A frequent complaint of EconLog commenter ThomasH is that the arguments made there (and, presumably, also here at Cafe Hayek) against the minimum wage are insufficiently empirical. Here’s part of his comment on David Henderson’s recent post on liberalizing the market for body organs – a post that ThomasH likes because that post examines both the costs and benefits of market liberalization:
This is very much unlike the minimum wage arguments that are made here [at EconLog], which is basically to argue that if there is any cost (in loss of employment) that settles the issue in the negative, rather than weigh the costs to those that loose or do not gain employment with the befits of the larger incomes of those that do.
It is better to address the real arguments rather than straw man arguments in order to actually be persuasive.
I’m a consequentialist.* I agree with ThomasH that consequences matter and, therefore, that as many as possible consequences ought to be examined rather than only a select few. But I disagree with his criticism of David’s and Bryan’s (and others’) arguments against the minimum wage. ThomasH’s complaint features two flaws that I wish here to address:
(1) he misses the context of nearly all arguments in support of the minimum wage, and
(2) he mistakenly assumes that the categories for judging the costs and benefits of minimum-wage legislation are more objective than they really are.
I deal with each flaw in turn.
(1) ThomasH misses the context of nearly all arguments in support of the minimum wage.
While some academics who argue for the minimum wage concede that minimum wages constrict the employment options available to low-skilled workers, the vast majority of the public pleas for the minimum wage either implicitly assume, or explicitly proclaim, that raising the minimum wage causes no job losses. Therefore, it is this assertion – the assertion that none of the ostensible intended beneficiaries of minimum-wage legislation will suffer any ill consequences as a result of such legislation – that good economists address. As I’ve written on earlier occasions here at Cafe Hayek, I have never heard a pro-minimum-wage politician admit that raising the minimum wage might well cause job losses for some of the very workers who allegedly are the objects of minimum-wage-proponents’ affections. In fact, whenever the prospect of job losses is raised by such politicians, it is raised only to be denied. (Robert Reich, in this video, calls those who warn that the minimum wage will cause job losses “scare mongers.”)
I infer from the virtual absence in public discourse of admissions of the possibility of job losses that even left-wing politicians and pundits worry that that voters will be far less likely to support hikes in minimum wages if voters believed that such hikes might well cause job losses among low-skilled workers. And if voters in fact would react in this way to proposals to raise the minimum wage, one plausible conclusion is that most voters’ value judgments lead them to oppose policies that yield benefits to some at the expense of others who are rendered indefinitely unemployed.
The mention of value judgments is a good segue into point (2) –
(2) ThomasH mistakenly assumes that the categories for judging the costs and benefits of minimum-wage legislation are more objective than they really are.
Ultimately, policies are “good” or “bad” only as judged according to human values. There is no objectively determinable “correct” minimum-wage policy (or trade policy, or environmental policy, or whatever policy) that is independent of human value judgments. A human (or group of humans) can quite sensibly judge a minimum-wage hike to be bad policy even if they all agree that the dollar-value gains (or even the subjective-utility gains) to the ‘winners’ of this minimum-wage hike exceed the dollar-value (or subjective-utility) losses to the ‘losers’ of that minimum-wage hike. The means used to achieve some end can be sensibly judged to be so undesirable that no further discussion of using those particular means is warranted: the means are rejected by value judgments. (See the ten scenarios in this earlier post.)
ThomasH also mistakenly presumes objectivity in the group of people who are the intended beneficiaries of the minimum wage. As I understand his position, the minimum wage is to be judged acceptable if the gains (measured in dollar values?) to the winners in the relevant group exceed the losses suffered by the losers in the relevant group – all with the relevant group being defined as low-skilled or poor workers.
Given his apparent wish to maintain the possibility of minimum-wage policy being socially beneficial, ThomasH correctly seeks to judge the welfare effects of the minimum wage only as these effects affect a select sub-group of society (namely, low-skilled or poor workers). Were we to look at society as a whole, and were we further to measure only that which can be objectively measured – changes in monetary incomes rather than changes in subjective utilities – then minimum-wage legislation would almost always generate net losses. Except in situations in which monopsony power is rampant among employers of low-skilled workers – which plainly is not the situation in modern-day America – raising the minimum wage will cause the value of total economic output to fall. Society at large will be poorer. But it’s possible that the bulk of these losses will be borne by people who our value judgments direct us to ignore or to discount – people such as “the rich” or “employers” or “corporate shareholders.” Assuming that ThomasH’s knowledge of economics is sufficiently robust to inform him that raising the minimum wage will cause the value of total economic output to fall, we must conclude that it is ThomasH’s value judgment that leads him to care only about (or overwhelmingly about) the effects of minimum-wage legislation on the poor or low-skilled.
Such a value judgment is fair (although it’s not one that I share). But who are “the poor”? Who, exactly, comprise this group of people whose incomes can be objectively measured before and after a minimum-wage hike in order to determine if the gains of the hike are greater than or less than the losses? Should the individuals whose incomes are to be so measured be chosen according to their annual monetary earnings? If so, what is the monetary cut-off that defines the relevant group? Is it $30,000 annually (such that we don’t care about – and, hence, in our empirical analysis we ignore – the effects of a minimum-wage hike on the incomes of anyone earning $30,001 or more annually)? Is it $25,000 annually? $50,000 annually? $10,000 annually?
