Neumark and Wascher on minimum wages and youth unemployment

Here is the abstract from their piece from 2003 (pdf):

We estimate the employment effects of changes in national minimum wages using a pooled cross-section time-series data set comprising 17 OECD countries for the period 1975-2000, focusing on the impact of cross-country differences in minimum wage systems and in other labor market institutions and policies that may either offset or amplify the effects of minimum wages. The average minimum wage effects we estimate using this sample are consistent with the view that minimum wages cause employment losses among youths. However, the evidence also suggests that the employment effects of minimum wages vary considerably across countries. In particular, disemployment effects of minimum wages appear to be smaller in countries that have subminimum wage provisions for youths. Regarding other labor market policies and institutions, we find that more restrictive labor standards and higher union coverage strengthen the disemployment effects of minimum wages, while employment protection laws and active labor market policies designed to bring unemployed individuals into the work force help to offset these effects. Overall, the disemployment effects of minimum wages are strongest in the countries with the least regulated labor markets.
More recently, Modeled Behavior has a relevant update on how minimum wages reduce the number of new teenage hires.  Brochu and Green you will find here, the effects on teen hiring are pretty clear.


"the effects on teen hiring are pretty clear."

Well, they appear to be. Nonetheless, there are a lot of people who will object to the conclusion regardless of the existing evidence.

We call them "libtards".

If you regress the teen unemployment rate against the total unemployment rate since 1950 the regression says that teen unemployment is 2.7 times the total unemployment rate. The r-square of the regression is 0.98.

Every significant rise in the US teen unemployment rate since 1950 has been associated with a recession. There is not a single example of the teen unemployment rate rising significantly during a business expansion.

If the minimum wage is raised in a business expansion --like in the 1960s, the late 1970s and the 1990 -- the teen unemployment rate falls, implying the wage hike has no impact on teen unemployment.

This is what the US data shows. Why should I ignore this robust data.

Does your data show:

1. How much the new minimum wage deviated from equilibrium price.
2. How much the new unemployment rate deviates from expected unemployment rate absent minimum wage increase.

If not, what value does your data have?

Every time I am shot, it hurts. When I am shot at the same time as I am hit with a hammer, getting shot still hurts. So why should I worry about getting hit with a hammer?

(guy who doesn't understand counterfactuals)

Hence jobless recoveries. A similar dynamic with private unionized workers. Wage and benefit packages negotiated during boom times. And a gradual disappearance of private unionized jobs.
What is the jobless rate of young workers right now?

But your thinking assume that higher worker wages and income detract from boom times by reducing the number of consumers, and that by cutting wages and incomes, profits rise, and the increased wealth creates new consumers from sheer demand from the higher supply. Obviously to you, workers are never consumers, so boom times leading to higher wages and incomes can't possibly extend of expand the boom times by continuing to drive increased consumption.

Great story but it isn't what happened.

If you regress Y on C, I, G, and NX you will get a high R squared too.

The teen unemployment rate is a component of the total unemployment rate. Of course they are highly correlated (the square root of R squared).

This is also a time series, and you havent tested whether the series are stationary and/or cointegrated.

And by omitting all of the other variables that explain the teen unemployment rate, the coefficient on the total unemployment rate is biased and inconsistent.

And you aren't testing for structural breaks, conditional heteroskedasticity, etc.

Recession dating is endogenous both to the choice to raise minimum wage and the unemployment rate.

The criticisms of the other commenters and these indicate you need to spend some more time with your nose up Walter Enders' ass before you throw around regression results here.

But by all means express an opinion.

I have been waiting years for a single study by those oppose to the minimum wage that incorporated the impact of the business cycle in teen or minimum wage employment.

Why is it that every study I see claiming that the minimum wage hurts teens, or other poor people completely ignores the impact of the business cycle.

Can anyone show me a good study by the opponents of the minimum wage that incorporates the business cycle in their results?

If you make minimum wage or teen employment a function of either the total unemployment rate or the unemployment rate of over 25 years old the current cycle behaves almost exactly how the regression says it should.

None of you have ever explained why a numerous minimum wage hikes during business expansions apparently had no impact.

I am not doing poor economics. You guys are doing bad economics by omitting relevant variables such as the teen population, or the business cycle in your studies of the minimum wage.

By the way, I am not expressing an opinion. I am reporting the data.
There is a big difference. You guy are the one expressing an opinion that is contrary to the data.

"This is what the US data shows. Why should I ignore this robust data."

I gave you about a dozen reasons not to rely on your own result.

Yes, it is bad economics. This analysis would get you a D or F in an econometrics course.

I didnt make any claims here about MW and employment. In another post, I made several suggestions on how to control for confounding factors including some of those above.

There is no burden on me to prove that your preferred policy, a distortion to markets, has a particular effect. The burden is on YOU to demonstrate that the good outweighs the bad.

