The Sound of Five Thousand Banks Collapsing

Acting as the lender of last resort to banks by making them collateralized loans from its “Discount Window” was a principal reason for the creation of the Federal Reserve in 1913.  It is still a key function, although at present, loans to banks represent only 0.06% of the Fed’s assets.  Nonetheless, it is a capability that has proven very handy in the many financial crises of the Fed’s career so far.  

With the passage of the Federal Reserve Act, many hoped that the new Federal Reserve Banks would make future financial crises impossible. William G. McAdoo, for example, the Secretary of the Treasury at the time—and therefore, under the original act, automatically the Chairman of the Federal Reserve Board—had this excessively optimistic prediction:

The opening of [the Fed] marks a new era…[It] will give such stability to the banking business that the extreme fluctuations in interest rates and available credits which have characterized banking in the past will be destroyed permanently.

Only one decade later, the still young Fed was facing the failure of thousands of banks, principally smaller banks across the country's agricultural regions. To be specific, from 1921 to 1929, 5,711 U.S. banks went broke. That is an average of 635 failures per year, or about 12 per week for a period lasting nine years, all in a decade that elsewhere featured the 1920s boom.

This now almost entirely forgotten banking bust, and the Federal Reserve Banks’ widespread use of the Discount Window during it, are instructively recounted and analyzed by Mark Carlson in his new book, The Young Fed—The Banking Crises of the 1920s and the Making of a Lender of Last Resort. The book reviews both colorful specific cases and the fundamental ideas involved. Calson makes it clear that the Federal Reserve Banks and Board officers were well aware of and thought carefully about the tensions and trade-offs inherent in their lender of last resort activities.

These include the systemic problem of having many banks in trouble at the same time, the problems of the moral hazard induced by central bank lending, distinguishing illiquidity from insolvency in the time pressure and uncertainty of a crisis, a banker’s first loyalty to depositors vs. the central bank’s shifting losses to depositors, and the central bank itself taking credit risk.

Thus, the book is both an interesting history and an exploration of some core concepts in banking and central banking.

Hundreds or Thousands of Banks in Trouble at the Same Time

As Carlson writes, “the disruptive effects of bank failures are particularly relevant when many banks are in trouble at the same time.” At that point, “the sudden liquidation of all assets of the troubled financial institutions would be disastrous.” Yes, all the troubled banks cannot sell their assets simultaneously when their lenders and depositors want their money back.  

An old Washington friend of mine once asked, “If a good pilot can in an emergency land a jet plane in the Hudson River like Captain Sullenberger did, why can’t we handle financial crises better?” I replied, “To get the analogy right, you would have to picture landing a hundred crippled jets in the Hudson River together.”

“In the 1920s,” says Carlson, “the troubles in the banking system were also widespread… Congress leaned heavily on the Federal Reserve to support the banks.” The book suggests that the new Federal Reserve did well under the 1920s circumstances. The Fed “experienced a large number of successes when providing emergency funds to banks,” Carlson observes, but “other banks failed with discount window loans outstanding, which put the Federal Reserve in uncomfortable situations.”

Why were so many banks in trouble at the same time? The root cause was the First World War, as war was often the key factor in big financial events. As Carlson explains, the “Great War” created a huge agricultural boom in the U.S., and the boom set up the bust. Because of the war, “prices of agricultural commodities soared globally. U.S. farmers bought more land, planted more crops, and raised more livestock amid expectations that the boom would last.... A significant proportion of the expansion was financed through borrowing. … The price of farmland rose notably. … To purchase the increasingly expensive land…farmers needed to borrow.” Whether they needed to or not, many did. So did land speculators.

Then: “the collapse was as dramatic as the run-up had been. … Foreclosures [and bank failures] surged.” Across vast swaths of the country, an agricultural banking crisis descended.  

Moral Hazard

In the 1920s, the Fed was fully aware that being ready to support troubled banks could induce more banking risk—the problem of “moral hazard.” This is the well-known general risk management problem in which saving people whenever they get in trouble makes them more prone to risky behavior.  

Carlson writes, “The greater availability of the discount window…meant that managers and shareholders had more incentives to take liquidity risks.” He quotes the Federal Reserve Bank of Dallas in 1927: “Those extensions of credit simply serve to create further opportunities to make the same mistakes of judgement and to further prosecute the same unsound policies.”  

The moral hazard dilemma of central banking, apparent 100 years ago and, indeed,100 years before that, will always be with us.

Distinguishing Illiquidity from Insolvency in the Pressure of the Crisis

In theory, lenders of last resort should address the problem of illiquidity, where the troubled bank is short of cash but the real value of its assets still exceeds the claims of its depositors and lenders. Thus, in theory, central banks should lend only to solvent borrowers. But to paraphrase Yogi Berra, in theory, illiquidity is different from insolvency, but in practice it often isn’t.  

In Carlson’s more scholarly language:

A lender of last resort will often find it difficult to fully determine the extent to which the need for support is the result of insolvency versus illiquidity. …The experiences of the Federal Reserve in the 1920s highlight that in some cases it will be impossible to determine whether a bank is solvent at the time it requests funds.

