Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper explores group size in joint liability lending, primarily in the adverse selection framework with local borrower information. A single, standardized contract that imposes full joint liability subject to a limited liability cap is optimal. Further, if gross returns to borrowing are moderately high, this contract results in perfectly efficient lending if groups are large enough. However, raising group size accomplishes nothing if there is no local borrower information. These results show that more is required for efficient lending than full within-group insurance, and highlight a complementarity between group size and social capital. Very similar results are shown in two different settings, ex ante and ex post moral hazard, though the type of social capital that complements group size varies across the settings. Taking a step toward modeling drawbacks of larger groups, it is shown that if information deteriorates sufficiently with group size, an intermediate group size does better than either extreme. Simulations suggest that most of the efficiency gains from larger groups are realized in group sizes below ten, and that outreach and efficiency can increase dramatically when a moderate group size threshold is crossed.