David R. Henderson  

Hooves and Priors

Interpersonal utility comparis... Voters in successful neolibera...

In "Yes Indeed: I Have My Priors About the Minimum Wage," Don Boudreaux at Cafe Hayek writes:

So an economist whose priors tell him or her that raising the hourly cost of employing low-skilled labor will cause employers to choose to employ fewer hours of low-skilled labor is falling into no unscientific or dogmatic trap. Such an economist is behaving no less scientifically than does, say, a biologist whose priors tell him or her that a zebra's stripes or a trout's dorsal fin were formed over time by natural selection. A renegade biologist who insists, say, that a trout's dorsal fins are unique among all animal features in having been formed by some process distinct from - and, indeed, at odds with - natural selection would have a huge burden of proof to overcome. Not only that, no one would expect other biologists to immediately, simply upon hearing one or a handful of biologists claim unique origins for a trout's dorsal fin, to start regarding natural selection as a process that works only sometimes. All good biologists, in fact - even those who do not specialize in studying fish - would correctly insist that science, far from requiring that they withhold judgment about whether or not natural selection formed the trout's dorsal fin, requires that they, given the current state of knowledge in biology, regard natural selection as applying to a trout's dorsal fin no less than it applies to other biological phenomena.

He says it well. I endorse his thinking, at least the part about economics. I don't know enough about natural selection to say that part is completely sound. I'm quite confident about the law of demand.

But I'm not writing to take issue with natural selection. I'm writing to make an analogy between what we're pretty sure of from economics and what doctors are pretty sure of. About 25 years ago, I had some symptoms that made me think I might be having a heart attack. I called my doctor on the weekend and his exchange connected me with one of his partners. The partner asked about my symptoms and what I had been doing just before having chest pains, and he concluded that the odds were high that it was just gas. He told me to eat some bread and see if the problem "passed," adding that the double meaning was unintended. It did. Later, when I went to see his partner, my regular doctor, to discuss something else, I mentioned that incident. He smiled and said that one of the most important lessons he learned from one of his teachers in medical school was:

When you hear hooves, think horses, not zebras.

Comments and Sharing

COMMENTS (21 to date)
Don Geddis writes:

Economics theory of course provides tremendous support that higher minimum wages would reduce low-wage employment. And that theory of supply and demand is supported by overwhelming empirical evidence.

All that said, however, the actual specific empirical evidence on the disemployment effects of minimum wages, is surprisingly ambiguous. It's no longer "simply upon hearing ... a handful ... claim" status. There are now sufficient studies, of sufficient quality, that the situation must at least be "we don't quite understand what is going on", rather than "there is no reason to doubt that a trout's dorsal fin came from natural selection." The doubts are well past the status of mere unsupported claims.

Daniel Kuehn writes:

I agree strongly with Don Geddis on the problems with Don's characterization of the empirical evidence, but I think the other important point is that the people that think the minimum wage does not have substantial disemployment effects are not (that I am aware of) suggesting the law of demand is wrong. (Just as Card, in his immigration work, is not suggesting demand curves don't slope down).

Not only does a prior about the law of demand not get you to the view of that the minimum wage has to reduce employment, there is also no sense in which low wage labor is a special case, as Don suggests. People who think bargaining power, market power, and frictions are important don't think these things only exist in the low wage labor market.

ThomasH writes:
So an economist whose priors tell him or her that raising the hourly cost of employing low-skilled labor will cause employers to choose to employ fewer hours of low-skilled labor is falling into no unscientific or dogmatic trap.

I agree, but that prior plus any amount of evidence that the demand curve for low paid labor slopes downward is not enough to conclude that the minimum wage is a bad way to transfer income to low paid workers. One must know the elasticity. If it is low then the number of fewer hours that employers will chose will be small and the many workers who remain employed will gain a lot more income that the small amount lost by the few who do lose their jobs. In economics, there is no such thing as a free lunch.

Of course there are better ways to transfer income such as the wage subsidy that Scott Sumner favors or raising the EITC, but strangely, opponents of the minimum wage (except Scott) seldom mention these better alternatives and help make them acceptable to "conservatives."