In short, the lender of last resort suffered in the 1920s, and in crises, it will always suffer a “fog of financial crisis”—just like generals in Carl von Clausewitz’s celebrated “fog of war.” It is inevitably very difficult to know what is really going on and how big the losses will be.

The Banker’s First Loyalty vs. Shifting Losses to Depositors

The 1920s Fed was clear about classic banking ethics. Carlson quotes the Federal Reserve Bank of Chicago as “firmly of the opinion that the prime obligations of any bank are—first, to its depositors; second, to the stockholders; and third, to its borrowers.”  Does the current Fed say anything that clear?

The Federal Reserve Bank of San Francisco agreed: “A bank’s first obligation is to its depositors. No course should be followed which jeopardizes a bank’s ability to pay its depositors according to the agreed terms.  A bank’s second obligation is to its shareholders, to those who have placed their investment funds in charge of the directors and officers”; the borrowers were an also-ran.

Yet the Fed of the time knew that by acting as lender of last resort, which meant (and means today) taking all the best assets of the troubled bank as collateral to protect itself, the central bank was pushing losses to the remaining depositors. “Reserve Banks officials were aware,” Carlson writes, “that lending to support a troubled bank could end up allowing some depositors to withdraw funds while leaving the remaining depositors in a worse position.” I would change that “could end up” to “ends up.”

Carlson continues, “The depositors that did not withdraw would only be repaid from the poorer assets of the bank; that would likely mean that their losses would be worse than if the Federal Reserve had not provided a loan.”  

The Federal Reserve Bank of San Francisco wrote in the 1920s that discount window lending should not “invade the rights of depositors by inequitable preferences to Federal Reserve Banks.” But it inevitably does.

With the advent of national deposit insurance in the 1930s, depositors as claimants on the failed bank’s assets were largely replaced by the Federal Deposit Insurance Corporation. The problem then became the Fed’s shifting losses to the FDIC. This was an important debate at the time of the FDIC Improvement Act of 1991, and it remains an unavoidable dilemma if the Fed can take the best available collateral at any time.

Credit Risk for the Fed

Today, the Fed tries to avoid taking any credit risk. It now gets the U.S. Treasury to take the credit risk as junior to it in various clever designs, like the “variable interest entities” formed for the 2008 bailouts. But we learn from the book that the Federal Reserve Banks actually suffered some credit losses on their Discount Window lending in the 1920s. That meant the borrowing commercial banks failed, and then the Fed’s collateral was insufficient to repay the borrowing. The combined Fed had credit losses of $1.1 million, $1.3 million, and $1.4 million in 1923, 1924, and 1925, respectively.

One hundred years later, in the 2020s, the Fed has zero credit losses but massive losses on interest rate risk. The aggregate losses arising from its interest rate mismatch are $225 billion as of March 27, 2025, and it has a hitherto unimaginable mark to market loss of over $1 trillion.  One wonders what the Federal Reserve officers of the 1920s would have thought of that!

I conclude with three vivid images from the book:

Getting Bank Directors’ Attention

Carlson observes that the Federal Reserve Banks in the 1920s sometimes required personal guarantees from the borrowing bank’s board of directors’ members for Discount Window loans.  When I related this to a friend, who is a successful bank’s Chairman of the Board, he replied, “That would get the directors attention!”  I’m sure it did then and would today.

Mission to Havana

In 1926, there was a bank panic in Cuba involving American banks there, and more paper currency was needed to meet withdrawals. Carlson tells us that the Federal Reserve Bank of Atlanta “scrambled to assemble the cash and ship it to Cuba. … Atlanta assembled a special three-car train with right-of-way privileges to rapidly make the journey from Atlanta through Florida all the way to Key West.  The train left from Atlanta late Saturday afternoon bearing the currency [and] Atlanta staff and guards.  [In] Key West, the money was transferred to the gunboat Cuba…. The gunboat reached Havana harbor at 2:00 a.m. on Monday, whereupon a military guard escorted the currency…[allowing] delivery to the banks before their 9:00 a.m. opening. …and the panic subsided. … Federal Reserve officials received significant praise from the Cuban government.”

The Fed as cattle rancher

“In addition to the losses,” Calson writes, “dealing with the collateral could sometimes cause considerable headaches.” The Federal Reserve Bank of Dallas “ended up owning a substantial amount of cattle after defaults by both banks and ranchers. … Efforts to sell them had a meaningful impact on the local market prices… there were a number of complaints from the local cattlemen’s associations.”  A 1925 memo from Dallas described the “challenges of managing several hundred head of cattle acquired from failing banks in New Mexico.”  This seems a good parting vision for Carlson’s insightful study of the early Federal Reserve wrestling with a systemic agricultural banking crisis.

Overall, The Young Fed is a book well worth reading for students of banking, central banking, and the evolution of financial ideas and institutions.

Alex J. Pollock is a senior fellow at the Mises Institute, the author of Finance and Philosophy—Why We’re Always Surprised, and co-author of Surprised Again! His work is available at alexjpollock.com.