N°26-18: Information Processing Costs and Nonbank Lending: Evidence from Machine-Readable Disclosures

We study how lowering the cost of processing public financial information—holding information content fixed—reshapes lender composition in syndicated loans. Using the SEC’s staggered adoption of eXtensible Business Reporting Language as a plausibly exogenous reduction in processing costs, we show that nonbanks increase participation and form new lending relationships, with no comparable effects for banks. The participation increase is concentrated among lenders lacking industry expertise. Following adoption, nonbanks substitute away from relationship-based signals toward standardized financial metrics, while their role in delegated monitoring remains unchanged. Consistent with machine readability as the mechanism, pre-loan automated EDGAR activity predicts greater nonbank participation.