The Myopic Empiricism of the Minimum Wage

You don't find Neumark & Wascher's critique of Card & Krueger's NJ-PA study persuasive? Neumark & Wascher pointed out that C&K used "full-time employment" (FTE) as their measure of employment. FTE counts each full-time worker as "1" and each part-time worker as "0.5", without regard to how many actual hours they work. But the neoclassical prediction about the minimum wage was never about the number of workers per se; it was always about the number of worker-hours. When N&W redid C&K's study using actual payroll records, C&K's positive employment effect turned into a negative effect (although the result was admittedly not significant at traditional levels of confidence). I find the critique of FTE devastating, irrespective of the empirical results.

Card and Krueger's best point was showing that the past empirical work on minimum wages was beset with data mining and publication bias.

I wonder how a law introduced to exclude women and others from the workfoce can some how now be to their benefit?

see Leonard, Thomas C. (2005) “Protecting Family and Race: The Progressive Case for Regulating Women's Work.” American Journal of Economics and Sociology 64(3): 757-791.

see too Leonard, Thomas C. (2000) “The Very Idea of Applying Economics: The Modern Minimum-Wage Controversy and its Antecedents.” In Roger Backhouse and Jeff Biddle (eds), Toward a History of Applied Economics, History of Political Economy, Supplement to Vol. 32, pp. 117-144.

As I remarked when Tyler Cowen blogged on point #1, the model is consistent with both a horizontal labour demand curve and the contention that the minimum wage increases low-wage employment:

This can be consistent with elastic labour demand: a rise in wages causes a large fall in labour demand, but at the bottom end, this effect is swamped by the decline in employee turnover and the easier retention of high-quality employees (low-quality employees are unambiguously expelled from the labour market; the C+K employment effect is merely that high-quality employees spend less time in search).

The variable that gets squeezed from both ends is the number of unoccupied vacancies. The minimum wage just forcibly internalizes the cost of search. Obviously, it only works where the pay and productivity are so low that the search cost is a significant fraction of it.

It is possible that both you and Borjas could be right? By that, I mean in the grand and general scheme of things the Labor Demand Curve is downward sloping, even for unskilled labor; but near and at the margin, it could be that the curve flattens.

JD -
Borjas is specifically concerned with this low-wage, unskilled group.

I think the downward sloping labor demand curve completely misses the point. Card and Krueger think the demand curve is downward sloping. What distinguishes them is that they think the supply curve facing the firm is upward sloping.

Sorry - should clarify. In the study Bryan cites Borjas looks at a range of skill cells. In other work he targets the low-wage low-skill groups and finds the same results.

This is not an argument over the demand curve, IMO.

How then can I continue to embrace (and teach!) the textbook view that the minimum wage significantly reduces employment of low-skilled workers?

I think the key word is *significant*. It's not hard to imagine that for most people, the social benefits of a minimum wage law may easily outweigh the disemployment effects.

Now, of course, this won't make sense for those who cannot see, understand or acknowledge the social effects of minimum wage laws, but given the widespread support for such laws among economists, It appears that most do.

@Tom west Is there widespread support among economists for these laws? Every time someone mentions raising the minimum wage, it seems to me that most economists agree that it would be a bad idea.

Do these same people actively campaign to abolish minimum wage laws? No, but I think all that says is that the personal cost to them of being called 'heartless' by non-economists is larger than the personal benefit of having the minimum wage abolished.

I think the key word is *significant*. It's not hard to imagine that for most people, the social benefits of a minimum wage law may easily outweigh the disemployment effects.

I doubt *most* people ever saw any calculation of benefits of minimum wage outweighing disemployment effects.

That said, I have a problem saying that forbidding the lowest-productive to work in order to slightly raise a pay of people with not-that-low-productivity could somehow result in 'net social benefit'. I mean, if it does, wouldn't killing these people convey even bigger net social benefit?

@Kevin: there is a panel of economists run by the University of Chicago school of business. Here are recent results regarding minimum wage: http://www.igmchicago.org/igm-economic-experts-panel/poll-results?SurveyID=SV_br0IEq5a9E77NMV
There isn't a consensus, but it's much more positive than negative.

I cannot agree that "David Card and Alan Krueger's famous research on the topic is well-done." It is incomplete, at best, as it elicits more questions than it answers.

