Competition or collaboration? The reciprocity effect in loan syndication

It is well recognized that loan syndication generates a moral hazard problem by diluting the lead arranger's incentive to monitor the borrower. This paper proposes and tests a novel view that reciprocal arrangements among lead arrangers serve as an effective mechanism to mitigate this agency problem. Lender arrangements in about seven out of ten syndicated loans are reciprocal in the sense that lead arrangers also participate in loans that are led by their participant lenders. I develop a model in which syndicate lenders share reciprocity through such arrangements in a repeated-game setting as monitoring effort enhances lead arrangers' ability to profit from participating in loans led by others. The model generates specific predictions that I then confront with the data. I find strong and consistent empirical evidence on the reciprocity effect. Controlling for lender, borrower, and loan characteristics, I show that: (i) lead arrangers retain on average 4.3% less of the loans with reciprocity than those without reciprocity, (ii) the average interest spread over LIBOR on drawn funds is 11 basis points lower on loans with reciprocity, and (iii) the default probability is 4.7% lower among loans with reciprocity. These results indicate a cooperative equilibrium in loan syndication and have important implications to lending institutions, borrowing firms, and regulators.