion of domestic deposits. Apply the differential in deposit interest rates, and those 10 banks appear to be saving nearly $30 billion a year thanks to their size.

Conclusion

So banks present a dilemma for libertarians. The asset transformations that banks can achieve (the provision of liquid transactions media that trade at par) while investing in longer-term loan portfolios enhance economic growth. But those very characteristics periodically result in systemic events in which contract suspension or bailouts are used to prevent the liquidation of the financial system. Deposit insurance and the Quiet Period lulled us all into thinking that systemic events were just interesting historical events. But how wrong we were!

Unlike other scholars, Gorton argues there are no solutions, just patches on the current system of deposit insurance that will also fail in the future in ways that we cannot now predict, combined with clean-up lender-of-last-resort activities by the Fed. Even though libertarians will have a predictably negative reaction to Gorton’s book, his analysis and views merit serious consideration.

Readings

• “An Interview with George Selgin.” Region Focus (Federal Reserve Bank of Richmond), Winter 2009.
• “Another Advantage for the Biggest Banks,” by Rob Cox and Lauren Silva Laughlin. New York Times, March 29, 2010.
• “In Praise of More Primitive Finance,” by Amar Bhidé. The Economists’ Voice, Vol. 6, No. 3 (February 2009).
• “Limited Purpose Banking—Putting an End to Financial Crises,” by Christophe Chamley and Laurence J. Kotlikoff. Financial Times, January 27, 2009.
• “Rethinking the Roles of Banks: A Call for Narrow Banking,” by Oz Shy and Rune Stenbacka. The Economists’ Voice, Vol. 5, No. 2 (February 2008).
• “Why Did Holdings of Highly Rated Securitization Tranches Differ So Much Across Banks?” by Eril Isil, Taylor Nadauld, and Rene Stulz. SSRN #2186174, December 2012.