Essays in financial economics and industrial organization: An examination of the statistical power of the test of CAPM, copyleft: An R&D game, experimentation in the IPO market.

摘要

This thesis examines some issues in financial economics and industrial organization. Chapter 1 looks at the two-pass empirical test of CAPM incorporating time-varying betas and different magnitudes of the variance of the market model residues. We conduct a simulation assuming an ideal efficient market which CAPM depicts, then run the two-pass test to check it's statistical power. It turns out that the power of the test is low due to the short duration of available market data and the low mean of the market premium, in addition to the EIV problem.;Chapter 2 explores the question of why firms go "copyleft" in a market with network externalities. "Copyleft" means a low or negative license fee for a firm's technology. We consider a market with an incumbent and one potential entrant in a 2-stage game. In the 1st stage, the two firms compete for downstream clients to buy their licenses and develop complementary products. In the 2nd stage, they sell to the consumers. The incumbent makes its monopoly profits when its network advantage is very large in both stages. The incumbent is only able to deter entry by going copyleft when network advantage is small in the first stage. The incumbent can only restrict the quality of the entrant's product to zero when its network advantage in the second stage is small. When both stages' network advantage is very small, both firms produce a positive amount.;Chapter 3 looks at the underwriter's role of deciding what firms to bring to the IPO market. We explicitly model serial correlation in the IPO prices in one industry as a reflection of investor sentiment. We then study the effects of correlation and underpricing on the underwriter's intertemporal decision. We find that more firms in the correlated industry tend to go public relative to firms in an uncorrelated industry. Moreover, even if the firms in the correlated sector are at a valuation disadvantage, as long as there is enough underpricing, the underwriter still inefficiently experiments with these firms.