Evaluating the Magnitude of Financial Market Frictions: Evidence from Firm-Level Data

We investigate the extent to which firm-level data are consistent with the microeconomic foundations of the benchmark financial accelerator model of Bernanke, Gertler, and Gilchrist (1999). To that purpose, we construct a new dataset that directly links firm-specific balance sheet variables to credit spreads on publicly-traded debt. The estimated leverage-spread schedule exhibits statistically significant nonlinearities that are consistent with the theoretical predictions of the model. We then use the observed data on financial leverage, spreads, and market-based measures of default risk to solve for key structural parameters of the model. Our results indicate that a substantial degree of financial market frictions is necessary to match the default probabilities implied by the model with market-based measures of default risk. Moreover, we quantify the magnitude of the model-implied external finance premium, show how it is related to various firm characteristics, and examine its behavior during the most recent economic downturn. Finally, we identify several directions in which the basic framework needs to be extended to account for other cyclical and cross-sectional features of the data