1) What happened to the workers whose skills were only worth $4.25? They would near certainly be removed from the labor market. Were they replaced by other workers who would have gone to college if their skills were only worth $4.25, but the $5.15 enticed them to take the fast food job instead of college? A displacement of one worker, and the stifling of the education of another are not positive social benefits.

2) Why do similar studies ignore the head counts on the welfare rolls, and those who have exited (or never entered) the labor market? Unemployment counts are not a complete count of the alternatives to work. I will have a hard time believing that increasing the minimum wage doesn't increase the welfare count, or those who are not counted in the potential labor market. Unless, like in 1) above, there are persons who choose welfare at the $4.25 wage alternative, but will enter the workforce at $5.15. In that case, I would expect the rise in minimum wage to create measurable, commensurate rise in productivity.

On another, perhaps similar, topic, can an economist explain the price/demand curve for Louis Vuitton products? If you've never been in one of those stores, it is worth a visit. They sell vinyl handbags for $1500. They are in extremely high demand BECAUSE of their high price. Psychology trumps economics.

The main reason why I potentially wouldn't believe those priors in the face of contrary medium-strength empirical evidence would be the paradox of toil - did you address that one here?

http://en.wikipedia.org/wiki/Paradox_of_toil

There are vast empirical literatures studying the effects of price controls of housing (rent control), agriculture (price supports), energy (oil and gas price controls), banking (Regulation Q) etc...If you object, "Evidence on rent control is only relevant for housing markets, not labor markets," I'll retort, "In that case, evidence on the minimum wage in New Jersey and Pennsylvania in the 1990s is only relevant for those two states during that decade." My point: If you can't generalize empirical results from one market to another, you can't generalize empirical results from one state to another, or one era to another. And if that's what you think, empirical work is a waste of time.

Because clearly generalizing from oil drilling to minimum wages is exactly as valid a generalization from minimum wages in one decade to another in the same state...

You say "top papers", but the only one anyone ever talks about is Card and Kreuger 1994.

What are these other top papers you speak of?

And if there is really only one paper of interest, how can anyone draw any conclusions whatsoever about the state of research a field based solely on one result?

Shouldn't the study be repeated by other researchers in different states to see if the results are consistent? Maybe this one is a fluke. Maybe there were confounding factors (say an improving overall job market in 1993-1994). If you really want to do this properly you have to do independent studies across different contexts.

That is a great piece.

Our ultimate goal should be to get rid of the errors that the minimum wage has come to represent.

Let's abolish intellectual laziness. Problem solved. ;)

"Why then do I make such a big deal about it? Because it is a symbol of larger evils."

ALMIGHTY and most merciful Father; We have erred, and strayed from thy ways like lost sheep. We have followed too much the devices and desires of our own hearts. We have offended against thy holy laws. We have left undone those things which we ought to have done. And we have done those things which we ought not to have done; And there is no health in us. ...

The "system of natural liberty," if I may call it that, presupposes certain moral dispositions or sentiments on the part of its participants. A fellow named Smith wrote a book about that once. Now, what you're saying about the minimum wage being "a symbol of larger evils" sounds like a moral prescription that wants to somehow confine itself to means rather than ends, substituting economic "efficiency" for political justice. How is that intention not itself a "symbol of larger evils"?

Labor unions represent workers who make more than the minimum wage. And they support increasing the minimum wage. Why? To cause an increase in the demand for the workers they represent, not out of the goodness of their intentions.

sounds like a moral prescription that wants to somehow confine itself to means rather than ends

If intellectual rigor and honesty is a "moral" prescription, so be it. I see no other means to any ends than an honest search for truth.

I fail to see that any amount of false or incorrect research that minimum wage laws will help low wage workers will actually cause minimum wage laws to actually help wage workers. Why isn't your first interest ALWAYS in the intellectual "means" to finding out what really works?

substituting economic "efficiency" for political justice

Presupposing that (a) increases in wages constitute political justice, and (b) that minimum wage laws will actually acheive that intended effect.

"Our ultimate goal should be to get rid of the errors that the minimum wage has come to represent."

By this, may I presume you are not advocating the abolition of wage labor?

