This dissertation investigates the effect of endogenous and exogenous events on firm behavior and performance. These are fundamental questions in economics. The contribution of this study is threefold. First, it provides estimates of the impact of mergers on railroad efficiency, which has important antitrust implications. Second, it provides new estimates of the effect of negative events on the market value of Johnson & Johnson, Bridgestone, and Toyota, which is important to the understanding of how markets punish corporate errors. Third, it develops better ways to estimate these effects. Chapter 2 uses the event study approach to determine how product recalls due to exogenous and endogenous shocks affect the value of the firm. Three recalls from Johnson & Johnson, Bridgestone, and Toyota have been studied in this chapter. The traditional event study method assumes that markets are efficient, a questionable assumption in the short run. Thus, the current stock value of a firm may not reflect its true market value. To address this potential problem, frontier based methods are used, including data envelopment analysis, corrected ordinary least squares, and stochastic frontier regression analysis. Stochastic frontier methods are shown to be more appropriate when market behavior is not fully rational. The evidence shows that endogenous events due to firm errors are more detrimental to firm value than exogenous negative events that are beyond the control of the firm. That is, the market is more forgiving of negative shocks that the company cannot control. Chapter 3 studies the effects of merger activity on the efficiency and productivity growth of U.S. Class I railroads from 1983 to 2008. In this chapter, I assess the effects of merger activity on efficiency, and identify the major factors associated with productivity growth. Unlike previous research, I use data envelopment analysis with an attribute-incorporated Malmquist productivity index. This approach allows firm specific measures of efficiency and productivity to be calculated for firms with differences in technology. The approach allows a decomposition of the attribute-incorporated Malmquist productivity index into technical, efficient and attribute components, the impacts of railroads mergers, and the real source and change of productivity. I find that (1) the technology efficiency performance of the seven survivor firms has grown through time; (2) mergers overall do not lead significant technology and scale efficiency gains, but there are differences across mergers; (3) mergers in the 1980s do not have significant different effect on efficiency change compared to those in the 1990s; and (4) the productivity gains are mostly attributed to the network and operation attributes change and industry technology improvement. Overall, the mergers have no direct impact on the efficiency gains or losses during our study period. The application of these techniques to product recalls and railroad merger models demonstrates how they can provide superior estimates over traditional estimation techniques. It is hoped that these applications will motivate the use of these techniques in other settings. / Graduation date: 2012