Minimum wage and teen employment: what 50 years of data show

If the minimum wage costs anyone jobs, it should cost teenagers first. They have the least experience and the fewest credentials, and many workers earning near the wage floor are young. That is why, for half a century, the teen employment rate has been the first place economists look when they argue about wage floors.

It is also the place where it is easiest to fool yourself. The teen employment rate has fallen a long way since the late 1970s, and it is tempting to line that decline up against minimum wage increases and declare the case closed. We think the minimum wage does real damage at the bottom of the job ladder, but the long-run chart alone does not prove it, and pretending otherwise would weaken our argument. Here is the data, and an honest account of what it shows.

The chart

The series below is the Bureau of Labor Statistics' employment-population ratio for Americans aged 16 to 19: the share of all teens, in school or not, who held a job. It comes from the Current Population Survey and is shown as annual averages from 1975 to 2025. The dashed lines mark each of the 13 increases in the federal minimum wage over that period, and the shaded bands are recessions as dated by the NBER.

Three things stand out. First, the teen employment rate is highly cyclical: every recession on the chart coincides with a drop. Second, the long decline did not happen all at once. Teen employment fell in steps, and it largely failed to recover after the 2001 and 2007–09 downturns, which it had done after the downturns of the early 1980s and early 1990s. Third, the rate has stabilized since the low point. It climbed from 25.8 percent in 2011 to around 32–33 percent in 2021–23, then eased to 31.1 percent in 2025. The latest seasonally adjusted monthly reading, for September 2026, is 31.7 percent.

Four rounds of increases

Federal increases cluster into four rounds. Taking them one at a time is more useful than staring at the whole line.

• 1975–1981 ($2.10 to $3.35). Six steps in seven years. Teen employment rose through the late 1970s and peaked at 48.5 percent in 1979, in the middle of the increases, then fell to 41.5 percent in 1982 and 1983. That fall coincides with the deepest recession of the period, so the chart cannot separate the two.
• 1990–1991 ($3.35 to $4.25). Teen employment was 47.5 percent in 1989. It fell to 45.3 percent in 1990, 42.0 percent in 1991 and 41.0 percent in 1992. Again, the increases arrived alongside a recession, and the decline outlasted that recession.
• 1996–1997 ($4.25 to $5.15). This one goes against our side, and we say so. Teen employment fell slightly from 43.5 percent in 1996 to 43.4 percent in 1997, then rose to 45.1 percent in 1998 and 45.2 percent in 2000. A booming late-1990s labor market swamped whatever effect the increase had.
• 2007–2009 ($5.15 to $7.25). The largest round since 1981 coincided with the Great Recession. Teen employment fell from 34.8 percent in 2007 to 28.4 percent in 2009 and 25.9 percent in 2010. Some of that drop surely reflects the recession; how much reflects the 40 percent increase in the federal floor is exactly what economists still argue about.

Notice what is not on the chart: a federal increase between 1997 and 2007. Yet teen employment fell from 45.2 percent in 2000 to 34.8 percent in 2007, with no change in the federal floor. Many states did raise their own minimums in that decade, but the national federal line cannot explain that drop. Anyone who presents this chart as proof that the minimum wage alone destroyed teen employment has to account for that decade, and they can't.

What else moves teen jobs

Defenders of the minimum wage are right that teen employment has many causes. Here are the main ones.

• Schooling. More teens stay in high school longer, take on more homework and extracurriculars, attend summer school, and enroll in college. BLS and other researchers have documented rising school enrollment, especially in summer, as a major reason teens work less. Some of the decline reflects choices made by families with more income, not jobs that disappeared, and nobody should mourn that part.
• The decline of the summer job. The classic summer job (lifeguard, camp counselor, ice cream stand) has faded. That fits the schooling story, but it also fits a demand story: summer jobs are short, low-productivity, high-turnover positions, exactly the kind a higher wage floor makes uneconomic first.
• Recessions. Teens are last hired and first fired. Every downturn on the chart hits them hard, and the post-2000 recoveries were slower for young workers than earlier ones. Without controls for the business cycle, the chart overstates what the minimum wage did.
• Immigration and competition from adults. Teens compete for entry-level jobs with less-educated adult workers, including immigrants. Research on low-skilled immigration, such as Christopher Smith's 2012 Journal of Labor Economics study, finds that it reduced youth employment, though how large the effect is remains debated. Older workers and retirees also take more part-time service jobs than they used to.
• Changing job mix. Retail and food service have changed how they schedule and staff, including self-service and automation. We cover that margin in our article on kiosks and robots. It is partly independent of the minimum wage and partly a response to it.