Neumark and Wascher are professional labor economists, far above your own level of knowledge. If I recall correctly, they recreated the natural experiment of Card and Krueger, fixing some of their blatant errors.

Do you suspect that Camden and Philadelphia faced different national business cycles? If not, then your criticism lacks all merit.

There have been hundreds of papers, Sparky, on minimum wage and employment. All of the good ones control for business cycles in one way or another. Or do you think you are smarter than all those researchers, referees, and editors?

Isn't the relationship between wages and business cycles obvious? Didn't Keynes talk about sticky prices, and labor is one of the stickiest.

In good times demand for labor increases, and would tend to raise wages. Those are great for young inexperienced workers who for a time at least are a cost rather than productive. During the Clinton boom unemployment was going down and folks in typically high unemployment areas were getting jobs as some kind of employer was using them. A minimum wage is not a factor when there is high demand.

During downturns the demand for labor drops. There will be more supply of low skill labor, and the higher the price, the less will be employed.

I don't quite see the need for a deep study, it is obvious to anyone.

A characteristic of a high labor cost economy is chronic youth unemployment. Most socialist or social democratic economies with generous social welfare have high youth unemployment.

Don't confuse him with actual thought.

He can't even properly interpret a regression coefficient much less model this correctly.

I have been waiting years for a single study by those oppose to the minimum wage that incorporated the impact of the business cycle in teen or minimum wage employment.

Your long, lonely wait has ended! Well, actually, it ended back in 1995 with Deere, Murphy, and Welch. Well, actually, it ended long before that with any number of other studies.

But let's talk about Deere, Murphy and Welch. They found a large effect of the 1990-1 minimum wage increase on both teens and high school dropouts. Fortunately for you, your parching thirst for business cycle controls can be slaked by considering their Table 4. In that glorious table---fulfillment of your anxious anticipation---you will find that higher min wages lower employment among teens and dropouts, controlling for overall unemployment.

Of course, all the rest of the literature does this too, so you can continue to slake your thirst for business cycle controls many, many times over (tens! scores! nay, hundreds of times!) by continuing to read in said literature.

I envy you the rapturous joy of having all your dreams fulfilled. If only my own thirsts were so meagre and easily sated.

"And by omitting all of the other variables that explain the teen unemployment rate, the coefficient on the total unemployment rate is biased and inconsistent."

Since its pooled cross sections (aka a panel) omitted variables don't make it biased.

Where in his comment does it even suggest that he used panel data?

He said he regressed teen unemployment on total unemployment and got an R-squared of .98. That doesn't sound like a panel - it sounds like a time series model of a simple linear regression.

Panel data allows you to control for SOME omitted variables that vary across cases but are time invariant. If that is what you are relying on in your experiment, you need to be prepared to explain which variables are omitted and how the model controls for them.

What are the other cases here?

1. US;
2. ???;
3. R^2 = .98

Then the dolt talks about behavior of the series during recessions but doesn't explain how, if at all, he controlled for recessions. If he used the 50 states as cases, don't they all have the same dates for national recessions?

The only thing more annoying than listening to sanctimonious drivel is listening to the rescue attempts.

You have obviously never read Engle-Granger. My guess is your IQ is too low to comprehend the math in their work on cointegration.

I don't recall exactly where I read this simple illustration (Don Boudreaux?):

Imagine that you want a new shed in your back yard. You can choose to hire lowi skilled workers at $10/hr or skilled workers at $30/hr. 5 Low skilled workers will complete the project in 5 hours for a total labor cost of $250. 2 skilled workers will complete the project in 5 hours for a total labor cost of $300. Now, raise the minimum wage to $15.

It takes a lot of effort to convince yourself that all of the basic principles of economics stop working when applied to labor. I'd guess that most people who believe that labor doesn't conform to equilibrium price have some sort of negative present-time ("things are always getting worse") bias which causes them to believe labor theory of value.

"It takes a lot of effort to convince yourself that all of the basic principles of economics stop working when applied to labor."

I'm not sure that's true... it seems to me that many people have intuitions which say they're different. Doing cost-benefit analysis in terms of dollars per life, for example, is something that many people oppose intuitively. Paying for organs is another. It is easier for an economist than a non-economist to view labor as just another commodity.

Aside from that, workers as suppliers of labor don't behave like other suppliers. The labor market models are detailed and sophisticated, but are unlike run of the mill supply and demand models. For example, backward bending supply curves, derived demand for labor, reservation wages, efficiency wages, moral hazard.

The leading model on the other side is a monopsony model, where low wage employers have market power. Whether it meets your definition of "basic economics" I don't know, but we do teach it to undergrads. Below you can find MIT students in their first course on economics learning the two leading model of the minimum wage. At minute 35 or so, the professor starts summarizing and gets to the empirical literature at about 42 minutes.

