Abstract:Since an underwriter sets an IPO's offer price without knowing its market value,
investors can acquire information about its value and avoid overpriced deals ("lemon-
dodge"). To mitigate this well-known risk, the bank enters into a repeat game with
a coalition of investors who do not lemon-dodge in exchange for on-average under-
priced shares. We  derive and test a quantitative IPO pricing rule (showing that
tech IPOs were not excessively underpriced during the boom of the 1990s); and
analyzing a sunique multibank data set, find strong upport for the conjecture that a
bank preferentially allocates shares to its coalition.