The Employment Effect in Retail Trade of California ’ s 1988 Minimum Wage Increase Lowell

In this paper, we study the outcome of an unusually clean natural experiment--California’s large minimum wage increase of 1988. Two different approaches to evaluating the experiment result in the same conclusion: the textbook analysis of minimum wages holds true. In particular, we find that employment growth in California’s low-wage retail trade industry was slowed by the minimum wage increase. The Employment Effect in Retail Trade of California’s 1988 Minimum Wage Increase Since the passage of the Fair Labor Standards Act of 1938, minimum wage laws have represented one of our country’s most prominent labor market interventions. Subsequent to the establishment of the Act, a number of amendments have raised the federal minimum. Most recently, after almost a decade with no changes, the federal minimum increased from $3.35 to $3.80 on April 1, 1990, and one year later to $4.25. There appears to be broad support in the new Democratic administration for raising it still further. Some favor rather substantial increases. For instance, one proposal, forwarded by House Budget Committee Chairman Martin Olav Sabo, would boost the minimum wage to $6.50 an hour. Many policymakers are enthusiastic about increasing the minimum wage because they feel it will raise the earnings of low-wage workers while not greatly damaging their employment prospects. This view has been bolstered recently by a series of interesting and compelling empirical studies which call into question economists’ conventional prediction--that an increase in the minimum wage moves equilibrium backward along the demand curve for labor, reducing the employment of low-wage workers. Each of these studies suggests that the most recent round of increases in the minimum wage in the United States in fact had little effect on employment among low-wage workers. One study, by David Card (1992a), explored the consequences for teenage employment of California’s minimum wage increase of July 1988. Card compared changes in teenage employment in California and in a group of southwestern and southern states from 1987 to 1989. Although the minimum increased in California during this period while remaining fixed in the other states, no impact was found on California’s teenage employment. (Indeed the effect seemed to be the "wrong way": employment increased.) In a second study, Card (1992b) exploited regional variation in wages across the United States and found little evidence that recent increases in the minimum wage reduced 2 low-wage employment. Lawrence Katz and Alan Krueger (1992) and Card and Krueger (1993), in studies of the fast-food industry, also detected little change in employment due to increases in the minimum wage. As the authors of these papers have suggested, the finding that minimum wage increases do not adversely affect employment indicates that economists should reevaluate the standard welfare analysis of minimum wage laws. The result also may indicate that low-wage labor markets are not well described by conventional models, and economists should consider alternatives to textbook theory in analyzing these labor markets. Because the most recent wave of empirical studies is important and perhaps somewhat controversial, and because much of the previous literature is equivocal, we have undertaken in this paper a new empirical evaluation of a recent increase in the minimum wage--California’s substantial increase (from $3.35 to $4.25) in July 1988. The increase of the minimum wage in California provides an especially clean "natural experiment" for evaluating the effects of a wage floor. We can think of California as a treatment group, and the rest of the nation as a control. Our focus is the effect of the wage change on employment in retail trade, a sector that employs a disproportionate number of low-wage workers. In particular, we seek to discover if the increase in the minimum wage resulted in systematic industry variation in employment growth and wage changes within the retail trade industry. Using data from County Business Patterns (U.S. Department of Commerce, various years), we find evidence that is not at all ambiguous or sensitive to specification. This evidence is consistent with conventional theory. In evaluating the period March 1988–March 1989, we show that within retail trade, in those industries where the wage increased most rapidly in California relative to the rest of the country, relative employment growth fell. We also pursue an alternative approach in studying the effect of the minimum wage increase. We note that within California there was substantial intercounty variation prior to the 1988 minimum 3 wage increase. Retail trade wages in some (primarily urban) counties were rather high; in these counties an increase in the minimum could have little impact. Retail trade employees were initially paid substantially less in other (often rural) counties. Conventional theory suggests that employment would be adversely affected in these latter counties. Our empirical results are in accord with this prediction. Interestingly enough, the magnitude of the effect of the minimum wage on employment growth estimated in our regressions is quite similar in the two approaches we take. 1. RETAIL TRADE IN CALIFORNIA AND THE UNITED STATES, 1986–1989 Our study focuses on employment in the retail trade industry. Studies concentrating on other industries may also prove useful, but the retail trade sector is of particular interest. Because wages paid by many employers in this industry are low, the minimum wage is particularly likely to present a binding constraint in this industry. Also, the retail trade industry is by far the largest employer of low-wage workers. Using CPS data, Card (1992a) estimated that in 1987 over 30 percent of California’s workers in retail trade earned at least $3.35 per hour but less than the new minimum, $4.25. Moreover, nearly half of all workers who earned $3.35–$4.24 were employed in retail trade. The years 1986 to 1989 represent an excellent period for investigating the independent effect of the imposition of a new minimum wage. During this time, inflation was very low, increasing only slightly toward the end of 1988 and into 1989. Overall unemployment in the mid to late 1980s was gradually declining as the national economy continued to recover from the recession of the early 1980s. Furthermore, rates of unemployment were similar in California and the United States as a whole. In 1988 and 1989 the unemployment rates in the United States were, respectively, 5.5 and 5.3. In California, the corresponding figures were 5.3 and 5.1. 