"It takes a lot of effort to convince yourself that all of the basic principles of economics stop working when applied to labor."


But it's funny how economists take a lot of effort to convince themselves that the basic principles of economics stop working when applied to labor immigrating.

Thus, we see the same economists citing David Card's Miami natural experiment * to prove that that mass immigration doesn't lower wages and denouncing David Card's New Jersey natural experiment to wave away evidence that the minimum wage doesn't lower employment.
* For an amusing explanation of what practically every economist has overlooked in thinking about wages in Miami in 1980-84, see:

It's instructive to compare the vast interest in the econosphere in the question of whether higher minimum wages lessens teen employment versus the crickets-chirping lack of discussion over whether higher immigration lessens teen employment.

If a significant share of MW recipients are teens, then (some would say) we can feel less bad about the disemployment effects... it's pocket money, not rent money. A reduction in pocket money is a shame, but not a tragedy. So by extension, if immigrants "take" the jobs that were formerly occupied by teens, we shouldn't feel too bad about that.

In other words, to the extent that the econosphere is discussing the question of the MW and teen employment, that discussion would probably end up supporting higher immigration.

I don't much care about the pocket money teen employees, either, although I must say that the much lower number of young adults who have had paying jobs before their mid-20s is a bad trend.

But here's two things that bother me a lot more.

1. That there's lots of kids who really need that part time job to help with their own expenses or to help their parents can't get one; and

2. There's millions of desperate unemployed adults who can't find a job, not even one for minimum wage.

Should our teens and adults have to compete with the massive number of immigrants that have come here in the past and have continued to come here after five years of economic bad times? I say no. I say it's high time to rein in the cheap labor employers and really reform our immigration laws.

I have heard the fast food restaurant owners cry boo hoo that they can't find any workers, that they must have immigrants to staff their business, and I have been nothing but disgusted.

Again, it is obvious to you that workers are never consumers.

Teens are consumers with infinite cash to pay for consumption, thus teens getting jobs at either the teen wage of $4.25 or $7.25 NEVER INCREASE CONSUMER DEMAND.

Likewise, immigrants are NEVER CONSUMERS so an immigrant getting jobs becomes a pure drag on the economy, sucking cash out of the economy into a black hole. Immigrants will never add to the economy by being consumers.

Wages and incomes are totally unrelated to consumption, obviously.

Of course they consume, but minimum wage redistributes income from business owners, consumers, owners of other inputs, and workers who lose their jobs/hours. Aren't they all consumers too?

There are two separate issues here. First is the welfare change and redistribution in the market for low skill labor and associated markets. Second, there are macroeconomic effects of a broad policy.

I assume that you have intelligent things to say about both of these, but it would help the debate to stick to one or the other. In this case, we are discussion the effects on employment of MW hikes.

Steve, there is no inconsistency for those who oppose barriers to voluntary transactions. A minimum wage, like immigration restrictions, are barriers to voluntary transactions.

The bringing of an immigrant employee into a country is not voluntary unless every owner of the country (i.e., every citizen) consents.

So much for your theory.

The issue is artificial interference in the price system, not maximizing teen employment. Min wages interfere with prices and are bad. We measure how bad by looking many thing including teen employment data. Policies that impact labor supply curves are wholly different.

Will robots be good even when they cause massive unemployment?

Dunno, if I could EVER find a teenager who babysits or is willing to shovel some snow I might be convinced that they've been pushed out of the workforce. As is, all the ones I know seem unemployed by choice and completely unwilling. And babysitting is expensive - you can make significantly more than minimum wage.

Most parents barter babysitting, although it can be hard to distribute 3+ kids when other barterers have only 1 or 2. Having grandparents and siblings nearby also helps (also helpful to opening a donut shop or cut-rate ethnic restaraunt).

Cash is necessary for misanthropes.

Just before the last round of minimum wage hikes some 92% of employed teenagers made more than the minimum wage.

Hi, Marie!
My kid babysits for $2 an hour. I suspect your experience is regional.

So, using 1976, a time when the economy was really really suffering and the number of teens entering the labor force was at record highs, and then 2000 when the economy was booming and the number of teens entering the labor force was hitting lows, we are obviously going to learn everything we need to know about teen employment when gasoline prices are at near record highs and then when gasoline prices are hitting historic lows.

Comparing 1976 to 2006 or 2010 would be more like comparing apples and apples, but still with big difference in the number of new jobs for youth needed to handle the increase in youth workers (16-25?).

Good point.

From 1950 to 2000 the percent of teenagers employed bounced around 44% of the teen population.

Only since 2000 has the share of teen employed fallen out of the 1950 to 2000 narrow range. Much of that decline occurred in the in the mid-1990s when the minimum wage was unchanged.

