What accounts for the differences in the performance of cooperatively-owned banks in the Dutch financial crisis of the early 1920s? This thesis measures and explains the (relative) performance of boerenleenbanken (rural Raiffeisen banks) and middenstandsbanken (urban Schulze-Delitzsch banks) during the Netherlands' interwar banking crisis by applying various economic methods to new historical evidence. The thesis asks: (1) what were the effects on risk-taking behaviour of differences in the religious attitudes of bankers and their customers? (2) what was the relationship between interbank competition and financial stability? and (3) what was the consequence of the liability choices made by shareholders for their banks' continued survival? Using a combination of economic theory, quantitative financial analysis and qualitative business histories, this thesis finds that: (1) banks serving small religious groups were less willing, despite being more able, to take on risks than those serving majority denominations; (2) those banks that were subject to the lowest competitive pressures enjoyed the most liquid investment portfolios; and (3) the choice of liability limitation available to bankers in uenced their balance sheet risks, for the worse. Together, these findings lead to the conclusion that social, organisational and institutional factors each explain part of the heterogeneity in the fate of the Netherlands' cooperative banks during a period which includes unprecedented debt- deflationary financial turmoil: hence, (1) strict membership criteria and the use of personal guarantors in loan agreements acted as strong devices to allow banks for minorities, regardless of their denomination, to screen and monitor their customers; (2) the switching costs associated with religious affiliation resulted in a competition- stability tradeoff during periods of extreme distress; and (3) the stakeholders of the banks which failed were probably less risk-averse than those of banks which did not, the consequence of endogenous group formation by risk type.