Christy Romer, writing in the New York Times, deems the Earned Income Tax Credit a more palatable alternative to the minimum wage. So do I. (So, I feel confident, do the great majority of economists). But there is almost no overlap between Romer’s reasons and mine. I believe her reasons are wrong.
First, Romer observes (correctly) that while the minimum wage tends to reduce employment (though perhaps not by very much), the EITC has the opposite effect. That’s because the minimum wage is essentially a tax on hiring unskilled labor, while the EITC is a subsidy. When you tax something you get less of it; when you subsidize something, you get more.
But, contra Romer, that’s no reason to prefer the EITC. Since when, after all, is it automatically better to have too much of something than too little? Underemployment and overemployment are both bad things. Indeed, if the minimum wage (for whatever reason) has very little effect on employment while the EITC increases it substantially past the efficient level, that’s a good reason to prefer the minimum wage.
When we teach elementary economics — and by “We”, I mean pretty much everyone who ever teaches this course, surely including Christy Romer if she ever does — we teach our students how to illustrate the deadweight losses from both a tax and a subsidy. Anyone who’s ever seen those pictures knows that there’s no a priori expectation that one is worse than the other.
Romer’s other mistake comes when she compares the redistributive effects of the two policies — and ignores much more than half the story. She worries about who gets the benefits from each policy without pausing to ask who pays the costs — which is where the really big differences lie. The cost of the minimum wage falls, in the short run, on the owners of businesses that employ unskilled labor, and in the long run on the customers of those businesses. The costs of the EITC fall on taxpayers generally. Minimum wage supporters often talk about why it’s desirable to help the poor, but they rarely explain why it’s desirable to make big transfers away from the sort of people who eat at McDonald’s in order to spare the sort of people who eat at Per Se.
This is not, then, an argument for increasing the EITC — but it is an argument for preferring an EITC increase to a minimum wage increase, and a sufficiently compelling argument, I think, that it ought to knock the minimum wage right out of contention. How odd, then, that Christy Romer, while trying to argue that the EITC beats the minimum wage, ignores this argument completely.
Her argument is that the EITC targets beneficiaries better. Her argument is complementary to yours — you are concerned about who pays, she is concerned about who benefits.
its best if everyone pays, and its best if the people who need the help most get it.
Why prefer a policy whose costs fall primarily on those who are already the most heavily taxed? In the interest of making the overall tax system flatter, and more like a consumption tax, why not favor the McDonald’s tax?
A mathematician’s argument.
(High praise.)
Two points
re comment 1
Steve explains this with the dead loss. He is saying that while Romer might be right, by chance as it were, in the present circumstances she could also apply the same argument at times where the deadweight loss would be greater with EITC. So he is pointing out her logic is not as compelling as it looks just because it seems to fit the current facts. His argument however applies in either case.
re comment 2
I don’t agree but this (rather than 1) is indeed a cogent reply to Steve’s argument.
I think you will find that most of her ilk would happily suffer a dead weight loss to help the poor IF they could somehow engineer policy such that the full weight of that loss is felt by the wealthy.
OK so GDP falls by 10% but the poor can now support their families comfortably and the only impact is that Warren Buffets’ wealth is reduced by 95% and still has the life that dreams are made of.
Granted, she doesn’t explicitly say that but I’ll bet my hat that’s what she thinks.
The “efficient level” of unskilled employment and unskilled wages is probably below the optimal level from the viewpoint of macroeconomic utility if you take negative externalities resulting from unemployment and low low-end wages into account (crime, lack of social mobility, loss of stake-holder mentality etc.).
Assuming (as many supporters of the minimum wage don’t) that the demand for labor is downward sloping in the range around the new and proposed minimums:
– The EITC as a subsidy – it will increase employment but also tend to lead to less investment in capital as a consequence. It will also lead to the employment of less productive workers and (perhaps) lead to acceptance of slacker work practices (since the employer doesn’t pay the full wage)
– The minimum wage is neither a tax nor a subsidy but a price floor. It will reduce employment. Some parts of its costs will also be met by taxpayers in benefits to the newly unemployed. However employers will have an incentive to invest in more capital as a substitute for low paid workers and workers will have an incentive to be more productive in order to get and keep jobs.
So when you look at the total costs and benefits of these 2 forms of intervention in the labor market then in the long run minimum wage may actually be better at increasing real low wages. It will tend to lead to more productive workers working in a more capital-intensive jobs than EITC.
” The cost of the minimum wage falls, in the short run, on the owners of businesses that employ unskilled labor, and in the long run on the customers of those businesses.”
While those groups certainly face some increased costs from minimum wage laws, why is the primary cost not on those priced out of employment?
Obviously meant “upward sloping”
I guess another way to phrase my comment #5 is that she would happily take a world where the total wealth is far lower if the minimum wealth is far higher at the cost of the top 1%.
Everything else is just window dressing.
Steve,
Is there any evidence to suggest, that America has a chronic underemployment issue at below minimum wage levels, and that her logic doesn’t just apply to our currently depressed economy? Maybe this is an argument for immigration rather than an EITC?
@Daniel 11, this BLS chart shows a correlation between unemployment and income: http://www.bls.gov/emp/ep_chart_001.htm . The implication of the chart is that they are not directly correlated to each other, but rather via correlation to education.