Since Lassalle's death, there has asserted itself in our party the scientific understanding that wages are not what they appear to be -- namely, the value, or price, of labor—but only a masked form for the value, or price, of labor power. Thereby, the whole bourgeois conception of wages hitherto, as well as all the criticism hitherto directed against this conception, was thrown overboard once and for all. It was made clear that the wage worker has permission to work for his own subsistence—that is, to live, only insofar as he works for a certain time gratis for the capitalist (and hence also for the latter's co-consumers of surplus value); that the whole capitalist system of production turns on the increase of this gratis labor by extending the working day, or by developing the productivity—that is, increasing the intensity or labor power, etc.; that, consequently, the system of wage labor is a system of slavery, and indeed of a slavery which becomes more severe in proportion as the social productive forces of labor develop, whether the worker receives better or worse payment.

@John B. Chilton, re: labor unions.

It is even more insidious or diabolical than that. In many union wage agreements, the minimum union scale is an amount (in $ or %) higher than the statutory minimum wage. So, they benefit DIRECTLY, as well as indirectly through decreased wage competition. So, a higher minimum wage can increase the pay of a union worker directly.

Hazel Meade asks, "Why isn't your first interest ALWAYS in the intellectual "means" to finding out what really works?"

Let me assure you that is ALWAYS my first interest. You seem to assume that I am a proponent of increasing the minimum wage. Actually, I view such measures as woefully inadequate. What I propose is an entirely different social accounting system that would make a concept such as minimum wage superfluous.

When Bryan Caplan opposes the "larger evils" symbolized by the minimum wage, I don't suppose the larger evils he has in mind have to do with poverty and social domination as much as with deviations from the ideal of a particular economic ideology, which I referred to as "the system of natural liberty" but that otherwise is known as economic liberalism or free market fundamentalism.

Bryan says:

"1. The literature on the effect of low-skilled immigration on native wages. A strong consensus finds that large increases in low-skilled immigration have little effect on low-skilled native wages. David Card himself is a major contributor here, most famously for his study of the Mariel boatlift."

As I pointed out in 2006, the Mariel boatlift to Miami in May 1980 was followed by several years of vast prosperity in Miami fueled not by immigration but by cocaine: e.g., Scarface 1983 and Miami Vice 1984. Of course, wages held up in Miami -- billions in illegal money was flowing in in the early 1980s.

http://www.vdare.com/articles/george-borjas-vs-david-cards-unworldly-philosophy

"This consensus among immigration researchers is so strong that George Borjas titled his dissenting paper"

No, the consensus among labor economists who have mastered their topic is Borjas's view. Instead, drive-by papers by guys who haven't thought through even their own studies, like Card or Peri, are popular with economists who aren't experts on labor economists for ideological reasons.

I second Daniel Kuehn's point that the debate isn't really about whether the demand for labor is downward-sloping. The key to the monopsony-power model is an upward-sloping supply curve facing the firm.

"Labour is a commodity like every other, and rises or falls according to the demand." – Edmund Burke

...every other commodity when brought to market, if you cannot get the price intended, it may be taken out of the market, and taken home, and brought and sold another day; but if a day's labour is offered on any day, and is not sold on that day, that day's labour is lost to the labourer and to the whole community... - William Longson

"Labour should not be regarded merely as a commodity or article of commerce." - Treaty of Versailles

"Labor is not a commodity." - ILO Declaration of Philadelphia

"We must now examine more closely this peculiar commodity, labour-power." – Karl Marx

Whether labor is or is not a "commodity like every other" is obviously a matter of considerable controversy. Therefore, whether or not one treats it as if it is a commodity-like-every-other in empirical economic analysis requires something more substantive than dogmatic assertion qua Burke.

To what extent do people suppose the Inca Empire could have survived without the khipu or quipu system of keeping accounts? To what extent could capitalism on a large scale prevail and prosper in the absence of double-entry bookkeeping? There is no equivalent systematic social accounting from the perspective of labor, only various ad hoc regulatory ointments and band-aids. John Maurice Clark hinted at such a social accounting for labor in his Studies in the Economics of Overhead Costs.

I invite those who are critical of ineffectual ad hoc band-aids to join me in innovating genuine remedies. Reciting the dogmatic assertions of some 18th century reactionary is not my idea of "intellectual rigor and honesty."

I like to err on the side of not getting my info on the interaction of crime/econ/immigration from fictional movies and TV shows.

But that's just me.

The analysis in the post and the comments does seem to undervalue the transitive effects that could lead to inelasticity in the short-term and elasticity in the long-term.

