Measuring the systemic risk in interfirm transaction networks

Using a unique and massive dataset that contains information on interfirm transaction relationships, this study examines default propagation in trade credit networks and provides direct and systematic evidence of the existence and relevance of such default propagation. Not only do we implement simulations in order to detect prospective defaulters, we also estimate the probabilities of actual firm bankruptcies and compare the predicted defaults and actual defaults. We find, first, that an economically sizable number of firms are predicted to fail when their customers default on their trade debt. Second, these prospective defaulters are indeed more likely to go bankrupt than other firms. Third, firms that have abundant external sources of financing or whose transaction partners have such abundant sources are less likely to go bankrupt even when they are predicted to default. This provides evidence for the existence and relevance of firms – called “deep pockets” by Kiyotaki and Moore (1997) – that can act as shock absorbers.