I prefer Scott's approach as when he criticizes the taxation of capital income; he has an alternative, a progressive consumption tax.

Don Boudreaux writes:


With respect, you miss the point. Of course the question of whether or not the total pay of low-skilled workers rises or falls as a result of raising the minimum wage is an empirical question - the answer to which, as you correctly note, depends upon the elasticity of demand for such labor over the relevant wage-change range.

But the basic economic case against the minimum wage is not that it reduces the total pay of low-skilled workers as a group. It is, rather, that the minimum wage reduces the employment options of some low-skilled workers (probably, although not necessarily, by causing some of these workers to be unemployed). The basic case against the minimum wage stands even if raising the minimum wage increases the total pay of low-skilled workers as a group.
If pro-minimum-wage politicians and others truly believe that the minimum wage is justified if it results in a net transfer of income to low-wage workers as a group even if, in the process, it pushes some such workers indefinitely into the ranks of the unemployed, let them say so publicly. Let these minimum-wage advocates admit publicly and unambiguously that raising the minimum wage will raise the incomes of some low-skilled workers but only at the expense of forcing other workers into the unemployment ranks.

I have never once heard a pro-minimum-wage politician recognize this trade-off publicly. The sale of the policy is always that it will help all low-skilled workers, perhaps at the expense of employers or at the modest expense of millions of consumers who pay higher prices. Never, ever though do politicians publicly recognize the likely cost borne by the workers who are the ostensible beneficiaries of the policy.

The public-policy debate on whether or not the minimum wage is a good device to transfer income to low-skilled workers as a group is entirely different from the debate on whether or not the minimum wage helps some low-skilled workers only by inflicting harm on other low-skilled workers. The debate that I generally participate in is the latter one (although I have strong opinions about the former, too).

My strong suspicion is that if most Americans believed (whether correctly or incorrectly) that raising the minimum wage actually throws some low-skilled workers into the ranks of the unemployed, they would oppose such legislation regardless of whether or not that legislation increased the pay of low-skilled workers as a group. I hold this suspicion because, again, politicians routinely insist that economists who point out that raising the minimum wage will destroy some low-skilled jobs are wrong.

Tom West writes:

If pro-minimum-wage politicians and others truly believe that the minimum wage is justified if it results in a net transfer of income to low-wage workers as a group even if, in the process, it pushes some such workers indefinitely into the ranks of the unemployed, let them say so publicly.

I'm no politician, but that's exactly what I believe. Within reason, the benefits of the increasing net transfer of income exceed the the costs of the job loss.

(I'll add that I believe that as an aggregate, we humans don't seem to be able to resist forming an association between what we pay for something and how its valued. My observation is that minimum wage makes workers more valued (by employers 8and* by the workers themselves) simply because they're paid more.)

However, as a politician, you are forbidden from ever acknowledging the costs of any policy you support. It makes open politics useless for realistic discussion of policy issues, but that's just how modern politics works on any side of the aisle.

Nathan W writes:

The problem in his portrayal is that economists agree less with dogma than biologists do.

If the supply response sufficiently outweighs the demand response then you can have more jobs at higher minimum wage.

Don Boudreaux writes:

Nathan W:

Help me out here, because I don't understand what you're saying when you write

If the supply response sufficiently outweighs the demand response then you can have more jobs at higher minimum wage.

The supply response to a higher minimum wage is an increase in the quantity supplied of low-skilled labor; the demand response - save in the curiosity case of genuine monopsony power among employers - is a lower quantity demanded of low-skilled labor. No amount of the supply response being larger or smaller offsets the fact that the higher the minimum wage, the lower is the quantity demanded of low-skilled labor. (Indeed, the larger the supply response, the greater is the amount of formal unemployment caused by minimum-wage legislation.)

In short, the problem with minimum-wage legislation is that it causes a demand response: specifically, it reduces the quantity demanded of low-skilled labor. The supply response - no matter its size - does nothing to "outweigh" the resulting destruction of jobs.

