Teen employment and minimum wages

by Spencer England

Last week Kevin Erdmann  at the idiosyncratic whisk blog published this chart to demonstrate how minimun wages have caused teen employment to suffer.

It is an interesting chart, but it suffers from the sin of omitted variables.

I constantly see similar charts from those opposing the minimum wage were they seem to assume that nothing impacts teen employment except the minimum wage.  They apparently believe that the business cycle never impacts teen employment or unemployment.   For example, we just suffered the second worse recession in US history that caused adult unemployment to soar.  But in their analysis, the great recession had no impact on teenagers. They claim that the last jump in teen unemployment was entirely attributable to minimum wage increases.

I suggest we look at a different chart that shows three variables:the teen unemployment rate, minimum wages and recessions.

First, notice in this chart that there were period in the 1960s, the late 1970s and the mid-1990s when the minimum wage rose and the teen unemployment rate fell.  Something you would never know from studying Kevin’s chart or other opponents of the minimum wage arguments against the minimum wage.  I don’t know how many times Ive heard the claim that raising the minimum wage always causes employment to fall.  The data massively contradicts this claim.

Next, observe in this chart that every rise in the teen unemployment rate was associated with a recession.  The only exception was one small rise in 1962.  This raises an important question.  Was the rise in the teen unemployment rate caused by the rise in the minimum wage or the recession?  In answering this question observe that every recession saw the teen unemployment rate rise and that the teen unemployment rate fell in every business expansion. On the other hand the relationship between teen unemployment and the minimum wage is inconsistent. The only time a rising teen unemployment rate has been associated with a rising minimum wage was when there also was a recession.

Note,  that there also were recessions when the teen unemployment rate rose but the minimum wage was was unchanged. A rising minimum wage during a business expansion has always been  accompanied by falling teen age unemployment rates– something that opponents of the minimum wage would prefer to overlook.  Since 1960 there have been eight recessions. In three of them –1970, 1981 and 1990 — there was no increase in the minimum wage, but the teen unemployment rate rose sharply. Moreover, in the business expansion of the 1960s, the late 1970s and the  mid-1990s rising minimum wages was accompanied by a falling teen unemployment rate. Since 1960 there have been 18 increases in the minimum wage, but only about half of the increases were accompanied by a rising teen unemployment rate and each of those was in a recession. A higher minimum age in a business expansion has always been accompanied by a fall in the teen unemployment rate.








I am not arguing in favor or in opposition to a minimum wage.  All I ask for is an honest discussion and it sure looks to me like Kevin Erdmannd  and others arguments are factually challenged.

A second way to look at the impact of the business cycle on teen unemployment is to compare the teen unemployment rate to the adult–over 25 years old — unemployment rate.  If the two rates have a different history it would suggest that different factors drive the teen and adult unemployment rate.  That is what you would expect if the minimum wage has a big impact on teen unemployment but a very small impact on adult unemployment. Since 1950 the teen unemployment rate has averaged some 3.1 times the adult unemployment rate. The correlation between the two unemployment rates is 0.98, strongly implying that the same economic factors drive both measures and that the business cycle plays a major role in determining the teen unemployment rate.



OK, it is a valid question to ask why the teen unemployment rate is 3.1 times the adult unemployment rate.  Opponents of the minimum wage will argue that the existence of the minimum wage generates this difference between teen and adult unemployment rates.  They may be right.  But it is unproven.

Kevin also presented data on teen employment that was also based on the premise that nothing impacted teen employment but the minimum wage.  I would like to introduce what I think is a major omitted variable that has had a major impact on teen employment. That variable is the teen population.  In the post WW II era we had a very unusual pattern of  teen population growth.   Because of the baby boomers aging. In the 1960s and 1970s baby boomers passing through their teens caused the teen population to rise rapidly. But after 1978 the baby boomers aged out of their teen years and were replaced by the baby-bust generation.
Consequently, the teen population actually fell in the 1980s.

The next chart shows the relationship between the teen population and teen employment. Population is on the left axis and employment is on the right axis.  Both axis have the same scale, but the employment axis starts above zero so you can see the tight relationship between teen population and teen employment.







As  you can see in the next chart from 1950 to 2000, on average, some 44% of the teen population was employed.  But in the 1980s the share of the teen population employed was above the 44% average. This is completely contrary to the claim that the minimum wage caused teen employment to fall in the 1980.  The falling teen population was what caused teen employment to fall in the 1980s.  See, if you claim everything was due to the minimum wage and ignore anything else that might have played a role, it is easy to demonstrate that the minimum wage caused every bad development.








The other interesting point this chart shows is that after 2000 the share of teens employed fell sharply. Note, this is annual data. It hard to blame this on the minimum wage since there were no minimum wage changes in the early 2000s. Actually, labor economists are well aware of the drop and there have been numerous academic papers on it.  Labor economist’s consensus is that because of the very high returns to education, teenager are increasingly realizing that working a minimum wage job is a very poor use of their time. They will be much better off in the long run spending their time in an activity that will look impressive on their college applications.  Lets face it, working part time at a minimum wage job does not carry much weight in an application to a better school. It is a rational economic decision for teens to avoid minimum wage employment.

There is one other set of data that is very relevant to any discussion of teen employment and the minimum wage.  The data is the share of teens earning the minimum wage.  This data only goes back to 2002, but it appears to be very important.  In the business cycle leading up to the great recession over 90% of employed teens earned more than the minimum wage. Even after the great recession and a round of three large minimum wage increases the share of teens earning more than the minimum wage still was over 75%.










First, this data strongly supports the thesis that teen participation in the labor market fell sharply in the early 2000’s because the minimum wage was below teens reservation wage.  Teens found more valuable uses of their time than working at a minimum wage job.

Second, the data only goes back to 2002, but if only a small portion of employed teens actually worked at the minimum wage in earlier cycles. it call into question the entire practice of using teen employment data to analyze the impact of minimum wage changes.

This data suggest that the problem with the minimum wage may not be that it is too high for employers to hire teenagers.  Rather it suggest that the problem with the minimum wage is that it is too low to induce teens into the labor market.

Finally, i regressed the teen unemployment rate against four variables.

1.  Adult unemployment

2. A dummy variable for recessions that is one when the economyis in a recession and zero when it is in an economic expansion.  This series is created by Haver Analytics and is in their databases.  I used it to create the recession shading in my first chart here.  Moreover, the St. Louis Fed also uses it to create recession shading in their charts.  So Im not playing games with the dummy variable.  It is a widely used and recognized variable among economist.

3. The teen population.

4. The minimum wage.

The regression says that the business cycle and teen population dominate the teen unemployment rate and that the minimum wage is insignificant.