4 We use data from the County Business Patterns (CBP) issued by the Bureau of the Census to track trends in employment and pay. These data indicate firms’ payrolls (based on the Employer’s Quarterly Federal Tax Return for the first quarter of the year, and the firms’ total employment for a pay period including March 12. In the retail trade industry, trends in employment were quite similar in California and the United States as a whole over the 1986–1989 period. Table 1 presents a summary. The third row of this table shows that growth in retail trade employment was not much different in California than in the rest of the United States. Employment grew slightly less rapidly in California in 1986–87 and somewhat more rapidly in 1987–88. The 1988–89 employment trend, which reflects any effect of the 1988 increase in California’s minimum wage, shows employment growing a little more slowly in California than in the United States. Table 1 also presents statistics for the reported pay per person in the retail trade industry. Most notable is the fact that wages grew much more rapidly in California than in the rest of the United States from the first quarter of 1988 to the first quarter of 1989. A natural explanation for this outcome is of course that the minimum wage increased in California (from $3.35 to $4.25) in July of 1988, while the wage floor remained constant in nearly all other states. These observations are consistent with Card’s (1992a) finding. While wages in retail trade did increase more in California than in the rest of the United States in 1988–89, presumably as a result of the new minimum wage law, employment was hardly affected. Growth in retail trade employment was a bit slower in California, but the magnitude of the deviation from the U.S. trend was small and no larger (in absolute terms) than in the previous two years. It is hard to make a good argument based on the data in Table 1, however, because they are aggregate data and thus may smooth over any systematic employment effects that may exist. 5 TABLE 1 Percentage Change in Employment and Wages in the Retail Trade Industry Percentage Change in Employment Percentage Change in Wages 1986–1987 1987–1988 1988–1989 1986–1987 1987–1988 1988–1989 California 4.46 2.95 2.48 1.28 2.75 6.15 United States 5.00 1.98 2.88 0.99 3.62 3.83 Difference between percentages -0.54 0.97 -0.40 0.29 -0.87 2.32 Source: County Business Patterns. 6 Therefore we looked at disaggregated data on sectors within the retail industry to see if there was more to these findings than meets the eye. Table 2 presents statistics similar to those given in the bottom row of Table 1, but for 2-digit SIC industry classifications within retail trade. Although the differences regarding wages in each sector vary considerably, the picture is quite consistent. In nearly every sector, notably Eating & Drinking places and General Merchandise stores, wages increased in California at a higher rate than in the country as a whole; in only one sector did wages rise more slowly in California. The figures concerning employment are more mixed: in three sectors, employment grew at a slower rate in California than in the United States, while in the other four sectors, employment increased at a faster pace. Moreover, the variation appears to be systematic: in industries in California that experienced the greatest relative growth in wages, employment suffered the greatest relative declines, and vice versa. This observation casts some doubt on the argument that the minimum wage increase had no effect on retail trade in California, and suggests that further investigation is warranted. 2. INTERINDUSTRY EFFECTS OF THE MINIMUM WAGE In the textbook model of the labor market, the relationship between wages and employment is simple: wages and employment are jointly determined by the supply of labor and labor demand. A regression of the form where ∆Ei and ∆wi are changes in employment and wages, respectively, within an industry, therefore makes little sense. Employment is related to wages, but any changes in employment are the result of 7 TABLE 2 Differences between the Percentage Growth in Wages and Employment in California and the United States: Retail Trade Industry, 1988–1989 Pay per Employee Employment General Merchandise 6.86 -6.16 Eating & Drinking 4.63 -1.24 Food Stores 2.36 -0.37 Apparel & Accessory 2.31 1.19 Building & Garden Supplies 0.77 3.15 Furniture 0.00 3.87 Auto Dealer & Service Stations -1.56 2.02 Source: County Business Patterns. Notes: Column 1 gives the percentage change in pay per person in California minus the percentage change in pay per person in the United States, while column 2 is the difference in the percentage change in employment in California and the United States. These categories represent all retail trade industries except "Miscellaneous Retail." 8 a shift in the supply curve or the demand curve (or both), and our regression will not distinguish between these. For a year in which there is an exogenous shock to the wage, however, there may be some hope of identifying a demand curve for labor. Suppose, for instance, that changes in employment in a specific industry i in California are due to movements along a conventional employment demand curve and possibly shifts in the demand curve as well. Assume that this can be represented by