This paper presents a micro data approach to the identification of credit crunches. Using a survey among German firms which regularly queries the firms' assessment of the current willingness of banks to extend credit, we estimate the probability of a restrictive loan supply policy by time taking into account the creditworthiness of borrowers. Creditworthiness is approximated by firm-specific factors, e.g. the firms' assessment of their current business situation and their business expectations. After controlling for the return on the banks' risk-free investment alternative, which is also likely to affect the supply of loans, we derive a credit crunch indicator, which measures that part of the shift in the loan supply that is neither explained by firm-specific factors nor by the opportunity costs of providing risky loans.JEL classifications: C23, E44, E51, G21