"evidence augments the case for shifting attention away from the minimum wage and towards the EITC as a means of improving the well-being of low-income households"

See How much have minimum wage increases contributed to the US employment slump? by Jeffrey Clemens and Michael Wither, economists at UC-San Diego.

"The merit of a minimum wage is a classic issue of contention in economics and is of particular interest during a contraction. This column uses worker-level microdata to investigate the effect of a federal policy change in the US that affected some states more than others. The authors evaluate not only the proximate effects on employment, but also follow workers for up to three years afterward to track career trajectory following a minimum wage hike."

"The minimum wage remains the focus of lively debate in both policy and academic circles.  In the 2014 US elections, four states passed referenda to increase their statutory minimums.  Earlier in the year, President Obama called for increasing the federal minimum from $7.25 to $10.10 per hour.  Among economists, minimum wage research has readily filled volumes (Card and Krueger 1995, Neumark and Wascher 2008).  Recent analyses arrive at disparate results, including work by Dube et al. (2010), Neumark et al. (2013), and Meer and West (2013).

An analysis focused on the Great Recession and subsequent recovery

In a new paper we explore the minimum wage’s contribution to labour market struggles following the Great Recession (Clemens and Wither 2014a).¹ We do so using data collected through the 2008 cycle of the Survey of Income and Program Participation (SIPP). The 2008 SIPP panel allows us to directly analyse the most recent change in the federal minimum wage, which occurred in July 2009.  We assess the contribution of this increase to the surrounding period’s sustained employment declines.  These declines included an eight percentage point reduction in the employment rate of young adults (ages 15 to 24), which remains far from fully explained (Clemens and Wither 2014b).

In July 2009 roughly half of US states were bound by the new federal minimum to increase their minimum wage rates by 70 cents, from $6.55 to $7.25 per hour.  The remaining states’ minimum wage rates changed by an average of roughly 10 cents.  The first dimension of our analysis thus compares changes in the employment of low-skilled workers in ‘bound’ relative to ‘unbound’ states.

Beyond making comparisons across states, the data allow us to identify targeted individuals with more precision than most past minimum wage research.  Specifically, we use 12 months of baseline data – from August 2008 through July 2009 – to identify workers earning wages in or near the affected range.  We similarly identify low-skilled workers with wages moderately above the new federal minimum, who were not directly affected.  We use these samples of slightly higher-skilled workers to bolster our confidence that our estimates disentangle the effects of minimum wage increases from the forces underlying the Great Recession.  Within each state, comparisons between the directly affected and the slightly higher-skilled workers help us to distinguish between trends impacting all of a state’s low-skilled workers and changes specific to those initially making below or near the new minimum. 

Our setting has additional advantages.  Methodologically, our approach enables transparent, graphical representations of the employment and income trajectories underlying our estimates.  Conceptually, it allows us to directly analyse the outcomes at the heart of the minimum wage’s intent.  That is, it enables us to analyse the minimum wage’s effects on the employment and income of low-skilled individuals. 

The distinction between individuals and workplaces matters for two reasons. 

• First, workplaces have varying capacities to substitute between low-skilled workers, higher-skilled workers, and capital.  Analyses of industry- and even firm-level data can thus have difficulty parsing out the minimum wage’s effects.  In such data, substitution between minimum wage workers and higher-skilled workers cannot typically be distinguished from increases in the earnings of the minimum wage workers themselves.
• Second, the minimum wage’s effects on career trajectories are likely more important for workers’ well-being than are its short-run effects on wages and employment.  Well known research by Murphy and Welch (1990), for example, showed that early career experience has particularly high returns.  Similarly, Smith and Vavrichek (1992) found that most minimum wage workers rise rapidly to employment at higher wage rates.  Our setting allows us to track workers’ outcomes for three years following the minimum wage increases we analyse.  We can thus provide evidence of the minimum wage’s effects on upward economic mobility.  