This angst about youth employment is better tended to when we have an unemployment rate of 6% or less. The question of timely policy relevance is: how do minimum wage changes affect the employment of 20 and 30 somethings? After we have figured this out, we can think about the kids.

Well, only 2% of minimum wage workers are over 24.


That's crazy.

you would be more believable if you occasionally posted about a study that didn't fit your ideology.

Good point. Obviously the authors of this blog are under no obligation to post research that doesn't align with their worldview(s), but I feel like 90% of the time that I know the topic I can predict exactly what position research linked here will support. ACA: not working; minimum wage: bad; ethnic restaurants: good; Paul Krugman: wrong; FDA: harmful; austerity: good.

Everyone has a bias to talk about things they don't like or would like to see changed (like me right now), but this space would be more appealing to a broader audience if it wasn't so a libertarian echo chamber. Of course there's plenty of interesting non-ideological stuff on here as well, and I much admire the blog. Just want to see it get better.

If teens can not get jobs doesn't that send them a signal they need to receive?

For some teens does unemployment send the very useful signal that they are worthless with their current skill set and that they better figure out a game plan for how to become value creators?

Like dealing cocaine.

Or staying in school. Certainly folks up on this topic must be familiar with Sutch's research on educational cascades?

Right. Staying in school during summer vacation.

You're a genius.


A signal that TEENAGERS are worthless? You think it's good to tell our young people that they are worthless? As opposed to giving them a signal that their abilities are temporarily (one hopes) worth less than $7.25 an hour?

But you are right that they are getting a signal that they are absolutelyworthless, and that signal comes from a government that imports people to compete with them for jobs at price points below, at, and above the minimum wage.

Regarding my earlier post. Did you read the linked abstract? "the engineered shortage of low-skill, low-paying jobs induced teenagers to invest in additional human capital – primarily by extending their schooling – in an attempt to raise their productivity to the level required to gain employment"

That's the positive analysis. It has nothing to do with my intelligence - though your assessment was correct - I am a genius.

How's your reading comprehension?

another good post from Dube:

two excerpts:

"In January 2014, a funny thing seems to have happened. Parts (though not all) of the econoblogosphere forgot why time series econometrics fell out of favor in the early 1990s when it comes to analyzing minimum wage policies. Besides the fact that there is a lot more variation in minimum wages than just the federal (or average) minimum wage in the U.S., the timing of the minimum wage increases is very uneven. For example, U.S. minimum wage increases tend to occur more frequently during the late phases of the business cycle."

" whether minimum wages cause teen employment to fall remains controversial. I think the clearest U.S. evidence comes from comparing across border counties with different minimum wages, using regional controls, or using synthetic control groups. All of these suggest the impact on teen employment is fairly small for the kind of minimum wage increases we have seen in this country. But whatever your position is on this question, you should probably steer clear of simple time series evidence that economists on both sides of the debate have wisely moved away from."

Could it be that a higher minimum wage encourages more non-teen people out of unemployment and into minimum wage jobs, thereby displacing teens? Thus a higher minimum wage is more effective at eliminating poverty?

Great example of the ideological Turing test--Cowen sees a study that minimum wage hikes benefit the non-teen poor, but doesn't see that minimum wage hikes benefit the non-teen poor.

This fascination with the minimum wage seems all out of proportion. Any idea how many workers who make more than the minimum wage would get an automatic bump if the MW were increased?

Pretty much all of them. I've made this point in two other posts.

...except for those above them at the other end of the not-so-large range who face the brunt of equal and opposite reaction: "automatic" compression. Has that point also been made in other posts, or even just an acknowledgement that teens displaced from work may not be beneficial (ex. people tend to start out as teens before becoming non-teens)? Or could this all be deeper Turing-sponsored irony?

Some "economists" are still researching the effects of minimum wage on...employment levels. Why bother convincing yourself of what you want to believe? It's as much wasted effort as focusing on employment levels, even for those on the other side that see the obvious. Employment levels are a drop in the bucket of potential responses and impact. Twenty plus years of focused MW research drawing pro-MW conclusions with support and endorsement from PhD and even Nobel recipients, and not one recognition of the obvious: that responses to and effects of MW policy are (minimally) uncontrollable and that this may possibly call into question net benefit. Some studies even show correlation to price increases (which influence cost of living for MW recipients) and compression of wage increase above MW (which influence the ability to work for increases in wage above MW) and still applaud MW as a net desirable outcome...while still failing to acknowledge the possibility of more responses. Even for someone like me that wants to see people's access to the economy at least not forcibly reduced, it's embarrassing to see so many economists fail so hard for so long to both understand economics and keep their prescribed political thought from rendering their expertise useless if not damaging.

Comments for this post are closed