In the runs on Credit Suisse this week, Silicon Valley Bank last week, and troubled financial institutions anywhere, the dilemma facing governments is whether and how to intervene.  When should a government simply let a financial institution fail as opposed to sending in cash to keep it alive?  There is a traditional answer and a current-day answer: the difference says a great deal about how economic orthodoxy has shifted over the years.


The Traditional Banking View

Under the traditional view, the critical question is whether the troubled institution is merely out of cash (illiquid) or is unable to meet its liabilities even if all assets were converted to cash (insolvent). Conventional thinking at the turn of the twentieth century was that in a panic, illiquid institutions should be supported, and insolvent institutions should declare bankruptcy. The iconic expression of this was Walter Bagehot’s Lombard Street: A Description of the Money Market, published in 1873.

In Bagehot’s day, Lombard Street in London was the center of the international money market, and the location of the Bank of England. Bagehot styled the book as a kind of primer for new professionals on Lombard Street. It was an extended argument in favor of a muscular central bank, one that would serve the liberal (and mercantile) policies of Britain and that would generally create the conditions for economic stability and advancement. An important responsibility of such a bank, he argued, was to fight financial crises by being a “lender of last resort” (LOLR).[i]

Bagehot’s advice for such lenders was that in a financial crisis, the LOLR should lend freely to solvent borrowers upon good collateral and at a “penalty rate” of interest. Consider the three elements of Bagehot’s instruction:

  1. Lend freely: Illiquidity is a hallmark of financial crises. When depositors run and withdraw their savings from banks—both the healthy and distressed ones—banks call in loans. Credit, the lifeblood of commerce, stops flowing. By lending freely into the financial system, the LOLR supplies the short-term funding that banks need and thereby helps to quell runs and forestall the credit contraction that sparks a debt-deflation spiral.
  2. Solvent borrowers upon good collateral: This is a “tough love” policy. To demand good collateral is to guard against losses by the LOLR. And it also means that the LOLR will really support only the solvent institutions and will necessarily allow insolvent ones to fail, preventing the persistence of “zombie banks” and the misallocation of resources they entail. It has been said that capitalism without bankruptcy is like Christianity without hell.[ii] A world without the risk of failure fuels moral hazard.
  3. At a penalty rate: Charging high rates of interest in a crisis may seem like extortion to stressed banks. But high rates may be useful. They discourage opportunists who would exploit the eagerness of the LOLR to lend freely and end the crisis. They discourage moral hazard by penalizing lenders who were too expansive. And they entice capital to come out of hoarding. Less-than-penalty rates might distort the efficient allocation of capital in an economy and/or fuel moral hazard.

In The Panic of 1907: Heralding a New Era in Finance, Capitalism, and Democracy, (2023, John Wiley & Sons), Sean Carr and I described the efforts of J.P. Morgan to quell the crisis.  Though historical archives did not explicitly cite Bagehot’s advice as Morgan’s guiding principle, Morgan’s behavior during the crisis was quite consistent with it.  Under Bagehot’s logic, Morgan would send auditors to scrutinize the books and assets of struggling institutions.  If the auditors determined that the institution was solvent (that its assets exceeded its liabilities), then Morgan found the money to keep it afloat.

Early in the Panic of 1907, Morgan faced the dilemma of whether to fund a rescue of the Knickerbocker Trust Company, one of the largest in New York City.  He sent his auditors over and they returned the next morning to say that it was not possible to tell whether Knickerbocker was solvent—they did not say that Knickerbocker was insolvent, but that they didn’t have enough time to get all the facts before the institution would run out of cash.  Perhaps relying on the opinion of other bankers who knew Knickerbocker better than he did, Morgan declined to assist the firm.  Knickerbocker promptly closed its doors and handed the keys to the New York State Superintendent of Banks.  Knickerbocker’s failure greatly inflamed the fears of depositors and worsened the crisis.  Thereafter, Morgan promptly responded to appeals for liquidity from troubled institutions.


Policies Adapt to Changing Conditions

Over the following century, orthodox thinking about financial rescues began to change, reflecting several developments.

