A Sage Policy Group study projects that raising DC’s minimum wage to $12.50 an hour for large retailers might cost the city jobs. However, the report is based on a misrepresentation of an earlier study on DC’s 1993 minimum wage increase by John Schmitt and David Rosnick of the Center for Economic and Policy Research.
In Schmitt and Rosnick’s study, they found that city-level minimum wage increases led to higher pay for workers without job losses in San Francisco and Santa Fe. The data for DC, the third city in their study, were less clear, however. This is because the data did not show that DC’s 1993 minimum wage increase had led to higher wages, for a variety of reasons. The study noted that the data for DC did not “allow us to draw conclusions about the employment effects of binding citywide minimum wages.” Yet Sage Policy Group cherry-picked one piece of data from the Schmitt and Rosnick report to come to a conclusion that the authors had not made.
John Schmitt issued a statement today in which he concludes “unequivocally that Sage Policy Group has misrepresented our findings.” The full statement is reprinted below:
Statement of John Schmitt, Economist, Center for Economic Policy Research
In Response to the Sage Policy Group Report on the DC Large Retailer Accountability Act
June 17, 2013
In a March 2013 report, the Sage Policy Group projects that the Large Retailer Accountability Act would have a negative impact on jobs in the District of Columbia. Their conclusion is based on recent research that I conducted with David Rosnick for the Center for Economic and Policy Research. As a co-author of that research, I can state unequivocally that Sage Policy Group has misrepresented our findings. The authors of the Sage Policy Group report appear to have made basic analytical and economic errors to arrive at conclusions completely at odds with the findings based on an analysis of the firm-level data presented in our original research.
The Sage Policy Group selectively and incorrectly used one part of our study to assert that the District’s 1993 increase of $1 per hour in the minimum wage had a negative impact on employment in the retail sector. In fact, as we stated plainly in our report, the data do not support that conclusion.
The Sage Policy Group’s erroneous conclusion is based on a highly flawed understanding of basic econometric analysis. In our report, we examined the impact of city-wide minimum-wage increases in the three cities with such laws at the time that we conducted our study (Washington, DC, Santa Fe, NM, and San Francisco, CA). The first step of any analysis of the impact of a minimum-wage increase is to determine whether the new minimum wage actually increased wages of low-wage workers. If wages did increase after implementing the new minimum wage, then, and only then, can we measure firms’ employment response to the increase. If, however, the minimum-wage increase was too low to raise wages, or if employers failed to pay the higher wage, then we cannot draw any conclusions about the impact of that particular minimum wage.
In Santa Fe and San Francisco, we found strong evidence that the minimum-wage did indeed raise wages in many industries that typically employ low-wage workers. In Washington, DC, however, we found that the smaller increase in the minimum wage implemented two decades ago had no discernible effect on wages in low-wage industries operating in the District. We presented evidence in our report that wages in the District at the time were already higher than the national average and therefore not much affected by the small increase. We also presented some evidence that suggested that employers may not have complied with the law.
In either case, the data are unequivocal that the small minimum-wage increase in 1993 did not raise the wages of workers in the retail sector. As a result, our analysis of the District’s 1993 increase cannot tell us anything about the effects of the minimum wage on retail employment. From an analytical point of view, if wages did not increase, it is as though there was no increase in the minimum wage. The minimum wage cannot have an effect on employment if it has no effect on wages.
Our analysis of confidential, firm-level data made available to us by the Bureau of Labor Statistics did find a fall in employment in retail in the District relative to retail employment the suburbs, but, as we make clear in our report, this observed decline could not have been due to the minimum-wage increase because the minimum-wage increase did not raise wages in the sector.
The Sage Policy Group report and its conclusions are built around one number from our report (-1.5), taken completely out of context. This number is what economists call an “elasticity,” which shows the relationship between changes in one variable (in this case, the minimum-wage level) and another variable (in this case, retail employment). If the 1993 minimum-wage increase had raised wages, then this “elasticity” would provide an estimate of the effects of a minimum-wage increase on retail employment in the District. In this case, however, the minimum wage did not increase wages, so the elasticity is meaningless and cannot be used to assess the impact of the increase. The observed fall in retail employment in the District was due to some other factor or factors, possibly including population losses, falling average incomes, bad weather, or some other factor beyond the scope of our investigation. What the analysis and the data clearly do, however, is rule out the minimum wage as a cause of the decline in retail employment in the years after the 1993 increase.
(A technical aside: Table 6 in our report, which is the source of the crucial -1.5 number used in the Sage Policy Report, clearly demonstrates that the observed decline in employment was in response to an economically small and statistically insignificant change in wages in the retail sector. Employment in DC retail fell relative to employment in the suburbs, but retail wages did not increase relative to the suburbs. The -1.5 number appears in our Table 6, panel (d), column 7, row 4. The corresponding wage increase associated with the 1993 minimum-wage increase appears in the same table, panel (a), column 7, row 4; the increase is small (0.01 log points, or about one percent) and statistically insignificant (standard error of 0.02 log points). Panel (a) reports nine different ways of measuring the wage impact of the DC minimum-wage increase. In four cases, wages appear to have fallen after the new wage was passed; in two cases, measured changes rounded to zero; in three cases, wages increased by small, statistically insignificant amounts. No slicing of the data provides any evidence that the 1993 minimum-wage increase raised retail wages in the District in 1993 — a crucial finding that the Sage Policy Group failed to mention or to factor into their analysis.)
We clearly stated these findings and their implications in our report: the 1993 increase in the Washington, DC, minimum wage “does not allow us to draw conclusions about the employment effects of binding citywide minimum wages.”
Meanwhile, the Sage Policy Group’s selective reading of our report led them to ignore important findings based on the experience of Santa Fe and San Francisco. In those cases, where wages increased, the overwhelming weight of results pointed to no negative effect on retail employment. In Santa Fe, where the size of the minimum-wage increase was much larger than it had been in the District, average wages in retail increased by statistically significant amounts ranging from 2 to 9 percent. Of the seven tests where we found increases in retail wages, retail employment in Santa Fe actually increased in four cases; in two cases, employment fell by a statistically insignificant amount; and in only one case did employment fall by a statistically significant amount. The evidence for Santa Fe, therefore pointed strongly toward no negative effect on retail employment. Meanwhile, in San Francisco, the minimum-wage increase only raised retail wages by a statistically significant margin in one of the nine cases we examined. In that case, retail employment rose, but the increase was not statistically significant. Again, the best conclusion to draw from our analysis is that the minimum-wage there had no negative effect on retail employment in San Francisco.
The Sage Policy Groups report also chose to hide our major findings: “The results for fast food, food services, retail, and low-wage establishments in San Francisco and Santa Fe support the view that a citywide minimum wages can raise the earnings of low-wage workers, without a discernible impact on their employment.”
As an economist and long-time resident of the District of Columbia, I am deeply troubled by the misrepresentation by the Sage Policy Group of my research. The Sage Policy Group’s report does not accurately reflect our conclusions and should not be used as an argument against the implementation of a living wage for retail workers in the District of Columbia.
To print a copy of John Schmitt’s letter, click here.