In other words, the analysis seems to take P = MC markets somewhat too literally, which then assumes that labor markets respond instantaneously. Instead, I think the depreciation and reinvestment cycle of business matters a lot to how quickly minimum wages lead to unemployment.

Let's say that a fast-food restaurant faces a perfect substitute in people cooking their own meals. Therefore, the fast-food restaurant is a price-taker and cannot earn the same margin when wages are increased.

But it does not make sense either to close the restaurant completely if the gross margin is still positive. Whether for the owners or creditors, the restaurant is a sunk cost and stays open. The same number of employees are employed, at a higher wage with the wages coming out of the return on capital.

Everything depreciates though and restaurants which do not make a good enough return on capital either close or they invest in more automation. Disemployment results...eventually.

The effects of minimum wages are somewhat different than nominal downward rigidity with slack demand. Demand is constant in the minimum wage case and only capital depreciation will cause the restaurant to close. If demand decreases and wages don't, the maximum cash flow for the owner is to immediately lay some people off.

It's like saying rent control doesn't put people out on the street overnight and it does benefit the current tenants. The buildings already built and a vacant building is not much use to the owner. If supply and demand are already balanced, the effect is merely a transfer from landlord to tenant. But then both people move and the capital stock needs to be replenished, and scarcity results. In both cases, there is a lag which could provide some conflicting evidence but long-term (NYC rent control or European labor markets), the effects are obvious.

The key to the monopsony-power model is an upward-sloping supply curve facing the firm.

the monopsony argument has always been a complex joint hypothesis of increased employment, increased output and lower prices. The reviews of the book in the Industrial Relations Review in about 1993 are all worth attention.

The new monopsony arguments is based on search and matching frictions

Manning in monopsony in motion (2005) accepts that above normal profits will attract entry.

If we start with no excess profits, Manning puts forward two opposing effects of a minimum wage: an employment effect as the wage rises – the oligopsony effect - and an employer exit effect where firm failures reduce employment.

Peter Kuhn in a great review of monopsony in motion pointed out the correct title was search fictions with wage posting and random matching in motion

Tom West's comment is one way of presenting what I think really is the last word, or bottom line, on this topic (some people are simply not in the business, or within the discipline, of recognizing certain aspects of the subject):

"How then can I continue to embrace (and teach!) the textbook view that the minimum wage significantly reduces employment of low-skilled workers?

I think the key word is *significant*. It's not hard to imagine that for most people, the social benefits of a minimum wage law may easily outweigh the disemployment effects.

Now, of course, this won't make sense for those who cannot see, understand or acknowledge the social effects of minimum wage laws, but given the widespread support for such laws among economists, It appears that most do."

Likewise with Sandwichman's comments. Most instructive.

Thought a lot about this one. I think at some level you're fundamentally begging the question. You want to prove to us that theory is correct, and you cite empirical evidence elsewhere and apply it to the minimum wage by assuming theory is right. Developed my thoughts in much more detail here:

http://ashokarao.com/2013/03/13/bryan-caplan-is-wrong-about-the-minimum-wage/

I think reducing low skilled labor is a good thing. Low skilled labor impedes the progress of automation technology. One of the ultimate goals of technological progress is to replace labor all together.

"Part of the reason is admittedly my strong prior."

Come to think of it, Bryan Caplan's "prior" is actually "priors", one of which is the notorious lump of labor! Caplan is "99%+ sure that a randomly selected demand curve will have a negative slope." But a demand curve for WHAT? For hours of work or for workers?

Caplan assumes that a negative sloping demand curve for hours of work transforms automatically into more jobs rather than, say, into more hours -- at lower pay -- for fewer workers. What's the empirical evidence for that assumption?

As Maurice Dobb pointed out long ago, even if the elimination of the minimum wage led to higher aggregate earnings for low wage workers, "It is not aggregate earnings which are the measure of the benefit obtained by the worker, but his earnings in relation to the work he does."

This is silly.

You could just do an Econlit search on the topic. You'd find Card-Krueger has been completely discredited.

The end.

A summary of the state of the art, if you're really interested.

http://www.cepr.net/documents/publications/min-wage-2013-02.pdf

http://www.cepr.net/documents/publications/min-wage-2013-02.pdf

As Schmitt points out:

"The minimum wage does not raise the cost of hiring workers – it raises the cost of hiring an hour of work performed by those workers. Even within the competitive framework, employers might choose to respond to a minimum-wage increase by reducing workers' hours, rather by reducing the total number of workers on payroll."

