Abstract

We view debt covenants as monitoring technologies that differ in intensity and allocation of control rights, thereby shaping corporate financial behaviour. We construct a novel dataset combining Compustat information with covenant data extracted from EDGAR filings through automated Python-based text analysis over 1995Q3–2020Q3. We classify covenants into maintenance capital covenants, maintenance performance covenants, and covenant-lite incurrence covenants, and firms according to their dominant financing behaviour, distinguishing leverage-adjustment-oriented firms from firms relying more heavily on financial flexibility through internal financing and debt maturity management. To account for the resulting heterogeneity, we estimate firm-level dynamic models and summarize the resulting parameters using a multivariate meta-analytic framework. We find that performance and limitation covenants are associated with greater financial flexibility, whereas capital covenants are associated with faster leverage adjustment. These findings suggest that covenant design shapes not only creditor protection but also firms’ dynamic capital structure choices.