Essays in Labor Economics

Abstract

My dissertation consists of three chapters in the area of labor economics. My first chapter looks into the puzzling decline in U.S. women's labor force participation after 2000. The second chapter examines the cross-border spillover effects of increases in state minimum wages. The third chapter studies dynamic employment responses to state-level minimum wage increases. My first chapter examines the recent declining trend in U.S. women's labor force participation. For many decades, the labor force participation rate (LFPR) of U.S. women rose along with their real wages. The prevailing economic interpretation was that the rise in women's labor force participation constituted a movement along a positively sloped labor supply function. By the late 1990s, however, women's LFPR had stopped increasing and, from 2000 to 2015, it declined from 61% to 57%. This reversal of trend is surprising over a period when women's real wages were commonly believed to be rising. This project explores this puzzle by taking a more detailed look at the trends in both labor force participation and real wages for U.S. women. First, by updating the evidence on women's LFPR to 2015, I find that almost two-thirds of the four-percentage-point drop in the LFPR is attributable to the aging of the adult female population. The other third of the decline comes from within-group decreases in labor force participation for those under the age of 55. The latter finding begs a question not answered in the existing literature: Why have women under age 55 reduced their labor force participation at a time when their real wages are still increasing? My most important finding, based on evidence from both the Current Population Survey and the Panel Study of Income Dynamics, is that the apparent real wage increase is more than accounted for by the substantial increase in women's educational attainment. My analysis that deflates wages by the CPI-U-RS suggests that, had it not been for the rise in women's educational attainment, their real wages would have gone down by about 6% since 2000. In that sense, women's real wage opportunities over this period may have worsened considerably (though not by as much as men's), so that the decline in women's LFPR is not so puzzling after all. My second chapter focuses on the cross-border spillover effects arising from states raising their minimum wages. Many studies that estimate employment effects of state increases in minimum wages do so by comparing employment developments between states where minimum wages increased and states where minimum wages did not change. Simple economic theories, however, indicate that employment developments in the latter “control” states might be “contaminated” by spillovers from neighboring “treatment” states. In that case, estimates that ignore spillovers into control states could be biased. This paper studies the spillover issue at the county level using the Quarterly Census of Employment and Wages between 1990 and 2014. I find that the spillover effects are modest, and that accounting for them has little impact on estimated disemployment effects. While many other methodological issues remain to be resolved in the minimum wage literature, bias from spillover effects need not be one of them. The third chapter looks into employers' dynamic responses to increases in state minimum wages. One concern raised in the debate about disemployment effects of minimum wages is an apparent tendency for employment to decline before the minimum wage is increased. Existing literature often chooses to control for these leading effects in arbitrary ways, such as including state-specific linear (or higher order) trends. This paper reconsiders the matter by testing if the pre-existing trends is resulted from anticipation. Most states legislate and announce their minimum wage changes a few quarters before they go into effect, and it also has become common for states to schedule their minimum wages for the next couple of years. If employers believe that these changes are permanent, and if the technology does involve adjustment costs, employers may adjust their production, particularly employment, before the minimum wage is implemented. In that case, lead effects are part of the genuine impact of minimum wage increases, not a pre-existing trend that ought to be controlled out. To test this, I adopt Sargent's classic model of dynamic labor demand with adjustment costs to the context of state minimum wage increases using the Quarterly Census of Employment and Wages over the 1990-2014 period. I find that the anticipation effect is significantly negative and has almost the same size as the current minimum wage effect. After accounting for anticipation, the long-run disemployment effect increases from -0.259 to as large as -0.311, and failing to consider anticipation biases the estimated disemployment effect by about 12%.