Raising the minimum wage is not a good way to help the poor

The Democrats are proposing to raise the federal minimum wage to $15 per hour. The current federal minimum wage is $7.25 per hour. Although most states set their minimum wages higher than this, a large number do not, so the jump the Democrats are proposing is quite considerable. In this article, I want to argue that raising the minimum wage is not a good way to help the poor. Moreover, I want to argue that this is true even if the minimum wage doesn’t have large disemployment effects.

Traditional economy theory says that raising the minimum wage will cause unemployment, except under certain circumstances (which I will get to). The reasoning is as follows. A firm and a worker enter into a labour contract, whereby the latter will work for the former, when each judges that it is in his interests to do so. It will be in the firm’s interest to hire the worker when the wage demanded by the worker is less than the benefit the firm expects to get in terms of extra productivity. Hence some economists claim that workers are paid their “marginal product”.

However, if the worker demands a wage that is higher than the productive benefit the firm expects to get, it will not be in the firm’s interest to hire him. A teenager cannot walk into McDonald’s, bang his fist on the counter and ask for $100 per hour, since he is unlikely to increase the firm’s productivity by anything close to that amount. Of course, more-skilled workers can demand higher wages than less-skilled workers, which is why investment bankers and city lawyers get paid more than the minimum wage.

If a worker has very few skills, the firm cannot expect to get a large benefit from hiring him. For example, if McDonald’s needs more cooks, but the prospective worker has never flipped a burger before, the firm will only hire him if he is willing to work for a low wage. This wage may happen to be lower than what other people in society consider fair. Suppose the government decides to pass a minimum wage law. If the minimum wage is greater than the wage that was agreed between McDonald’s and the unskilled worker, it will no longer be in the firm’s interest to employ him. He may therefore get fired.

According to traditional economic theory, the minimum wage is effectively a “price control” on labour. It says to firms: “you must not purchase labour at a price less than the minimum wage; if you do, you will be in violation of the law”. Likewise, it says to workers: “you must not sell your labour at a price less than the minimum wage; if you do, you will be in violation of the law.” As a consequence, any worker whose marginal product is less than the minimum wage will not be hired, leading to unemployment.

However, there are certain circumstances where the above reasoning does not apply. These circumstances are referred to as “monopsony”. Recall that “monopoly” is a situation in which there are many buyers but only only seller (e.g., there is a town with only one company providing water). “Monopsony” is the same, but instead of there being many buyers and only one seller, there are many sellers and only one buyer (of labour).

For example, suppose there is a town with only one factory, but many people who want to be machine operators. The factory owner has much more bargaining power than the workers because if he offers them a low wage, they cannot threaten to go and work in another factory. He may therefore offer to pay them less than the “market wage” – the wage that he would have offered in a competitive market (such as a town with several factories). Since the workers do not have many options, some of them may accept this low wage; others may decide not to work at all.

In these circumstances, a minimum wage can increase workers’ wages without having any disemployment effects. The reason is that, so long as the minimum wage is not raised above the market wage, it will still be in the factory owner’s interests to hire the workers. He will be forced to pay them the market wage, and most or all of the workers will accept that wage (including those who weren’t willing to work at the low wage he was originally offering). Under monopsony, the minimum wage can actually create employment.

Monopsony is an example of “market failure”. Since there is only one firm buying workers’ labour, the market for labour cannot be considered “competitive”. Hence there is a case for government intervention to correct the market failure in the form of a minimum wage.

Few economists dispute that the minimum wage can have disemployment effects. As a consequence, almost no economist advocates a very high minimum wage (e.g., $50/hour) even though it would be nice if everyone got paid that much. Where experts disagree is the extent to which labour markets are competitive versus monopsonistic. Experts who believe they are generally competitive tend to oppose the minimum wage, whereas experts who believe they are somewhat monopsonistic tend to support the minimum wage.

Of course, labour markets don’t have to be completely monopsonistic (as in the factory example above) for the minimum wage to be a good policy. If the labour market is dominated by a few large firms that employ different kinds of workers (e.g., factories and restaurant-chains), it may be costly for workers to switch between firms. They will therefore have less bargaining power, since the threat of going to work for another firm is less credible. In addition, the minimum wage raises the payoff to finding a job, which means that workers may put more effort into getting one.

There is still disagreement among economists as to exactly what impact the minimum wage has on employment. Some economists, such as Arindrajit Dube, argue that the minimum wage has little or no disemployment effect in countries like Britain and the US. Others, such as David Neumark, argue that most studies do show disemployment effects, especially for young and less-educated workers. However, regardless of whether you side more with Dube or more with Neumark, evidence for disemployment effects is certainly not overwhelming.