None of this lets the minimum wage off the hook. It means the minimum wage has to be judged with methods that hold these factors constant, and that is what the research literature tries to do.

What the research says

For the full list of studies with neutral summaries, see our research tracker. On teens specifically, the debate comes down to three generations of work.

The early consensus. Time-series studies through the 1970s, surveyed by Charles Brown, Curtis Gilroy and Andrew Kohen in the Journal of Economic Literature in 1982, settled on a modest but negative effect: a 10 percent increase in the minimum wage reduced teen employment by roughly 1 to 3 percent. That range became the benchmark the rest of the debate refers back to.

Neumark and Wascher. Starting with their 1992 ILR Review paper on state panel data, David Neumark and William Wascher estimated teen and young-adult employment elasticities of roughly −0.1 to −0.2, which falls inside the earlier consensus. In their 2008 book Minimum Wages, they reviewed the post-1990 literature and concluded that most studies find negative effects, and that the evidence is strongest for teens and the least skilled. They have been the most consistent voice for the view that minimum wages cost young, low-skilled workers jobs.

Allegretto, Dube and Reich (2011). In "Do Minimum Wages Really Reduce Teen Employment?" (Industrial Relations), Sylvia Allegretto, Arindrajit Dube and Michael Reich argued that states that raise their minimum wages differ systematically from those that don't, in regional economic trends among other things. Once they compared each state with others in its census division and allowed each state its own trend, the negative teen employment effect shrank toward zero and lost statistical significance, while the effect on teen wages remained. This paper, along with the border-county work by Dube, Lester and Reich, is the core of the "little or no effect" view.

Neumark, Salas and Wascher (2014). In "Revisiting the Minimum Wage–Employment Debate: Throwing Out the Baby with the Bathwater?" (ILR Review), Neumark, J. M. Ian Salas and Wascher replied that those controls throw away useful information. Neighboring states aren't automatically good comparisons, they argued, and state-specific trends can absorb the very employment changes the minimum wage causes. Choosing comparison states based on the data instead, they found that negative teen employment effects reappear, in line with the earlier literature. The Allegretto–Dube–Reich side answered in turn (with Ben Zipperer, ILR Review, 2017), and the methods argument has not been settled.

Since then. Two newer contributions frame the current debate. Cengiz, Dube, Lindner and Zipperer (2019) counted jobs just below and just above each new state minimum and found few missing low-wage jobs overall, though they examine all low-wage workers, not just teens. Neumark and Shirley (2022) gathered the preferred estimates from dozens of U.S. studies and found that a clear majority are negative, with the strongest evidence for teens and young adults. Critics disagree with how "preferred" estimates were chosen. Both papers are in the research tracker.

The honest summary is not "economists agree." It is that the more a study relies on teens and on large increases, the more often it finds job losses, and the most careful studies still disagree about how big those losses are.

Where we come down

We read the evidence this way. The teen employment-population ratio did not fall by half because of the minimum wage alone. Schooling, recessions and competition for entry-level work all played major roles, and the 2000–2007 decline happened with the federal floor unchanged. Anyone who says otherwise is overselling.

But "not alone" is very different from "not at all." In the episodes where the minimum wage rose sharply into a weakening economy (1980–81, 1990–91, 2007–09), teen employment fell hard and recovered slowly. Most of the long literature, from Brown, Gilroy and Kohen through Neumark and Wascher, finds negative effects concentrated on exactly this group. Even the studies that find little effect on headcount do not show that teens gain jobs when the floor rises. They show an effect too small to measure reliably with their methods. If you are the 17-year-old whose first job never got posted, that is cold comfort.

The cost of a wage floor also extends beyond one year's employment rate. As we argue in The first job ladder, an entry-level job builds habits, references and skills that pay off for decades. A policy that removes even a modest share of those first rungs reaches far beyond the teen years. Better tools exist to raise low incomes, such as the EITC, wage subsidies and lower payroll taxes, and none of them raises the cost of hiring someone with no experience.

Data and method

• Series. BLS publishes official annual averages for the teen employment-population ratio on the not-seasonally-adjusted series LNU02300012 (period "M13"). The seasonally adjusted series LNS12300012 is monthly only, so we chart the official annual averages and cite the SA series for the latest month. Seasonal adjustment does not materially change an annual average.
• Retrieval. Values were extracted on October 11, 2026, from the BLS LN time-series database flat file. The 2025 average covers 11 months. All 51 annual values, the 13 federal increases and the six recession periods are in data/teen-employment.json.
• What we did not do. We did not adjust for inflation, plot state minimums or estimate any causal effect. The chart is descriptive, and the causal claims above rest on the cited studies, not on the line.