Averting Financial Crises: Advice from Classical Economists

Classical Teachings

What was classical LLR theory? By classical here, I mean the work primarily of two Englishmen, namely Henry Thornton (1760-1815), a prominent banker, member of Parliament, evangelical reformer, anti-slavery activist, and all-time great monetary theorist writing in the early years of the 19th century, and Walter Bagehot (1826-1877), a financial writer and longtime editor of The Economist magazine who wrote in the century's middle decades.

Classical LLR theory referred to the central bank's duty to provide emergency injections of liquidity to a banking system facing massive cash withdrawals when no other liquidity source is available. The central bank fulfills this duty either through discount window loans to stressed banks or through open market purchases of Treasury bills, bonds, or other assets. Because open market operations were infrequently used in 19th century England, classicals instead advocated discount window loans, albeit at high interest rates so as to discourage too-frequent resort to the loan facility, to creditworthy, cash-strapped borrowers offering good collateral. The goal was to prevent bank runs that cause sudden, sharp contractions in the money stock, and thus declines in spending and prices. Given downward inflexibility of nominal wages, these declines lead to rising real wages and corresponding falls in profits that induce collapses in output and employment, collapses the classicals fervently sought to avoid.

But classicals noted that the LLR has no business bailing out unsound, insolvent banks. Its mission is to stop liquidity crises, not insolvency ones. Nevertheless, if the LLR acts swiftly, aggressively, and with sufficient resolve, it can prevent liquidity crises from deteriorating into insolvency ones. By creating new money upon demand for sound but temporarily illiquid banks, the LLR makes it unnecessary for those banks, in desperate attempts to raise cash, to dump assets at fire-sale prices that might render banks insolvent.

Two lessons emerge from classical LLR theory. Lesson number one: Filling the market with liquidity — or, even better, credibly pre-committing to do so in all current and future panics — is sufficient to still panics and end crises. Liquidity provision by itself is enough to do the job. There is no need also to bail out insolvent, poorly managed institutions or to charge below-market subsidy interest rates on LLR loans.

Lesson number two: The panic- and run-arresting duties of the LLR are part and parcel of its monetary stabilization responsibilities. The two tasks are not mutually exclusive. They are one and the same. By keeping the money stock — or better still, that stock adjusted for shifts in the demand for it so as to preserve money supply-demand equilibrium — on track in the face of shocks, panics, and crises that otherwise would shrink it, the LLR preserves nominal income and spending at their full capacity, non-inflationary, non-deflationary paths.

Thornton's Contributions

Although Walter Bagehot is the economist most often identified with classical LLR theory, Henry Thornton, writing decades before him, can lay claim to being its true father. What did Thornton do? For starters, he identified the LLR's distinguishing feature as its open-ended power to create base or high-powered money in the form of its own notes and deposits. The Bank of England possessed this power in spades during the Napoleonic Wars when the government had released it from the obligation of maintaining gold convertibility of its currency.

Thornton also noted that the LLR has a macroeconomic duty to the entire economy, or the "general interest," as he called it. This duty differentiates the LLR from an individual banker whose duties extend only to his bank's owners and customers. Let a panic occur. The individual banker will seek to contract his loans and deposits knowing that such contraction will boost his safety and liquidity without much affecting the whole economy. By contrast, the LLR, because it governs the entire money stock whose shrinkage will have widespread adverse effects, can make no such assumption. Thus, when panic hits, the LLR must act opposite to the banker, expanding its operations at the very time the banker is contracting his.

Another thing Thornton did was to identify the LLR's chief purpose as a monetary rather than a banking or a credit one. To be sure, the LLR acts to forestall bank runs and avert credit crises. But these credit-market actions, although vitally important, are not the end goal of policy in and of themselves. Rather, these actions are the means, albeit the most expedient and efficient means, through which the LLR pursues its ultimate objective of protecting the quantity, hence purchasing power, of the money stock. The crucial task is to prevent sharp and sudden shrinkages of the money stock since hardship follows from these rather than from bank runs and credit crises per se.

Why did Thornton see the LLR's function as a monetary rather than a credit one? Simple. He thought that money does what credit cannot do, namely, serve as the economy's unit of account and means of exchange. Since money forms the transaction medium of final settlement, it follows that its contraction, rather than credit crunches and collapses, constitute the root cause of lapses in real economic activity and of breakdowns of the payments mechanism.

Bagehot's Contributions

Building upon Thornton's earlier work (although never once citing him, for which I have no explanation), Bagehot added four propositions of his own.

First, the LLR, when quelling panics, should lend to all sound borrowers — nonbanks as well as banks — offering good security, namely assets that would be deemed creditworthy and valuable in ordinary or normal times if not in panics.

Second, the LLR has no duty to bail out unsound borrowers, no matter how big or interconnected. Such bailouts produce moral hazard: They encourage other banks to take excessive risks under the expectation that they too will be rescued if their risks turn sour. To Bagehot, lender-borrower interconnectedness and the purported associated danger of systemic failure constitute no good reason to bail out insolvent banks. Better to let bad banks fail and prevent their failure from spreading to the sound banks of the system. And the best way to do this is to pre-commit to pour liquidity without stint into the market in a crisis.