The minimum wage’s effects on employment and income trajectories

In the context discussed above, we find that minimum wage increases significantly reduced the employment of low-skilled workers.  By the second year following the $7.25 minimum's implementation, we estimate that targeted workers' employment rates had fallen by 6 percentage points (8%) more in ‘bound’ states than in ‘unbound’ states.  In addition to reducing employment, we find a 2 percentage point (12%) increase in the likelihood that targeted individuals work without pay. 

The latter effect is concentrated among early-career individuals who obtain at least some college education.  We interpret this as suggestive evidence that many of these workers' entry-level jobs can be posted as internships.  For low-skilled, low-education workers, the entire change in the probability of having no earnings comes through unemployment.

The primary threat to our estimation framework is the possibility that the low-skilled workers in the ‘bound’ and ‘unbound’ states were differentially affected by the Great Recession.  Aggregate data characterising the Great Recession’s severity reveal that the housing crisis was more severe in unbound states than in bound states.  Consequently, the recession would tend, if anything, to bias estimated effects of the minimum wage on employment towards zero.   We show that our estimates are robust to taking a broad range of approaches to controlling for macroeconomic conditions and the characteristics of the individuals in our sample. 

We consider our estimates' implications for the effects of this period's minimum wage increases on aggregate employment.  While a variety of caveats apply, our setting is reasonably well suited for this kind of extrapolation.  Our empirical approach allows us to estimate the minimum wage’s effects on a relatively representative sample of the population of targeted workers.  This is preferable to, for example, conducting such an exercise using estimates specific to selected demographic groups such as teenagers or high school dropouts. 

Over the late 2000s the average effective minimum wage rate rose by nearly 30% across the United States.  Our best estimate is that these minimum wage increases reduced the employment of working-age adults by 0.7 percentage points.  This accounts for 14% of the employment rate’s total decline over this time period and amounts to 1.4 million workers.  A disproportionate 45% of the affected workers were young adults (aged 15 to 24).

We next estimate the minimum wage increases’ effects on low-skilled workers’ incomes and income trajectories.  We find that binding minimum wage increases reduced low-skilled individuals’ average monthly incomes.  Targeted workers' average incomes fell by an average of $100 over the first year and by an additional $50 over the following two years.  While surprising at first glance, we show that the short-run estimate follows directly from our estimated effects on employment and the likelihood of working without pay.  The medium-run estimate reflects additional contributions from lost wage growth associated with lost experience and training.

Because most minimum wage workers are at early stages of their careers, lost opportunities for accumulating experience can be quite costly.  We provide direct evidence that such losses translate into meaningful reductions in upward economic mobility.  Two years following the minimum wage increases we study, low-skilled workers had become significantly less likely to transition into higher-wage employment in bound states than in unbound states.   Over this recent historical episode, the minimum wage’s effects on career paths thus appear to be quite important.

Concluding thoughts on directions for low-income support policy

As documented by extensive research on income inequality, low-skilled workers’ employment opportunities have declined in recent decades.  Work by Autor et al. (2008) and Autor et al. (2013) has identified technological change and expanding trade as leading contributors.  Our findings suggest that, along with housing and financial market crises, these forces have increased the sensitivity of low-skilled workers’ employment to increases in the minimum wage.   

We conclude by emphasising that the minimum wage is one among many policies that seek to improve low-skilled workers’ incomes.  Most notably, the Earned Income Tax Credit (EITC) has been used for precisely this purpose, and to great effect, for several decades.  Research on the EITC seemingly uniformly recommends it as a tool for increasing the employment and incomes of low-skilled workers (Eissa and Liebman 1996, Eissa and Hoynes 2006).  By supplementing incomes without burdening employers, it has meaningfully offset some of the inequality-inducing trends discussed above (Liebman 1998).  Our evidence augments the case for shifting attention away from the minimum wage and towards the EITC as a means of improving the well-being of low-income households."