ThomasH writes:

Professor Boudreux

First let me (as one of the "others" you challenge) publicly state that I "truly believe that the minimum wage is justified if it results in a [substantial] net transfer of income to low-wage workers as a group even if, in the process, it pushes some [fairly low number of] such workers indefinitely into the ranks of the unemployed." That is the reason that I prefer a EITC increase to a minimum wage increase as a way to increase the incomes of low paid workers and I suppose that most economists would agree.

I have not followed the debate carefully, but I was not aware that proponents of the minimum wage had denied at least the possibility that some people could loose jobs. (I did not read Krugman, for example, that way when he crowed about Walmart having discovered that there were benefits from paying above minimum wage -- benefits that might disappear if other firms were paying the same higher wage.) But if they have, then it's those specific claims that ought to be held up for discussion, not the failure to argue that "the minimum wage is justified because it results in a net transfer of income to low-wage workers as a group even if, in the process, it pushes some such workers indefinitely into the ranks of the unemployed."

I think that we need to be careful to avoid accepting the other side of any argument for which a faulty argument is made. [I say this with some contrition, having supported the invasion of Iraq because the "It's about oil" argument was (and is) silly.]

Don Boudreaux writes:


Economics, as such, can say absolutely nothing about whether or not a policy that artificially raises the wages of the Joneses by casting the Smiths into the ranks of the unemployed is good or bad. Value judgments are necessary to reach any such conclusion, whether in favor or opposed to that policy.

These value judgments can and ought to be informed by facts and sound reasoning. But even if unassailable empirical studies show that the result of raising the minimum wage by X percent will be a Z percent net increase in the incomes of low-skilled workers as a group, economics cannot inform you or anyone else if that policy is worthwhile or not. This reality doesn't change even if the net increase in the incomes of low-skilled workers is 2Z percent - or 1,000Z percent.

Most people (or, at least most Americans), though, do have a value system that rejects a policy of artificially restricting the economic opportunities of some people in order to enhance the incomes of others - which is why I believe that many more people would oppose minimum-wage legislation if they understood that a nearly inevitable result of such legislation is that some low-skilled workers are pushed into the ranks of the unemployed.

I've a mental experiment for you. Suppose that, in order to increase the hourly wages paid to low-skilled workers, government requires that all low-skilled workers – all teenagers, all immigrants with no more than a grade-school education, and anyone who has completed a maximum of 12 years or fewer of formal schooling and who has never before held a job in the U.S. – first obtain a government permit before searching for work. The government declares illegal the employment of any such person who does not have a valid government-issued work permit. And the government enforces this permit policy vigorously.

Further, the government intentionally limits the number of work permits it issues. The number of work permits is purposely kept well below the number of such low-skilled people in the population. By keeping the number of such permits lower than the number of such low-skilled people in the population, the government’s work-permit system – again, by design – artificially reduces the supply of low-skilled labor and, thus, artificially increases the wages paid to those low-skilled workers who are among the lucky ones to be issued permits. (Assume that the permits are distributed randomly among low-skilled people.) Each and every low-skilled person who is unable to acquire a permit is not permitted to work. These permit-less people must remain unemployed.

Would you support this scheme if you could be assured that it results in a substantial net increase in the income earned by low-skilled workers as a group?

The effects of this hypothetical work-permit scheme are nearly identical to that of minimum-wage legislation as it is implemented in practice today. (Actually, this scheme would work a bit better for the most vulnerable low-skilled workers than does real-world minimum-wage legislation - but ignore this point.) Do you believe that the workers who cannot get a permit (because the luck of the draw is against them, despite each individual having had an ex ante fair chance to get a permit) are treated fairly by the system?

One further question: how do you weigh the stream of incomes that would have been, but in fact are never, earned by the workers who are rendered unemployed by minimum-wage legislation? Remember: not only does such legislation cause some workers to lose their current incomes; it also - and in many cases more importantly - causes them to fail to gain work skills and experience that they would have gotten had the minimum wage not priced them out of the labor market. In other words, the stream of incomes lost as a result of minimum wage legislation include not only the incomes from the current jobs that are destroyed but also the higher incomes from the future jobs that these workers would have eventually had - but in fact never get because of minimum-wage legislation. Do you know of any reliable empirical studies that quantify these lost future streams of incomes?