Unrelated to the topic at hand, I would like to request a posting and discussion of QE, if I may. I was recently reading an article by Nouriel Roubini that discussed the state of the Velocity of Money, which caused me to look at this chart: http://en.wikipedia.org/wiki/File:Velocity_of_M2_Money_Stock_in_the_US.png
What is the implication of continued growth in the money supply, if GDP does not grow in parallel? It seems to me that QE isn’t accomplishing anything – or rather that it is not accomplishing what the Fed is trying to achieve. Is QE a factor in the run up in stock prices? As a non-economist, I am curious.
Al V. 12:
I’m not really sure why that matters? Are you suggesting that people with lower education are unemployed because they want to be unemployed? Then they really shouldn’t be counted as unemployed and we have a measurement problem. Regardless, increasing the EITC would still increase the supply of labor at the lower end and increase overall employment. Are you somehow suggesting that people with lower wages don’t respond to incentives to work?
Steve:
Also, when in the past 30 years have you felt we’ve had an over-employment problem? Over-employment can be adjusted for by raising interest rates, unemployment arguably can not always be adjusted for by lowering interest rates when we reach a zero lower bound.
‘…why is the primary cost not on those priced out of employment?’
Or even on those low skill workers who find themselves subject to harsher supervisory practices, to ring greater productivity out of them after a hike in the minimum wage?
Arnold Kling’s 3 axis model might explain why people prefer minimum wage increases to EITC. Those people want the employers to bear the cost because they view those employers as oppressors on the oppresor/oppressed axis and they want to punish them. The few minimum wage advocates that realize that customers will all bear costs rationalize it as a sort of sin tax (“poor people shouldn’t eat at McDonalds anyway”).
John Schmitt argues that efficiency gains are the most important channels absorbing minimum wage increases. For example, after a mimimum wage increase, staff turnover drops drastically. Since the costs of hiring and training new staff are a significant cost factor (and staff turnover is also a major drag on productivity), it’s quite possible that perhaps the majority of the higher wage cost is in fact absorbed by lower hiring and training costs and higher productivity.
http://www.cepr.net/documents/publications/min-wage-2013-02.pdf
@Daniel 14, what I was (inartfully) suggesting is that there is low demand for unskilled labor in the U.S., as shown by the unemployment rate; and that an oversupply of unskilled labor keeps wages down. That there is an oversupply of unskilled labor is good for businesses that make use of unskilled labor.
The low unemployment rate, and high wages, for highly skilled labor shows that there is a shortage of highly skilled workers. Thus, it seems fairly obvious that one way to address this is to restrict immigration of low skilled workers, and encourage immigation of highly skilled workers. Or, we could improve the skills of low skilled workers. How to do that? Hmmm….
I would doubt, though, that increasing the minimum wage does anything for either problem. I would predict that increasing the minimum wage will benefit some businesses over others – for example, benefit Chipotle over Taco Bell: Taco Bell has to pay workers more, increasing costs, and causing them to increase prices. Chipotle pays above minimum wage already, so they don’t have to increase prices, and thus the price-sensitive buyer is more likely to now choose Chipotle over Taco Bell.
Al V,
I don’t think what you showed necessarily means that there’s low demand for low skilled workers. I think what you may have showed is that wage earners at the bottom might not be willing to work for that low wage. However if they are subsidized (as would be the case with the EITC) they would be more likely to work.
Also if there’s indeed low demand for low skilled workers than the EITC would increase demand for low skilled workers since they would be willing to work for less wage.
You bring up the minimum wage, which I did not advocate for. I prefer the EITC. In the case of the EITC, both the employer and the low skilled worker benefit from the subsidy, the only one harmed is the average tabs payer.
This is from the article, and says exactly what Steve is saying about costs? What am I missing?
Romer also points out, by the way, that some of the benefits of the EITC will go the businesses hiring these workers and eventually to the customers of those businesses.
An too much low skilled employment may be preferred to too little, given our other existing programs to assist the poor. The EITC may allow for a decline in these other programs, while a high minimum wage may cause a greater use and cost of those other programs.
You may know more about per se than I do, but are we sure they do not pay minumim wage to their bus-boys and kitchen porters, and even waiters?
I mentioned this is the other post, but why does EITC not push down wages?
A guess at the answer to my last question – provided by a look back to the other post
“therefore that the wages of unskilled labor are necessarily, via the forces of competition, set equal to the marginal product of unskilled labor, for reasons explained in every elementary economics textbook.”
“there is no monopsony power in the market for unskilled labor and hence no potential for exploitation.”
What if all employers know that unskilled people were prepared to work for $3 an hour in preference to starving, whatever the value of their marginal product? Could not this create a de-facto monopsony, or even a monopoly?
I’m confused why so many people are trying to link the effects of the minimum wage to unemployment. Shouldn’t our primary concern be to link the effects of the minimum wage to consumption? Who cares if the minimum wage makes individuals in aggregate more or less employed? If the end result of a policy increases total consumption by all individuals, I would view that as good policy, regardless of what it does to the unemployment rate.
I, for one, am skeptical that either the minimum wage or the EITC increases total consumption across the economy. But, it would be nice to get some actual data that verifies/refutes that claim. Anybody know of a study that does just that?
Invisible Hand:
You are pointing to an issue that I’ve raised repeatedly re Paul Krugman in particular. He has a long-standing habit of telling us that such-and-such a policy is good because it has one particular benefit, without making any attempt (or at least any attempt he’s willing to tell us about) to enumerate the costs, weigh the benefits against those costs, and give us some framework for evaluating the net benefit. As you say, the minimum wage is ultimately beneficial on net only if it increases consumption streams by more than enough to compensate for any additional work it brings forth — and its proponents have an obligation to explain why they believe that will happen.
Thanks for this comment; it hits squarely on the head a nail that deserves to be hammered repeatedly.