One highly cited paper was even titled ‘Why Does the Minimum Wage Have No Discernible Effect on Employment?’ And as Noah Smith points out in an article supporting the Democrats’ proposal, surveys of economists suggest that expert opinion has gradually shifted in favour of the minimum wage, which he attributes to the lack of evidence concerning its supposed disemployment effects. On the other hand, libertarian economists such as Bryan Caplan and Tyler Cowen have contested this new consensus, arguing that it is too myopic.

In the remainder of this article, I want to argue that raising the minimum wage is a bad way to help the poor even if it doesn’t have large disemployment effects. My argument is based on the fact that the absence of disemployment effects does not imply the labour market is monopsonistic. The reason why is that firms can respond to minimum wages in many ways other than by firing or not hiring workers. These are called “non-employment margins”.

In an excellent new paper, the economist Jeffrey Clemens discusses various non-employment margins, as well as the empirical literature surrounding them. As he notes in the conclusion: “Estimates of how a higher minimum wage affects wages and employment are not, as is commonly presumed, sufficient to assess its desirability. It is unfortunate that standard textbook analyses convey an erroneous intuition on this important point.” So, what non-employment margins are there?

First, firms can simply evade the minimum wage. This is particularly true in areas where there are large numbers of illegal workers, who – in virtue of their status – have very little bargaining power. If an illegal worker tried to sue his employer for paying less than the minimum wage, he could risk being deported. As a consequence, he is very unlikely to do so.

Second, firms can pass the costs of the minimum wage onto consumers in the form of higher prices. One might assume that this just represents a transfer from relatively wealthy consumers to relatively poor workers. But it’s not quite that simple. Clemens notes that the customers of firms that employ large numbers of minimum-wage workers are often poorer than average. (Your typical Walmart customer isn’t a CEO.) He reviews evidence suggesting that “pass-through effects”, as they are known, can make the minimum wage much less progressive than it would first appear.

Third, firms can alter workers’ non-cash compensation and other job attributes. In most jobs, workers don’t just go home at the end of the day with an envelope of cash, having done a given amount of work. They receive a uniform, meals, discounts and other fringe benefits. In the US, the most important fringe benefit that workers receive is employer-provided health insurance, which often makes up a non-trivial portion of total compensation. All these forms of non-cash compensations are things that firms can tweak in response to a higher minimum wage.

For example, if the minimum wage goes up, firms may force workers to pay for their meals and their own uniform. Alternatively, they may offer workers a cheaper and less comprehensive health plan. In addition to changing the worker’s non-cash compensation, they can alter his “non-compensation amenities”. For example, they can require him to work harder or give him less input into his schedule. What firms care about is the total cost of employing an additional worker, not just the wage they have to pay him (even though that may be easier for economists to measure).

Fourth, firms can alter their mix of labour and capital. When the minimum wage goes up, it may not be in a firm’s interests to immediately fire a large number of workers. Going forward, however, the firm may choose to rely more on capital and less on labour. For example, it may decide to invest in some form of labour-saving technology, such as self-checkouts. As Clemens notes:

Consider a fast-food chain, for example, which must choose between automated ordering kiosks and employee-operated cash registers. Each choice involves an initial investment that will depreciate gradually. When wages are low, labor-intensive cash registers might minimize average production costs. When minimum wages rise, new firms might minimize average costs by acquiring automated kiosks. New firm entrants to the fast-food market may thus adopt less labor-intensive technologies soon after a minimum wage increase.

If firms respond to minimum-wage increases by changing their mix of labour and capital, there will be disemployment effects, but they will occur gradually as the new technology is phased-in (over, say, a number years). Hence it may be difficult to detect these effects in empirical studies, which typically look for discontinuities within a relatively narrow window of time around a policy change.

There are more examples, and lots more details, in Clemens’s paper itself – which I strongly recommend reading. The implication, it seems to me, is that trying to raise people’s wages through government fiat does have unintended consequences, and unless there are good reasons to believe a labour market is monopsonistic, you should avoid interfering in wage negotiations. Since the federal minimum wage applies to all labour markets in the country, there is a strong case against raising it to $15 per hour (especially given the cost-of-living differences across states).

Small businesses have had a tough year. After being shuttered for months during the pandemic, many were trashed or looted during the summer’s riots. Raising the minimum wage seems like the last thing you’d want to do under these conditions. (Note that Amazon, which already pays its employees $15/hour, has taken out full-page newspaper ads calling for an increase in the minimum wage. One suspects this less to do with helping workers than with driving its smaller competitors out of business.)

Of course, workers have had a tough year too, with record levels of unemployment during the pandemic. However, there are ways to help workers that don’t place an undue burden on small businesses (e.g., expanding tax credits or doling out stimulus checks). Let’s hope Biden and the Democrats see sense, either by abandoning their proposal altogether or by compromising on a less dramatic increase up to, say, $10 or $12 per hour.

Image: Gustave Courbet, The Stonebreakers, 1849

---

Thanks for reading. If you found this newsletter useful, please share it with your friends. And please consider subscribing if you haven’t done so already.