What Does the Minimum Wage Do?

2 - Employment

21 2 Employment Employment—specifically, the number of jobs—is square one for disputes about the minimum wage and its effects. Support for the minimum wage is premised on its improving the lives of those most vulnerable in the labor market. If a minimum wage leads to job loss for many of those same people, serious questions arise with respect to its relative benefits and costs, especially if third parties, such as employers, also bear some of the costs. The disagreement here is not so much whether the minimum wage ever leads to some loss of jobs but whether it always or usually does. If it does not, then legislation setting a minimum wage is not necessarily a bad idea. The NMWR has yet to come to any widely accepted resolution on this issue, and we will not presume that we can settle the matter here. This chapter presents a comprehensive review of research on the effects of a minimum wage on employment, describing the contributions and mutual criticisms of the best-known early protagonists in the NMWR, followed by a detailed look at more recent research. EARLY NEW MINIMUM WAGE RESEARCH The employment effect of the minimum wage is the topic of four of the five (revised) papers from the conference that appeared the following year in the Industrial and Labor Relations Review. Among them, they contained many of the features that were to become common in the NMWR: state-level panels and establishment data rather than national aggregates, quasi experiments in addition to regressions, and ways of gauging minimum wage changes other than the Kaitz index (or similar measures).1 Neumark and Wascher (1992) draw on the framework of the earlier aggregate time-series research on teenagers to study a panel of state-year observations. Card (1992a) performs a differences analysis of state-level data to examine the response of teenage employment to the federal minimum wage increase in 1990; the minimum wage Belman and Wolfson.indb 21 Belman and Wolfson.indb 21 6/16/2014 12:05:11 PM 6/16/2014 12:05:11 PM 22 Belman and Wolfson variable is the fraction affected, the percentage of a state’s teenage employees who were earning less than the higher minimum wage in the three quarters before the increase.2 Katz and Krueger (1992) perform a differences analysis of establishment data to examine the employment response of Texas’s fast food industry to the federal minimum wage increase in 1991; the minimum wage variable is a wage gap measure, how much an establishment’s starting wage would have to rise from its value five months before the increase in order to comply with the new minimum wage. Card (1992b) uses the 1988 increase in California’s minimum wage as an opportunity for a (proto-) difference-in-differences analysis of several employment responses: of teenagers, of the retail sector as a whole, and of one specific part of the retail sector, eating and drinking establishments.3 Only Neumark and Wascher (1992) find a negative employment response, consistent with the earlier literature. A year and a half later, an exchange between Card, Katz, and Krueger (1994) and Neumark and Wascher (1994) appeared. Card, Katz, and Krueger (1994) criticized aspects of Neumark and Wascher’s (1992) analysis, including the reconciliation with Card (1992a,b) that Neumark and Wascher believe shows that Card’s (1992b) specification was unable to detect much of the employment response. Card and Krueger (1994), the most well-known and controversial analysis in the NMWR, appeared almost immediately following this exchange. Building on Katz and Krueger (1992) and Card (1992b), Card and Krueger (1994) use the 1992 increase in New Jersey’s minimum wage to construct a quasi experiment. The treatment group consists of fast food establishments in New Jersey, and the control group consists of similar establishments in nearby Pennsylvania counties. This generated a final exchange of criticism (Neumark and Wascher 2000) and response (Card and Krueger 2000). In describing this work in detail, the focus will be on what has turned out to be most influential: Neumark and Wascher (1992), Card and Krueger (1994), and the exchange that each generated. The Conference: Staking Out Positions Prior to NMWR, studies of the minimum wage and teenage employment in the United States relied on aggregate time-series data or, much less frequently, a cross section of states. Neumark and Wascher (1992) introduce to this literature state-by-year panels based on the CPS Belman and Wolfson.indb...