Here it would be remiss not to note that even on the moral hazard issue, Thornton had scooped Bagehot 70 years before the latter published Lombard Street. In a prescient footnote on page 188 of his 1802 book An Enquiry into the Nature and Effects of the Paper Credit of Great Britain, Thornton wrote that it was not up to the central bank "to relieve every distress which the rashness of country banks bring upon themselves." Relief instead should go to protect "the general interests" and not "those who misconduct their business." The latter must be left to suffer "the natural consequences of their fault." Thornton noted that unsound banks "no matter how ruinous their state" would nevertheless plead that rescuing them was necessary to save the general interest.

Bagehot's third point was that the LLR should charge above-market or penalty rates of interest on its accommodation. This is the famous Bagehot Rule: Lend freely but at a high rate. The high rate does several things. It discourages unnecessary resort to the discount window. It encourages would-be borrowers to exhaust all market sources of liquidity and even to develop new sources before applying to the central bank. It discourages overcautious hoarding of scarce cash. It attracts gold from abroad and encourages gold's retention at home, thus protecting Bagehot's cherished gold standard while bolstering the monetary base. A high rate also rations liquidity to its highest-valued uses. It serves as a partial test of borrower soundness since only solvent banks can afford to pay the penalty rate, even though unsound banks facing credit risk premia in excess of the penalty rate-market rate differential may be tempted to try. It also appeals to distributive justice on the grounds that it is only fair that banks pay handsomely for the security and protection provided by the LLR. And it encourages prompt repayment of LLR loans — and removal and extinguishment of money used to pay them — at panic's end, thus eliminating inflationary monetary overhang.

Fourth, not only must the LLR act promptly, vigorously, and decisively so as to erase all doubt of its determination to end current panics. It must also pre-announce its commitment to lend freely in all future panics. Such credible pre-commitment dispels uncertainty and promotes full confidence in the LLR's willingness to act. It generates a pattern of stabilizing expectations that ease the LLR's task. Confident that the LLR will deliver on its commitment, the public will not run on the banks, perhaps obviating the need for emergency liquidity in the first place.

The Thornton-Bagehot precepts served England well. After 1866, the nation suffered no bank runs until 2007. By contrast, in the United States, the Federal Reserve honored the classical doctrine as much in the breach as in the observance, and the nation suffered dearly for it. The Fed disregarded the classical advice altogether in the 1930s and so failed to stop a massive monetary contraction that contributed mightily to the Great Depression. Most recently, however, the Fed seems to have absorbed some, but not all, of the classical wisdom. In the recent financial crisis, the Fed followed the Thornton-Bagehot prescription regarding liquidity provision while departing from other of its precepts.

Classicals on Fed Crisis-Management Policy

What would the classicals have thought about the Fed's handling of the crisis? Certainly they would have applauded the Fed's filling the market with liquidity. Likewise, they would have approved of the Fed's expansion of its balance sheet and of the monetary base. These things were precisely what the classical prescription called for — expanding the monetary base to match corresponding increases in the public's and bankers' demand for money.

At the same time, classicals might have noted that the Fed's expansion of the monetary base, while sufficient to offset the panic-induced fall in the multiplier relationship between base and bank money in a fractional reserve system, was insufficient to counter falls in velocity caused by the public's flight to money as the safest liquid asset. The result of this increased money demand (or fall in velocity) was a shortfall of the supply of broad money below the demand for it, leading to a prolonged fall of spending, output, and employment below their pre-crisis paths. [Editor's Note: For elaboration on this view and those that follow, see Readings.]

Likewise, the classicals would have approved of the Fed's Bagehot-like actions to lend to a wide variety of borrowers on a wide array of assets. But they would have looked askance at the Fed's acceptance of opaque, dubious, hard-to-value collateral that arguably would have been deemed questionable even in normal times. The same holds for the Fed's direct purchase of tainted assets.

Most important, Thornton and Bagehot would have condemned both the Fed's bailout of arguably insolvent, too-big-to-fail firms such as American International Group Inc. and Citigroup and its charging of subsidy rather than penalty rates for its assistance.

And they would have scolded the Fed for extending its loan deadlines beyond very short-term (week- or at most month-long) intervals, for its failure to pre-commit to ending all future crises, and for not spelling out the conditions and indicators that would trigger its actions in future crises.

Thornton, who sharply distinguished between the monetary and credit rationales of LLR policy, would have disagreed with the Fed's credit-market rationale. To Thornton, the LLR's purpose was to protect the money stock from contraction and to expand it to offset falls in velocity. This was in sharp contrast to the Fed's stated LLR rationale, which was to free up credit markets, shrink panic-widened yield spreads, and get banks lending again. Thornton would have shunned the Fed's credit-market rationale even though it achieved much the same result as his monetary one.

Finally, classicals might have opposed the Fed's payment of positive interest on excess reserves. The Fed implemented this measure in 2008 to prevent its credit interventions from resulting in monetary expansion. And it retained the interest-on-excess-reserves measure even when it later shifted to a policy of monetary expansion. Such payments, which boost demand for idle reserves and keep them immobilized in reserve accounts rather than getting them lent out into active circulation in the form of bank deposit money, would be inconsistent with the classicals' goal of expanding or maintaining the stock of broad money as required to keep economic activity at its pre-panic level. Bankers' demands for reserves already are extraordinarily elevated during crises. Paying interest on excess reserves only raises those demands further.

Despite claims to the contrary, the Fed never acted as an unmitigated classical LLR in the recent financial crisis. Instead, it adhered to parts of the classical prescription while deviating from others. So when you hear the Fed described, often by Fed policymakers themselves, as a classical LLR, be skeptical.