Author: David Demilly (Banque de France, CRED & Paris-Panthéon-Assas University)
Economic historians often describe the 1920s in France as a period of remarkable banking stability. This new article argues that this calm was, at least in part, an illusion. Drawing on previously unexploited archival data from Crédit Lyonnais, the Banque de France and French commercial courts, the paper shows that the banks which did fail between 1918 and 1928 were, on average, better capitalised than those that survived. High capital ratios created a false sense of security and, paradoxically, encouraged the very risks that would bring those banks down.
A puzzling banking landscape in interwar France
The 1920s were turbulent years for the French economy. Three successive currency crises, the depreciation of the franc, speculation against public debt, and the eventual stabilisation under Raymond Poincaré reshaped the financial landscape. Yet, bank failures were rare during 1920s. This apparent calm has long led historians to treat the decade as a period of consolidation. A closer look at the balance sheets of 277 French banks — collected by Crédit Lyonnais in its remarkable “Albums” — tells a different story. Behind the aggregate stability, a small but telling share of institutions had engaged in risky, illiquid activities that their capital buffers could not protect.
A counterintuitive finding: failed banks were better capitalised
Standard banking theory predicts that better-capitalised banks are less prone to fail. Shareholders with more skin in the game have stronger incentives to limit risk, and equity absorbs losses before they threaten depositors. The 1920s French evidence points the other way. Failed banks held capital ratios of roughly 30 percent of their balance sheets, against 20 percent for surviving banks. Cox proportional-hazards and logit regressions confirm the pattern: both subscribed capital and paid-up capital are positively associated with the probability of failure, even after controlling for other balance-sheet characteristics and for the macroeconomic environment.
By contrast, liquidity — measured by banks’ holdings of cash and central-bank reserves — is strongly associated with survival. Banks that kept a larger share of their assets in immediately available liquid resources were far better placed to meet withdrawal requests when shocks materialised. In the French context of the 1920s, capital mattered less than the composition of a bank’s assets.
The illusion: how capital encouraged risk-taking
Why would more capital lead to more failures? Qualitative archival evidence — from commercial-court failure files and from Banque de France supervisory reports, largely unexplored until now — helps reconstruct the mechanism. In the absence of modern capital requirements linking equity to risk-weighted exposures, high capital ratios were interpreted as signals of solidity. They made it easier for banks to attract depositors, reassure counterparties, and raise further funds from shareholders. Armed with these resources, some banks drifted away from the traditional French model — financing public debt through short-term, liquid instruments — and towards speculative securities, equity participations, long-term loans to a narrow set of clients, or real-estate ventures.
Three recurring patterns of failure emerge from the archival case studies. Some banks over-concentrated their lending on a handful of insolvent counterparties, as with the Caisse Générale d’Escompte et de Crédit or the Banque Lafont. Others bet heavily on a single industrial venture — the Banque Certes et Marty devoted most of its resources to the Figeac coal mines — before being unable to meet repayment demands when cash flows disappointed. A final group, including the Moniteur Financier and the Banque Privée, crossed into outright fraud. In most cases, the trigger was not insolvency in a narrow sense but liquidity: when losses materialised or reimbursement requests intensified, illiquid balance sheets gave way.
The feeling of stability conveyed by high capital was, in this sense, illusory. Capital provided a buffer against credit losses, but it did not shield banks from the liquidity risks that came with illiquid or speculative investments.
What 1920s France tells us about banking stability today
These findings speak directly to contemporary debates on banking regulation. Much of the post-2008 reform agenda has emphasised higher and better-quality capital, most notably through Basel III. The French experience of the 1920s suggests that capital alone is not enough — and can, under some institutional conditions, actively encourage risk-taking. It is the joint requirement of capital and liquidity ratios, such as the Liquidity Coverage Ratio and the Net Stable Funding Ratio introduced by Basel III, that provides a more robust framework.
More broadly, the paper is a reminder that banking stability depends on the “rules of the game” — the interaction between balance-sheet structures, the macroeconomic environment, and the regulatory framework. In 1920s France, that framework left banks free to choose the composition of their balance sheets, and a context of monetary instability and speculative opportunities did the rest. The majority of banks stayed safe by holding public debt; a minority were seduced by seemingly stable, but ultimately illusory, strategies. The illusion of stability belongs to the 1920s — but the lessons are not confined to the past.