ThomasH writes:

Professor Boudreux:

I certainly agree that economics per se cannot decide questions of policy. I do not believe that anything I said leads to that inference.

I agree that most people think that it's bad to limit the opportunities of some people and good to expand the opportunities of other people. I do not think that everyone would agree that no amount of expansion of the opportunities of some could justify any reduction in the opportunities of others. Maybe I'm an odd ball, but for me it would depend on more than just the magnitudes (which opportunities, why is there a link between the expansion and restriction, starting positions, etc.) but it would depend at least in part on the magnitudes.

I do agree with you that most people who support increasing the minimum wage think about it mainly as a sales tax and think, "Yeah, I don't mind paying a few pennies more for my hamburger" so that if they were aware of the employment loss effect, however small, they'd be less supportive of a minimum wage increase. But having gotten people that far, getting them to a wage subsidy/EITC/exemption of the first $x from SS taxation might be that much easier. Too bad all Liberals do not have Krugman's "conscience."

As for your work permit thought experiment, I'll answer that again it would depend on the numbers of people not gaining/losing jobs -- you say the number of permits is "well below" -- compared to the number of jobs and amounts of income gained by the beneficiaries. One might also worry about implementation failures, administrative costs, unintended consequences, etc. I might accept it tactically thinking that with such an illiberal scheme I could surely get some Libertarians to join an EITC coalition. :)

I don't see why the discounting of future income lost and gained would change the calculus.

In sum, I'd need more information than just the sign of the unskilled worker demand function.

Don Boudreaux writes:


Please feel free to call me Don.

Thanks for your thoughtful response. And please forgive the brevity of this reply. My poor excuse is that it's late in the day and I have other matters to attend to before turning in.

Question: Suppose that the group of people whose income as a group that we're talking about are not those people classified (by some criteria) as "low-skilled" but, instead, people with blue eyes. Would you potentially support a government policy that artificially raises the incomes of some of blue-eyed people while casting other blue-eyed people indefinitely into the ranks of the unemployed if you could be assured that on net this policy would raise the incomes of blue-eyed people as a group? If not, why would you potentially support a policy that does such a thing when the group of people under consideration are those who are classified as low-skilled?

Gene Marsh writes:

Hi Don Boudreaux,

Econ novice here.
Am I to understand that you object to any policy that may result in layoffs?
Your readiness to go to the wall for the least-skilled has me wondering if i need to update my stereotypes.

Thomas Boyle writes:

I've found that many supporters of the minimum wage are college grads who want to feel good about themselves. Often, they play to send, or are sending their own kids to college.

I have asked them this: You support a "livable wage" for low-skilled workers, with no low-skilled worker being allowed to take a job for less than a livable wage. OK. As we know, when someone graduates college, they often have debt they need to service, so their "livable wage" is higher. For a college grad, a livable wage would be, say, $50,000, or $25/hour. Would you support a law that said that if your child finds they cannot get a $25/hour job, it should be illegal for them to take a job for less? A law that says, no matter what, no matter their degree, no matter their circumstances, they must remain unemployed, either until they get a job above the minimum wage for college grads, or for life?

Don Boudreaux writes:

Gene Marsh:

You ask of me (fairly):

Am I to understand that you object to any policy that may result in layoffs?

The answer is unambiguously no (assuming that by "policy" is understood to be any government stance toward the economy). Nearly all economic change results in a shifting of resources, including labor, from some tasks to other tasks. So nearly all economic change results in layoffs. And I'm emphatically a fan of economic change as long as it's driven by market decisions - that is, as long as this change is driven by the freedom of consumers to spend their money as they choose and the freedom of entrepreneurs and businesses to compete for consumer dollars (or yen or yuan or euros or pesos or whatever).

What I do strongly oppose is any government policy that artificially prevents people from contracting on terms that they voluntarily choose.

Layoffs that result from market competition are literally unavoidable if economies are to grow - and such layoffs, when driven by market forces, occur only when, over the long-run, net prosperity increases. Layoffs that result from government-imposed restrictions are, in contrast, completely avoidable; they not only are not necessary for economies to grow, they stifle the growth of economies. Such layoffs are part of polices that reduce net prosperity.