Which is essentially the point I made in my comment above!

http://econlog.econlib.org/archives/2013/03/the_vice_of_sel.html#254511

Jim Rose: The new monopsony arguments is based on search and matching frictions

From what I understand, there's quite a variety of arguments. Burdett and Mortensen (1998) talk about frictions. But there are quite a few others.

In Stigler's 1946 model, firms are not price-takers in the labor market, but have a degree of market power (or monopsony power) over the price, and thus labor was being paid less than the MRP and MC and AC cost of labor differ (firms are extracting rents). If the minimum wage increases AC without increasing MC, the rents decrease, but the optimal level of employment stays the same (or can even increase). Rebitzer and Taylor's 1993 model uses shirking, where a higher wage increases the disincentive for workers to shirk (much like efficiency wages).

The above two fit within the neo-classical framework. There are many others approaches, like that of the Institutionalists or approaches that simply drop the assumption that firms and workers are behaving rationally. Eithne Murphy has a 2006 survey, "Minimum Wages and Employment: A critical appraisal of two theoretical frameworks" which discusses these.

Sorry, link here.

Matt Waters raises an important point. What do we expect to see? I asked in the economics subReddit, suggesting a transmission mechanism that looks laggy enough to make empirical work very hard.

Unlike Bryan, I had no prior bias on the min wage, but I now am firmly convinced it's a terrible idea. David Neumark's book Minimum Wages (app 300 pages) simply demolishes the case for a min wage. Not only does it increase unemp, it actually increases poverty and reduces skills.

"The Demand Curve for the U.S. Minimum Wage" March 8, 2013

How much will President Obama's 2013 State of the Union proposal to increase the federal minimum wage to $9.00 per hour affect the teens and young adults who make up roughly half of all those who earn the minimum wage or less in the United States? ...

... we estimate that the number of 15-24 year old Americans with incomes would decline by over 1.8 million from 2011's figure of 26,014,000 to the 24,192,580 figure estimated by our tool above, ...

* * *

From 1994 through 2011, ... the U.S. Census Bureau reports that there has been a net decrease of 1,012,000 teens and young adults with incomes. ... virtually all of the negative change in the number of Americans Age 15-24 with incomes has occurred at annual incomes that fall below $15,000.

* * *

... after adjusting for the effect of inflation, the total amount of income earned by American teens and young adults in 1994 and in 2011 is virtually identical. Increasing the minimum wage does not increase the amount of money available to pay wages and salaries ...

@ Walter Sobchak

That page is hilarious.

Their regression is based on data '94-'11, but they exclude the years '97-'03, because "they are clearly the result of an atypical situation for the U.S. economy."

But they include the years '09-'11, which apparently are a typical situation for the U.S. economy.

Can you think of a more egregious instance of cherry picking?

Switch the exclusions, and the regression points in the opposite direction.

C'mon: really???

Congratulations Steve, you completely missed the point.

I would be happy to support the abolition of wage controls, such as the minimum wage.

In its place we should institute a negative income tax. That way individuals can choose to work at any wage their employer proposes and labor would reflect the benefit to their employer but no one would starve.

It seems that Caplan's contention assumes a fixed labour demand curve. This is an assumption worthy of some debate.

In respect of the M. boatlift, could it be that increases in the supply of low-skilled labour raised the demand for said labour (i.e. an outward shift of D curve)? After all, these new arrivals will spend much of their wage income locally. This would, to some extent, offset the wage depressing effects of the increase in supply.

A similar argument can be made for the min. wage. If higher wages lead to higher consumption expenditure (a reasonable assumption at this end of the income distribution) then it is reasonable to expect some increase in labour demand.

In sum, shifts in labour demand can reconcile these seemingly conflicting results.

I want to stress that I am not a zealot on either side of this debate: I remain very much open to persuasion based on the evidence. But I do question overly simplistic textbook analyses that insist on treating the labour market like the market for any other product: there are endogeneity issues in the labour market that are absent in the market for potatoes.

More data is clearly needed to settle this debate conclusively (which requires more policy experimentation). This is less likely to be forthcoming whilst some economists peddle ideologically driven (a charge to which Caplan confesses above) analyses based on questionable assumptions.