  • Increased complexity. Banks and financial markets are complex institutions, which makes it difficult to know exactly what is going on. Complexity creates a host of information asymmetries among managers, regulators, depositors, long-term creditors, stockholders, and bank employees—each group has a different take on the bank’s condition and communicates that at digital speed to others. The resulting information asymmetries spawn problematic behavior (such as runs) that arises from adverse selection[iii] and moral hazard.[iv]
  • Greater connectivity. Yet allowing some institutions to fail might worsen a crisis. The rescue of Continental Illinois Bank in 1984 was premised on the assertion that it was “too big to fail,” that its failure would impoverish too many depositors.  Then in 2008, the distress of Bear Stearns, Citigroup, and AIG Group sharpened that rationale to be “too interconnected to fail,” meaning that the failure of any of those institutions would trigger the failure of too many others.  Interconnections exist through transactional commitments among financial institutions, most importantly, in the funding markets by which banks lend or borrow cash to each other.
  • Adaptability. The financial sector morphs over time, reflecting the embrace of financial innovations in markets, institutions, instruments, and processes.  Like the generals who are always prepared to win the previous war, crisis-fighting playbooks tended to reflect the previous But modern financial innovations render previous playbooks obsolete.  In the case of J.P. Morgan in 1907, and of Federal regulators in 2020 and 2023, we saw responses that were more ad hoc and experimental, reflecting adaptation of policies to conditions in real time.
  • Possible market inefficiency. During a crisis, do market prices truly reflect the intrinsic value of financial assets?  Prices driven by fear and fire sales of assets may severely understate fundamental values.  This was an acute problem in the crisis of 2008, when some kinds of financial securities proved hard to value.  A court case adjudicated during the Panic of 1857 (Livingston v. Bank of New York) established a precedent invalidating the closure of banks if they could not provide cash to depositors during a panic—this case was cited during the crisis of 2008 in the defense of banks seeking to avoid regulatory administration.
  • Risk.  The larger context will certainly influence the rescue decision.  Is the nation at war?  Is it facing a pandemic?  The presence of existential crises will motivate a government to take dramatic steps to quell a financial crisis.  It will focus on liquidity and rescues first, and solvency later. Demanding penalty interest rates might work against the goal of restoring liquidity to financial markets and institutions.
  • Interests. This is the dark underbelly of all bank rescues.  Lurking in the background might be favoritism of the few at the expense of many.  These few could include those led by cronies of the decision-makers—such was the allegation toward J.P. Morgan’s rescues as asserted by Senator Robert La Follette in 1908 and Representative Charles Lindberg Sr. in 1911.  Such assertions led to Congressional hearings about the existence of a “money trust.”  Or a rescue could be motivated by the strategic interests of the country, a region, or an industry—examples would be the government takeover of seven bankrupt railroads in 1976 and Federal loan guarantees for the almost bankrupt Chrysler Corporation in 1980.  The real politique of government rescues of businesses often comes down to who goes to bat for you and how big a bat is swung.

…and so on.  A complete discussion would cover other factors beyond the scope of this post. But by now, you can see that the current-day decision to rescue a financial institution will be more contingent and less doctrinaire, and probably erring on the side of more rescues rather than fewer.  Rescue policies today might be labeled “Bagehot Plus,” meaning continued commitment to liquidity of the financial system (“lend freely”), but less about “penalty rates” and more about tailoring of crisis response aimed at restoring normal conditions quickly.

The larger point of this traditional-versus-modern comparison of rescue policies is that orthodox thinking about how best to quell financial crises has changed over time and will continue to do so.  Traditional orthodoxy made sense when currencies were pegged to a gold standard, capital flowed slowly around the world, London was the undisputed hegemon of global finance, and citizens held low expectations for the intervention of national governments in financial crises.  But times have changed.  The dollar and other currencies float freely; capital moves among banks and across borders at the click of a few computer keys; news moves instantaneously; financial innovation has deepened vastly the interconnection among banks and countries; and the Fed in Washington and the New York financial community are dominant players in the world but lots of hungry competitors want their power and influence. Most importantly the citizens of developed economies expect their governments to restore financial order, much as they maintain law and order on the streets.

A final point: this shift in orthodox thinking is not lost on people who worry about the vitality of democracy.  The delegation of heightened powers to stabilize the financial system and commit large sums of taxpayer money without Congressional authority is attracting growing concern by critics on the right and left of the political spectrum. You should expect to see more debate over the authority of regulatory agencies.



[i] Bagehot did not invent the phrase “lender of last resort.” In 1802, Henry Thornton published An Enquiry into the Nature and Effects of the Paper Credit of Great Britain, in which he proposed that the Bank of England should be a backstop for the financial system. Bagehot’s novel contribution was to suggest how the central bank should perform such a function.

[ii] This statement has been attributed to Frank Borman, a former astronaut who became CEO of Eastern Airlines, a company that ultimately went bankrupt. He allegedly added, “But it is hard to see any Good News in that” (J. Madeleine Nash, Bruce Van Voorst, and Alexander L. Taylor III, “The Growing Bankruptcy Brigade,” Time, October 18, 1982).

[iii] Adverse selection arises if a better-informed party in a transaction can exploit information to the disadvantage of the less well-informed party. Think of buying a used car (the seller knows more about the condition of the car, possibly a “lemon”) or selling health insurance (the buyer knows more about his or her health outlook). Concern about adverse selection may drive parties out of the market, thus diminishing liquidity and the ability of the market to clear. Also, adverse selection might drive quality goods out of the market because sellers of high-quality goods cannot obtain the prices they deserve. In finance, “Gresham’s Law” (i.e., bad money drives out good money) is an example of adverse selection.

[iv] In the case of moral hazard, a party to an agreement fails to act in good faith and shifts risk onto counterparties. For instance, debtors who believe that the government will always bail them out in a crisis may simply borrow more, ultimately shifting risk onto taxpayers.