In a simple demand-supply analysis, it can be shown that setting the legal minimum wage at a level above the market-clearing level will result in a fall in employment level relative to the competitive employment level. This is not necessarily the same as an increase in measured unemployment, as workers who do not get minimum-wage jobs may not always engage in job search.
There are heterogeneous workers in a labor market. The minimum wage will be binding for some workers and will not be binding for others. For example, minimum wage tend to have the greatest effect on teenage workers (as these workers lack human capital and have low market-clearing wages). Until the early nineties, it was a professional consensus in the U.S. that a 10 percent rise in the minimum wage would lead to a 1-3 percent fall in teenage employment.
Skilled workers (whose market-clearing wage are above the legal minimum) may be substitutes for unskilled workers (whose market-clearing wage are below the legal minimum). If this is the case, a rise in the price of unskilled workers may raise the demand, and hence the wage and employment, of skilled workers.
Minimum wage laws typically have exemptions for some sectors, and compliance with the law is not complete. If minimum wage reduces employment in the covered sectors, then the excess labor may move to the uncovered sectors. This lowers wages in the uncovered sectors, while leaving total employment relatively unaffected.
Many minimum wage workers (e.g., teenagers) are second earners in the family. A rise in minimum wage for such workers may have little effect on reducing poverty. One study of the 1990-1991 increase in the minimum wage in the U.S. found that, of those who earned between the old and new minimums, only 22 percent lived in poor families. Assuming no employment effects, only 19 percent of the estimated earnings increases associated with the minimum wage increases went to poor families.
How does minimum wage affect workers' expected income? If the disemployment of a 10% increase in wage is 1-3%, one may think that workers' expected income will rise following an increase in minimum wage (110% x 97% > 100%). This reasoning is not correct. In a population of 100 workers, the minimum wage is binding for only a fraction (say, f) of them. Suppose a 10% rise in minimum wage reduces total employment to 97. Then 100-f workers are unaffected because the minimum wage is not binding. f-3 workers experience a wage gain of 10% (and maybe less). 3 workers lose their job. So change in total income (for f affected workers) is
Up to now, we have assumed that labor contracts are uni-dimensional. In fact, in additional to the wage rate, labor contracts specify (at least implicitly) other things such as fringe benefits, training opportunities, and expected effort level. Fixing the wage rate may cause employers and employees to adjust these other dimensions. For example, employers may provide less training to employees or require them to exert greater effort. If adjustment occurs in these dimensions, the dis-employment effect of minimum wage would be smaller.
When the federal minimum wage first went into effect in 1938, wages in the South were much lower than wages in other parts of the U.S. In the seamless hosiery industry employment fell by 5.5 percent in southern mills but rose by 4.9 percent in northern mills. Moreover, employment fell by 17 percent in mills that had previously paid less than the new minimum wage, while it stayed virtually the same at higher wage mills.
The minimum wage also seemed to have accelerated the switch to more automation. There was greater investments in converted-transfer machines and fully automatic machines. In addition, the machines were used more intensively than before. A night shift was added at many mills, and these workers did not receive extra pay for working this undesirable shift. Finally, total imports of seamless hosiery surged by about 27 percent within two years of the minimum wage's enactment.
The federal minimum wage was raised from $3.35 in 1989 to $3.80 in 1990 and further to $4.25 in 1991. The following table shows the trends in teenage employment.
employment rate index (1985=100) year ages 15-17 ages 18-19 men ages 35-44 -------------------------------------------------- 1985 100.0 100.0 100.0 1986 104.4 99.6 99.8 1987 107.8 102.2 100.2 1988 108.6 104.6 101.0 1989 109.2 104.7 100.9 1990 97.8 99.6 100.1 1991 89.2 95.2 98.3 1992 86.6 94.7 97.7
But one might argue that 1990-92 were recessionary years, and teenage workers might be particularly vulnerable to recessions.
