Date of Award
Doctor of Philosophy (PhD)
Mroz, Thomas A
Maloney , Michael T
Thomas , Charles J
Wilson , Paul W
This dissertation encompasses two papers. The first paper examines the impact of measurement errors in potential experience and reported education on Black-White wage differentials. I show that measurement error is not mean zero, is distributed differently for Black and White males and, for experience, is correlated with the value of the variable measured with error. Possible conditional distributions of the true values of education and experience, given reported values, are estimated using auxiliary data. This paper introduces a Maximum Likelihood method to deal with these errors, and evaluates this method as well as other methods currently used in the gender wage differential literature. The use of the Maximum Likelihood estimation method along with traditional Multi-Sample Two-Stage Least Squares reveals that a significant portion of the estimated Black-White wage differential in a classic Mincer-style regression is due to measurement error in reported educational attainment and (especially) potential experience. Use of predicted or probabilistic measures in lieu of reported education and potential experience reduce the estimated racial wage gap in the 2000 Census from 34 percent to less than 20 percent.
In the second paper, I examine how the introduction of competition from Southwest Airlines affects airfares in a variety of market structures. While the consensus of the airline literature is that entry by Southwest results in substantially reduced fares on the entered and nearby routes (known as the Southwest Effect), little attention has been paid to the differing effects across routes. This paper fills this hole in the literature in two ways. First, I use difference-in-difference estimates to determine the causal effect of Southwest entry on fares using a natural experiment (the repeal of the Wright Amendment) that allowed for competition from Southwest on routes where such competition was previously illegal. I show that, consistent with the literature, the average effect on fares of competition from Southwest is substantial. However, the per-route effect varies substantially, from a fall in fares of roughly 40\% to a slight increase in fares. A fixed-effects regression centered around the repeal uncovers some of the factors behind this difference. Specifically, the presence of existing low-cost carriers and the presence of the ticketing airline in the origin and destination airports are the most important factors behind the ``Southwest Effect,'' while the route Herfindahl and ticketing airline's share of passengers on the route matter little. These results illustrate the hazards of using a small-scale case-study approach to estimating the fare effects of entry by Southwest, as well as a need for a deeper understanding behind the mechanics of the Southwest Effect.
Huff, Jared, "Measurement Error and the Black-White Wage Differential" (2012). All Dissertations. 949.