RPLong writes:

The above discussion in the comments is excellent. Thanks to all of you for leaving such great comments for the rest of us.

One thing that stands out in this discussion is a problem I've always had with utilitarianism: Suppose the lives of 9 low-income individuals can be made much, MUCH better off by killing a 10th.

Utilitarianism has no built-in mechanism that keeps us from becoming opportunistic killers, so long as an adequate case can be made on behalf of "the group."

Minimum wage has long been "unemploying" some for the benefit of some others. The problem is that we who are debating the issue have no real stake in the matter. We can sit back from the comfort of our armchairs and move the chessmen in positions that we think better suit them. Short of adopting some other moral code, though, there is nothing to stop us from falling victim to a problematic moral conceit that none of us would tolerate if the job on the chopping block were our own.

ThomasH writes:


Thanks for the opportunity for this kind of exchange. I think that few people really know what we think about an issue until we have to discuss it.

Your question about blue-eyed people is interesting. My answer is no because I'd have no reason to want to benefit blue eyed people at the expense of non-blue-eyed people even if no subset of blue-eyed people were disadvantaged in the process.

My starting point is that the last not so few year have not been good for people with few skills. [I think the main reason for that has been bad macro management -- the failure of the Fed to keep NGDP on track and the failure of Congress and the President to have a system that invests in infrastructure at the state and federal level when the returns to those investments meet the old fashioned NPV greater than zero criteria -- a failure of longer standing than the Fed's. But that's another matter and it's not going to get fixed soon.] Therefore, I'm favorably inclined toward policies that transfer income from the rest of us who have done OK to low-paid workers. This inclination does not work for blue-eyed people.

For many people who feel this way, the first (and only) thing they think about is a law to force firms to pay low wage-workers more. That's unfortunate because they do not take into account that such a law can cause some of the very low paid workers they want to help to become unemployed. It's unfortunate because there are other ways to transfer income to low paid workers that do not cause them to become unemployed. Therefore between an increase in the EITC (especially as it functions as a wage subsidy) and a higher minimum wage, I'd oppose an increase in the minimum wage. [I know some people think (rightly I suppose) that some part of an increase in the EITC with no increase in the minimum wage means that some of the EITC is "lost" in a fall in market wages. This does not particularly trouble me.]

But (in my view of current political reality) that is not an extant choice. The actual political question is an increase in the Federal minimum wage (increases at the municipal level raise other issues) or not. Given this choice, I think the way to make a decision is to look at the amount of income that will be transferred to low paid workers (those that remain employed and will become employed at the new minimum wage) in comparison to the amount of income that will be lost by those that will become unemployed/not become employed because of the minimum wage -- and higher prices paid by consumers and lower returns on investment of firms employing low paid workers.

A starting point of that analysis is to know the elasticity of demand for low income workers. If that elasticity is close to zero -- even if it
is negative -- then very few low-paid workers will lose/not gain jobs and little of their income will be lost. [This means the minimum wage will mainly be a kind of excise tax on products produced with low paid labor, not a good idea, either, but that's part of the cost of not being able to chose some other way to finance the income transfer.] The efficiency wage argument -- that some firms were mistaken in the wage they were offering low income workers -- even if true, ultimately is just another factor that would make the elasticity low.

As I see it, none of these considerations arise in an (apparently) arbitrary proposed policy to transfer income to some blue-eyed people at the partial expense of other blue-eyed people.

I often grouse on this blog about the amount of attention that the minimum wage gets in relation to the thousands of other departures from a perfectly Libertarian polity, but it does seem to be useful in clearly bringing out differences in how (at least some) Liberals and (at least some) Libertarians think about policy that would not be so clear in a discussion of, say, the corporate income tax. If so, it's worth our time.

Too often Liberals think Libertarians are simply choosing arguments to support an objective of maintaining the existing income distribution (if not make it even less egalitarian) and Libertarians think that Liberals are interested only on signaling "virtue" with either ignorance or reckless disregard of the actual consequences of their policies (if they are not in fact interested only in increasing state power because they think they will be the ones who exercise it.)

