1	How Did the $8.50 Citywide Minimum Wage Affect the Santa Fe Labor Market? Dr. Aaron Yelowitz University of Kentucky www.Yelowitz.com
2	Introduction “Living wages” are thought of as ordinances that affect businesses with a direct relationship to a city. Since 1994, living wage ordinances have become very popular. More than 130 localities have either “narrow” or “broad” ordinances that affect at least some workers. Wage floors are substantially higher than federal minimum wage of $5.15/hour. For example, city of Santa Barbara is currently considering a $14/hour living wage for local companies contracting with the city.
3	Introduction Citywide “minimum wages” are far less common. These laws mandate that all firms – regardless of whether they have a relationship with the city – provide a higher wage floor. San Francisco – Passed $8.50 minimum wage in November 2003, effective February 2004, then indexed to CPI. Santa Fe – As will be discussed later, passed $8.50 minimum wage that was implemented in June 2004. Raised to $9.50 in January 2006. 65% increase in wage floor. Washington DC -- $6.60 per hour, effective since 1993.
4	Introduction Other localities have tried, but failed in implementing citywide minimum wages: New Orleans, LA -- $6.15/hour in 2002, blocked by Louisiana legislature. Madison, WI -- $5.70/hour in 2004, Wisconsin legislature banned citywide minimum wages in exchange for raising statewide minimum wage. Milwaukee, Lacrosse, and Eau Claire, WI – all preempted by statewide minimum wage. Santa Monica, CA -- $13/hour in 2001 in costal zone, blocked by citywide referendum. Albuquerque, NM -- $7.50/hour on ballot in 2005, narrowly defeated.
5	Introduction Ongoing minimum wage campaigns: Santa Cruz, CA – proposal for $9.25/hour. Others?
6	Introduction Citywide minimum wages differ in a number of important respects from both living wages and statewide minimum wages. Coverage is clearly different from living wages. In addition, the public sector may be far less competitive than private sector. Ability to escape the citywide minimum wage by moving out of the jurisdiction is greater than for statewide minimum wages.
7
8	Santa Fe's Experience Timeline February 27, 2003: City council approves minimum wage by 7-1 margin. $8.50 starting in January 2004 $9.50 starting in January 2006 $10.50 starting in January 2008 Indexed to CPI starting in 2009. All employers with 25 or more workers (simple headcount each month). Initially, bill covered firms with 10 or more workers, but this changed. Both for-profit and not-for-profit firms.
9	Santa Fe's Experience State Court Litigation March 10, 2003: Santa Fe Chamber of Commerce and coalition of business groups – called “New Mexicans for Free Enterprise” – filed lawsuit against Santa Fe's law in state court. December 11, 2003: State judge throws out most, but not all, of NMFE lawsuit. December 18, 2003: State judge temporarily blocks ordinance. April 12-16, 2004: Wage trial in state court. June 24, 2004: State judge upholds ordinance, becomes effective immediately. November 30, 2005: New Mexico Court of Appeals upholds decision. Political process March 14, 2003: New Mexico Senate passes bill by 21-13 margin to prohibit local governments from establishing a minimum wage that exceeds the federal rate. Doesn't move further, however.
10	Predictions about the ordinance Predictions around the time of passage in January 2003. The higher minimum wage in Santa Fe's private sector would “not have that much of an effect” on the city's historically low unemployment rate. “Prices will rise,” Waldman explained. “There will be businesses that will try to recoup profits by raising prices. That is more likely to happen than major layoffs.” Larry Waldman, Senior Economist, University of New Mexico, Bureau of Business and Economic Research
11	Predictions about the ordinance These predictions highlight the fact that labor market adjustments are one of several mechanisms that firms can use to adjust to minimum wages. Layoffs Relocation Labor-capital substitution Higher consumer prices / lower quality Lower profits Change in corporate structure / outsourcing / temporary workers Land values – commercial & residential Is Waldman's prediction about prices reasonable in the Santa Fe context?
12	Preliminaries: How binding is $8.50? Labor market effects ultimately an empirical issue. Depends on: Firm sizes Ability to move outside of city limits Wage distributions Ability to pass along costs to consumers Only 10% of businesses were affected due to firm size, yet 55% of workers were covered.
13
14	Preliminaries: How binding is $8.50? Figure 2 suggests that passing on the higher costs of the ordinance to consumers is likely to be limited. In accommodation & food services, two-thirds of firms – the smaller ones -- were not covered. Three-quarters of workers were covered. Card and Krueger, “Myth and Measurement,”1995, p. 55, find “no evidence that prices rose faster among New Jersey restaurants that were most affected by the increase in the minimum wage. One explanation … is that restaurants in New Jersey compete in the same product market.” This strongly suggests that higher consumer prices are an unlikely outcome. Thus, the labor market adjustments may be more pronounced.
15	Preliminaries: How binding is $8.50? How did Santa Fe’s “historically low unemployment rate” fare? Figures 3 and 4 present time series evidence on relative unemployment rates in Santa Fe and the rest of New Mexico. Figure 3 is the metropolitan statistical area, which more accurately reflects a labor market. Figure 4 shows similar figures for the city.