And is the relevant income level that of an individual or of a household? That is, do we classify (for purposes of our scientific, objective cost-benefit assessment) as “poor” all individual workers earning annual pay of less than (say) $25,000? Or do we exclude from our relevant group individual workers who earn less than $25,000 annually but who live in households with minimum annual incomes of (say) at least $75,000? (Is my 18-year-old full-time college student son a poor worker? Should he be classified as one, such that the effects on his earnings of a hike in the minimum wage become part of our empirical calculus? My son’s annual income now is in the low four figures. That’s hardly enough for him to live on alone – but, fortunately for him, he’s a member of an upper-middle-class household whose annual income is quite high.)
What about age? Should we exclude from our relevant group all low-wage-earning workers who are, say, younger than 20 – on the grounds that these workers haven’t yet had time to gain the skills that will eventually enable the vast majority of them to command higher wages? Or should we include such workers in our relevant group? What about low-wage-earning people over 65 years of age who receive the bulk of their annual incomes from investments that they’ve accumulated over the years? Should such workers – many of whom are retired from their careers yet who choose to work at ‘small’ jobs either to supplement their retirement incomes or simply to avoid withering away at home from boredom – be included in our relevant group?
The point is that ThomasH mistakenly presumes that there is some objectively (or objectively enough) defined group of workers or poor people whose collective fortunes or misfortunes we can measure in order to empirically determine if the minimum wage does or does not generate net benefits. Even if we agree that the relevant group does not include everyone in society, there is no such relevant group. Saying “poor people,” “poor workers,” or “low-wage workers” is easy, but these terms do not map objectively onto some well-defined collection of individuals whose incomes can be measured in order to make an objective ‘cost-benefit’ calculation of the minimum-wage’s distributional effects.
I end this post by returning to the first point raised in bold above and posing to ThomasH the following scenario:
Suppose we define the relevant group to be every American other than Bill and Melinda Gates and their children. (Yes, this definition is arbitrary – but so, too, is every conceivable definition other than one that includes all living people.**) Further suppose – not terribly unrealistically – that we can murder all the Gates, liquidate all of their assets, and distribute the value of these assets evenly to all Americans at the paltry cost of $1M. The result will be that each and every American will receive a one-time payment of about $250 dollars. Every one of us in the relevant group – that is, all Americans other than the Gateses – will be made objectively better off to the tune of $250 if this policy of murdering the Gateses and seizing and distributing their assets equally among all other Americans is pursued.
A question for ThomasH: is this a policy that you believe passes moral muster? All – not just some! – members of a sub-group of people are made objectively, measurably better off. Does science counsel us to seriously consider the pursuit of such a policy?
No, of course not, for murder is involved, and no one this side of sociopathy condones murder. So what about stealing all of the Gateses’s wealth – or stealing and then redistributing all but, say, $10M of the Gateses’s wealth? (The resulting gain to each and every American will still be close to $250 – and the Gateses will be left with $10M to live on. This sum is more than enough, for sure, to survive rather comfortably in modern-day America.)
I don’t know for sure if ThomasH would object to such outright theft, but if he’s like most people, he probably would (at least in the blatant manner described here, for it doesn’t appear to be part of a “policy” that is consistent with the rule of law). So what about, instead, a policy – duly enacted by elected politicians – of forcing all the firms in which the Gateses are invested to raise to $10 the hourly wage of all workers currently paid less than $10 per hour? What if it were shown – contrary to economic theory (but play along with me) – that this policy would result in absolutely no losses to any of these low-skilled workers. The entire cost of the enforced wage hike will come out of the pockets of the Gateses. Would ThomasH endorse this policy? I suspect he would – at least if “Gateses” is replaced with “the wealthy” or “businesses.”
My point is a serious one: it is to ask where is the line drawn that separates acceptable confiscations of some people’s property and wealth in order to ‘beneficially’ transfer it to other people? I see no good place to draw such a line. The line I draw is in a very different place: that which separates voluntary, contractual arrangements built on mutual consent from involuntary arrangements built on force – force used, not to enforce contractual agreements and property rights, but to alter such agreements or the property arrangements that are the consequences of such agreements.
In short, the minimum wage
(1) has unseen negative consequence that the vast majority of its champions and fans remain ignorant of – consequences that include job destruction, worsened work conditions, and employment discrimination;
(2) makes the economy as a whole less productive and, hence, poorer;
(3) does indeed produce some winners, but such winnings come invariably at the greater expense of others;
(4) ensures that many of the ‘others’ who pay the cost of the winners’ bounty are themselves poor workers (for example, by being priced out of jobs);
(5) also ensures that many of the winners are precisely the sorts of workers – such as high-wage skilled workers – who most people do not regard as worthy objects of intrusive legislation that imposes costs on poorer workers;
(6) cannot be determined to be good or bad “scientifically” – that is, independently of value judgments.
* I’m a rule-consequentialist and not an act-consequentialist. The best we limited-knowledge humans can do is to assess the likely consequences of various rules as these are followed over time. Our knowledge is simply to meager to allow us to correctly assess the likely consequences of individual actions. Rules that, compared to alternative rules, are likely over time and over many ‘plays’ to produce net benefits are good rules. Rules that by these criteria are likely to produce net losses are bad rules. Moreover, because our limited knowledge prevents us from doing anything better than wisely speculating about what would be the full range of consequences of alternative rules, we must admit the possibility that rules once believed to be good are, in fact, not good (and vice-versa).
** Even “all living people” might be arbitrary. Why exclude from the relevant group those individuals who will be born, say, within the next 20 years?