The high social costs of financial crises imply that economists, policymakers, businesses, and households have a tremendous incentive to understand, and try to prevent them. And yet, so far we have failed to learn how to avoid them. In this article, we take a novel approach to studying financial crises. We first build ten case studies of financial crises that stretch over two millennia, and then consider their salient points of differences and commonalities. We see this as the beginning of developing a useful taxonomy of crises – an understanding of the most important factors that reappear across the many examples, which also allows (as in any taxonomy) some examples to be more similar to each other than others. From the perspective of our review of the ten crises, we consider the question of why it has proven so difficult to learn from past crises to avoid future ones.

Excessively easy bank credit – visible in unusually small credit risk spreads and rapid loan growth – is often posited as a root cause of unsustainable asset price booms. This paper considers whether an increase in bank risk tolerance drove high loan growth that coincided with Florida's land boom of the mid-1920s, the first Florida housing boom in which buyers from around the nation participated. Estimates suggest that an astounding 20 million lots were offered for sale in Florida at that time. Our detailed narrative and empirical evidence suggest that the facts do not require the assumption of increased risk appetite during the boom. We find that most Florida banks that failed were associated with the Manley-Anthony chain and did not exhibit increases in observable indicators of risk during the boom. Instead, their increases in risk mainly reflected hidden choices either to lend to bank insiders on a preferential basis or to fund other banks that were engaged in such risky and often fraudulent activities. Bank regulators seem to have been complicit in the hidden risk-taking. Even informed investors would have been left in the dark about the amount of risk that was growing in Florida.

After an unprecedented number of banks suspended operations in the during Panic of 1893, the head regulator of banks chartered by the United States government allowed about 100 banks to reopen after certifying their solvency. We evaluate whether actions by bank owners to change management, contract with depositors to extend liability maturity structure, write off bad assets, and/or inject capital affected bank survival and deposit retention. This historical episode is particularly informative because there was no expectation of government intervention. We find that contracting with depositors provided short-term benefits while dealing with bad assets was key for long-run viability.

How should sovereign bailouts take account of the effects bailouts have on policy reforms? Conflicted recipient governments complicate bailout choices because some reforms that spur growth reduce rents that benefit government decision makers. Our model takes account of whether bailout generosity and policy reforms are strategic substitutes, strategic complements or both, and each case implies a different optimal bailout contract, which generally cannot achieve First Best. Conditional forgiveness of some loan payments when economic outcomes are sufficiently favorable can achieve outcomes closer to First Best, and this is so for a small ex ante amount of the bailout subsidy.

We construct a new measure of deposit insurance generosity for many countries, empirically model the exogenous international influences on the adoption and generosity of deposit insurance and use a novel econometric method to explore the causal chain from the expansion of deposit insurance generosity to increased overall lending, increased lending to households, increased banking system leverage, and more severe and frequent banking crises. Greater deposit insurance generosity robustly produces greater overall lending relative to bank assets and more lending to households relative to both bank assets and GDP, and results in higher banking system leverage. Our estimates, however, are not conclusive regarding whether greater deposit insurance generosity resulted in greater total loans relative to GDP or in more frequent or severe banking crises.

We examine changes in banks’ market-to-book ratios over the last decade, focusing on the dramatic and persistent declines witnessed during the financial crisis. The extent of the decline and its persistence cannot be explained by the delayed recognition of losses on existing financial instruments. Rather, it is declines in the values of intangibles – including customer relationships and other intangibles related to business opportunities – along with unrecognized contingent obligations that account for most of the persistent decline in market-to-book ratios. These shifts reflect a combination of changed economic circumstances (e.g., low interest rates reduce the value of core deposits; meager growth opportunities reduce the value of customer relationships) and changed regulatory policies. Together, these changes in the business environment since the financial crisis have led investors to associate little value with intangibles. For example, changing market perceptions of the consequences of leverage have affected the way investors value banks; prior to the crisis, higher leverage, ceteris paribus, was associated with greater value (reflecting the high relative cost of equity finance), but during and after the crisis, as default risk and regulatory concerns came to the fore, lower leverage was associated with greater value. Reflecting the rising importance of regulatory risks (e.g., the uncertain consequences of the Volcker Rule), after controlling for other influences, dividend payments (a signal of management and regulatory perceptions of the persistence of financial strength) matter for market prices much more after the crisis, while increases in recurring fee income matter less.

This paper offers a historical perspective on the evolution of central banks as lenders of last resort (LOLR). Countries differ in the statutory powers of the LOLR, which is the outcome of a political bargain. Collateralized LOLR lending as envisioned by Bagehot (1873) requires five key legal and institutional preconditions, all of which required political agreement. LOLR mechanisms evolved to include more than collateralized lending. LOLRs established prior to World War II, with few exceptions, followed policies that can be broadly characterized as implementing “Bagehot's Principles”: seeking to preserve systemic financial stability rather than preventing the failure of particular banks, and limiting the amount of risk absorbed by the LOLR as much as possible when providing financial assistance. After World War II, and especially after the 1970s, generous deposit insurance and ad hoc bank bailouts became the norm. The focus of bank safety net policy changed from targeting systemic stability to preventing depositor loss and the failure of banks. Statutory powers of central banks do not change much over time, or correlate with country characteristics, instead reflecting idiosyncratic political histories.

Liquidity shocks transmitted through interbank connections contributed to bank distress during the Great Depression. New data on interbank connections reveal that banks were vulnerable to closures of their correspondents and their respondents. Further, banks were less responsive to network liquidity risk in their management of cash and capital buffers after the Federal Reserve was established, suggesting that banks expected the Fed to reduce that risk. The Fed's presence weakened incentives for the most systemically important banks to maintain capital and cash buffers against liquidity risk, and thereby likely contributed to the banking system's vulnerability to contagion during the Depression.