Brian writes:

"Libertarians think that Liberals are interested only on signaling "virtue" with either ignorance or reckless disregard of the actual consequences of their policies (if they are not in fact interested only in increasing state power because they think they will be the ones who exercise it.)"


Yes, that would be because the vast majority of "liberals" (read minimum-wage supporters--they don't have to be liberal) support the minimum wage for the purpose of signaling. It's their way of saying "we should be concerned for the plight of the poor--the government says so." How could this policy be anything BUT signaling, given that very few know or care what the evidence says? On the other hand, libertarians could be accused not of trying to maintain the existing income distribution, but of being opposed to all unnecessary government intervention, regardless of the evidence. I think that would be a fair criticism of libertarians.

But let's face it. Regardless of the evidence, a national minimum wage is a fundamentally illogical policy. If the purpose is to ensure a "living wage" by a single worker supporting a family of four (or whatever criteria one wants to use), it's not going to happen with a single national wage, given the differences in standards of living. A living wage in New York City or San Francisco will be very different from one in some rural town in Mississippi. Any minimum wage that is not locale specific is incapable of achieving the stated objective (producing a living wage). Its only justification, then, is as a signaling device.

In my opinion, government-by-signaling is bad government. If we are going to be stuck with some government, then at least it should work to actually improve society with evidence-based policies rather than simply make people feel better with signaling-based policies.

robert writes:

Are the companies and people that do not follow the law included in the minimum wage impact studies?

The reason I ask is because I worked most of the time I was in High School, and my first job was as a dish washer, and I made significantly less than $15 per hour. I took the job because I wanted to become a busboy like the older boys, and if I did a good job as a busboy, I could become a waiter, which were the people who did really well.

It seems that fewer and fewer teenagers are affected by the minimum wage law, and this seems to be benefit to the politician since teenagers cannot vote and may even be more ignorant when they are old enough to vote. If I were an employer, and I had to pay some below the law, I would rather pay someone already breaking the law since I am less likely to get in trouble. Aren’t most restaurant jobs in NYC being done by immigrants?

On the one hand, the immigrant is better off. On the other hand, the benefit of the minimum wage could lead to other social issues, such as a sectioning off of a class of society since as soon as the immigrant and their family is legal a new immigrant would be needed, income inequality, and younger people with fewer skills.

I know there will be more automation since automation will become less expensive relative to the option, i.e. fixed cost vs marginal cost, especially considering insurance and lawsuits.

I hope that the analysis is well done and accurate.

ThomasH writes:


Regardless of the evidence, a national minimum wage is a fundamentally illogical policy

A national minimum wage linked to the cost of living by locale could work better than an single national minimum but even the latter would not, in my view, necessarily be "fundamentally illogical" if evidence shows that the elasticity of demand for low skilled labor is low and better ways of redistributing income are excluded.

Of course I'd prefer that more Liberals would signal their virtue by supporting an increase in the EITC or a wage subsidy. I think it would be a better signal, too. :)

Brian writes:


It's illogical if the goal is a living wage. On the other hand, if the goal is merely some level of redistribution, the minimum wage is an extremely inefficient way of doing it, considering that 1) many poor people don't work in the first place, or only work part time, 2) only a small fraction of those who do work make the minimum wage, and 3) many minimum wage workers live in households that are far above the poverty line. In other words, the population of full-time minimum wage workers is not what most people imagine it to be. It is really impossible, in my view, for "better ways of redistributing income" to be "excluded" because almost anything is better.

LD Bottorff writes:

I suspect that many people support the minimum wage as a way of redistributing money from business owners to workers. An increase in the minimum wage has the potential to hurt the small-business owner. This sounds fine if you believe all business owners are rich and deserve to have some of their money transferred to their oppressed workers. But many small-business owners are investing their entire savings and borrowing from family. A reduction in the earnings of the small-business owner to pay a higher wage to college-bound teenagers doesn't sound equitable to me.

The original post had to do with priors. My prior is that any government imposed rule makes it more difficult for people to get the best job, not just any job.

Comments for this entry have been closed
Return to top