16
17
18	Preliminaries: How binding is $8.50? The figures clearly show that Santa Fe’s historical advantage relative to the rest of the state shrunk. It also appears, but not as conclusively, that the relative gap shrunk after the minimum wage was implemented. My empirical results find effects for the full population – as is illustrated here, but the results are much more pronounced for the less educated.
19	Preliminaries: How binding is $8.50? What did the wage distribution look like? Examine nominal hourly wages in monthly CPS data. Fairly small sample sizes. Yet, Card and Krueger (p. 32) find dramatic shifts in the starting wage range due to the New Jersey increase. Figure 5 shows the Santa Fe MSA and rest of state.
20
21	Preliminaries: How binding is $8.50? Unlike Card and Krueger’s findings, no obvious increase in wages in Santa Fe. Using nominal wages. Measurement error. Small sample sizes. Not all firms & workers are affects. Surprising, still. Confirmed in empirical analysis, however.
22	Empirical model The implementation of the Santa Fe minimum wage creates a straightforward quasi-experiment, using both time-series variation in wage floors and cross-sectional variation across the state. Estimate models of the form:
23	Empirical model In practice, the labor market outcomes include: Employment (to population) Labor force participation Unemployment Long term layoffs > 26 weeks Involuntary part-time employment Usual hours of work per week Wages Workforce composition
24	Empirical model LWO is a dummy variable equal to one if the CPS respondent lived in the Santa Fe MSA from June 2004 onward. By including 33 dummy variables for each month/year combination (January 2003 … October 2005), as well as 3 dummy variables for the geographic location within New Mexico (Santa Fe, Las Cruces, Albuquerque, Rest of State), the coefficient estimate on LWO is the “difference-in-differences” estimator. Comparisons between less educated & more educated can be thought of as “triple-difference” estimates. Standard errors are corrected for clustering at the county-month-year level of aggregation.
25
26	Empirical model Santa Fe clearly does differ in some – but not all – respects compared to the rest of the state. The location dummies should largely control for these fixed, time-invariant characteristics. To the extent that there are time-varying, location-specific factors (other than the minimum wage) that affect the labor market, they could bias the interpretation on the LWO variable. If these factors do not vary by education level, however, then the triple-differences should still be valid.
27	Empirical model Other covariates include: Marital status Head of household Sex Educational attainment Race/ethnicity Veteran status Household size
28	Data Monthly Current Population Survey interviews around 50,000 households. Analysis presented here uses data from January 2003 to October 2005. Updated findings through December 2005 confirm results here. Identifies Santa Fe, Albuquerque, Las Cruces separately. Rest of New Mexico residents not identified at the local level. Roughly 1800 individuals (total) in New Mexico in any given month. Total sample includes 33,139 observations on non-elderly adults, divided nearly equally by 12 or fewer / 13 or more years of schooling.
29	Data Summary statistics are shown in Tables 1, 2, and 3, broken out by educational attainment. Some highlights: 5.9% unemployment over the time period; Less educated twice as likely to be unemployed. Usual hours of work 39.6 hours, with modest differences by education. 3% of sample classified as subject to LWO. Wages vastly different by educational attainment. Measured in constant July 2005 dollars, wages were $18.34 for more educated workers, and $11.78 for less educated workers.
30
31
32
33	Empirical Results Tables estimate, successively, models for full-sample, less educated sample, and more educated sample. Probit models for most labor market outcomes (except hours of work and wages). Estimate hours using OLS, and try a variety of approaches for wages.
34
35
36
37
38
39
40
41	Putting it all together Key policy parameter is the elasticity of employment with respect to the wage floor. Card & Krueger note that many “first generation” studies found very large, negative elasticities. CK find an elasticity close to zero. Neumark and Wascher, using payroll data on PA and NJ, find elasticities in the range of -0.22. A 10% increase in the wage floor reduces employment by about 2.2%.
42	Putting it all together Other researchers – Pollin and Wicks-Lim – have replicated my Santa Fe findings, but claim that near-zero effects on employment-to-population means there are no adverse labor market effects. Card & Krueger, Neumark & Wascher, and others use full-time equivalent employment however, not a simple headcount. CK count part-time workers at 0.5 of a full-time worker. NW have hours data. Using this, hours fell for the less educated by 8.4%, for a 65% increase in the wage floor. This would generate an elasticity of -0.13. As the figures showed, however, only 55% of workers were actually covered due to the firm-size requirements. In additional, a small percentage of workers are probably outside of the city itself, but within the MSA. Accounting for the fact that only workers at larger firms were covered, the elasticity would be -0.24.
43	Putting it all together Even this elasticity – -0.24 – does not reflect the full picture however. Labor substitution has distributional effects. Business movements from Santa Fe city to Santa Fe county are not detectable in the CPS. This would likely understate the labor market consequences. Average wage levels were considerably higher than $5.15 to start – meaning the percent change in the actual wage level was less, understating the elasticity.
44	Conclusions Results taken as a whole suggest negative consequences on the Santa Fe labor market. Have evaluated the full $8.50 period (June 2004-December 2005), and find similar results. Most pronounced on less-educated workers. Appears that there is some labor substitution toward workers who probably possess more skills or ability. $9.50 ordinance went into effect, though future increases must be voted on by City Council.