The following table shows the fraction of low-wage (less than or equal to $4.25 in 1989) workers for different demographic groups and the one-year employment change since the minimum wage was raised in 1991.
men women Group low wage emp. change low wage emp. change ------------------------------------------------------------------ age: 15-19 44.5 -15.4 51.8 -12.9 20-24 14.2 -5.6 19.0 -4.3 25-64 3.3 -2.5 8.8 -0.3 65-69 14.0 -4.3 21.0 +3.5 race: black 11.0 -4.8 16.9 -3.4 white 7.2 -3.1 13.0 -0.6 Asian 5.4 +0.7 9.3 -0.3 ethnicity: Mexican 15.6 -4.8 21.9 -5.5 other Spanish 8.8 -3.3 16.4 -0.7 non-Spanish 7.1 -3.2 12.9 -0.8 schooling: < 12 20.7 -6.6 35.4 -7.3 = 12 6.0 -4.0 13.5 -2.2 > 12 3.5 -2.8 6.4 -0.7 marital: single 15.1 -4.4 18.4 -3.2 married 2.7 -2.4 9.0 +0.7 states: 10 lowest wage 13.1 -1.5 23.7 +0.0 middle wage 7.9 -3.4 14.2 -0.7 10 highest wage 3.5 -4.0 5.6 -2.9 all 7.5 -3.2 13.3 -1.1
For all categories except partition by states or by gender, groups with more low-wage workers experienced greater employment decline, as predicted by theory. The employment rate of women does not fall relative to men, though there are more low-wage workers among women than among men. This has to be understood in the background of persistent growth in female labor participation over the past few decades. What about classification by states? Some authors have used such classification to argue that minimum wage does not have negative employment effects.
During the 1980s and 1990s, low wage states (e.g., southern states) grew faster than other states. So the small employment decline in low wage states may be due to good regional conditions. To check this idea, Deere, Murphy, Welch run a regression using employment rate of a selected group (teenagers, high-school dropouts) as the dependent variable. The independent variables include year effects and employment rate for men aged 15-64. The latter variable serves as a control for local economic conditions. Observations are aggregated by state and year (1985-1992). The following shows the results for men:
teenagers, 15-19 high school dropouts, 20-54 ----------------------------------------------------------------------- male employment rate 3.29 1.19 year effects minimum=$3.80 -4.78 -1.49 minimum=$4.25 -7.29 -3.13 p-value for other year effects 0.87 0.19
These results indicate that employment of teenagers and high school dropouts are highly sensitive to aggregate employment (with elasticities greater than one). However, even controlling for aggregate employment, the years with minimum wage increases were associated with significant employment decline for teenagers and for high school dropouts. For teenagers, a 17% (30% in 1991) increase in minimum wage was associated with a 4.78% decline (7.29% in 1991) in employment rate. The implied elasticity is about 0.24-0.28, quite close to the estimates obtained from research done before the late 1980s.
In the late 1980s and early 1990s, David Card and Alan Krueger did a number of studies that disputed the dis-employment effects of minimum wage laws. In a well known study (Card and Krueger 1994), they compared employment levels of fast food restaurants in New Jersey and Pennsylvania before and after New Jersey raised the minimum wage from $4.25 to $5.05 (there was no change in the Pennsylvanian minimum wage). They found no evidence that the change in minimum wage reduced employment. The issue of minimum wage effect on employment is now again a subject of active research in the profession.
Several features of their study are noteworthy:
Across states, the empirical results can be summarized in the following table:
PA NJ Difference _____________________________________________________ FTE employment before 23.33 20.44 -2.89 FTE employment after 21.17 21.03 -0.14 change in FTE employment -2.16 0.59 2.76
Within New Jersey, some stores (those paying over the new minimum wage) are not affected by the legislations. They form another natural control group for comparison.
wage=4.25 wage=4.26-4.99 wage>5.00 _______________________________________________________________________ FTE employment before 19.56 20.08 22.25 FTE employment after 20.88 20.96 20.21 change in FTE employment 1.32 0.87 -2.14
Neumark and Wascher re-examine the issue using administrative payroll records drawn from the same geographical areas and the same chains. They have two main findings: