Does a higher minimum wage cause unemployment?

by Jehu

SPOILER ALERT: The answer may be, “Yes”

The great thing about the debate over whether the minimum wage reduces employment is how easy it is to dismiss it.

First off, the minimum wage is not, and was never meant to be a jobs program; instead, it is designed to protect the subsistence of the lowest paid strata of employed workers and place a floor under wages generally. In some cases, however, the content of the debate suggests the two sides expect the minimum wage to do something it was never designed for: add to or at least not prevent an increase in aggregate employment.

The second problem with the debate, is that it is obvious businesses are not primarily concerned with wages of their workers, but with their own profits. A business will gladly pay higher wages if it can make a bigger profit doing so. This suggests there is, at best, only an indirect relationship between wages and employment.

In theory, higher wages might reduce aggregate employment only if they can be shown to reduce the profits of the capitalists. This is not always the case.

For example, at the top of an expansion the mass of wages and the mass of profits are both at their peak; while both wages and profits fall to their lowest level together during a depression. A higher minimum wage may, therefore, be correlated with higher profits, lower profits or leave profit completely unchanged.

Wages, profit and labor time

Nevertheless, a higher minimum wage must be inversely correlated with employment, because, at the most basic level, higher wages are clearly inversely correlated with profits. At some level an inverse correlation between wages and profits has to be true because the aggregate labor day itself is divided into paid and unpaid labor time. If the duration of the labor day is fixed, wages correlate with the portion of the labor day for which the worker is paid, while profit correlates with the unpaid portion of the labor day. Assuming the aggregate duration of the labor day is fixed, when the paid portion of the labor day increases, the unpaid portion must decrease.

However, as we already know, the labor day is not fixed. Both the paid and unpaid portions of the labor day can increase or decrease together, and they can increase or decrease at different rates, or one can increase while the other decreases.

Thus, what appears at first to be a fairly simple relation between wages and profits appears now to be a highly complex one. If wages rise, profits might also rise or they might fall or remain unchanged. In theory at least, the change in the state of one aspect of the relation in any direction has no necessary impact on the state of the other.

Since employment is a function of profit not wages, we can’t know what effect any increase in wages will be unless we know what impact the given increase in wages has on profits. But here the relationship is also highly complex: profits may increase while employment increases, remains unchanged or falls.

For instance, a capitalist introducing more advanced means of production may increase his profits even as he reduces his labor force. He may increase his profits by penetrating into new markets and, thus, may either increase his labor force or leave it unchanged. He may work his existing labor force for a longer duration of time and thus increase his profits without increasing his labor force at all.

From the above, it can be seen that trying to correlate a simple straightforward change in the minimum wage to a change in aggregate wage employment is nowhere near as simple as it is made out to be.

Does the relation between wages and profit work both ways?

Let’s add another wrinkle to this problem: In general a relation between two aspects of a relation may be commutative or non-commutative. A relation between any two factors is commutative if a change in the state of either aspect necessarily results in a definite predictable change in the other aspect of the relation. For the relation between the minimum wage and aggregate employment to be commutative, the change need not be in the same direction; e.g., higher wages may lead to lower employment and vice-versa. It simply means whatever the change we specify has to operate no matter which aspect of the relation is taken as the independent variable. If the relation between wages and aggregate employment is commutative, an increase in wages will lead to a decrease in employment; but a decrease in employment would also lead to an increase in wages.

If higher wages leads to lower employment, the principle of commutation implies lower employment leads to higher wages. If this principle does not hold, then we can say the relation is non-commutative. Higher wages may decrease employment, but lowering wages will not boost employment; or, higher wages may reduce employment but reduced employment (in the form, for instance, of reduced hours of labor) may not boost wages.

Interesting enough, in addition to oversimplifying the complex relation between the minimum wage and aggregate wage employment, people seldom ask if a lower minimum wage will create more jobs. Assuming a higher minimum wage decreases aggregate employment, does it follow a lower minimum wage will increase employment? Moreover, if a higher minimum wage reduces aggregate employment, does reduced aggregate employment (reduced hours of labor) increase wages?

If the relation is non-commutative, a positive or negative change in wages does not have the same predictable inverse impact on employment. In the second example a change in employment does not have a predictable impact on wages as a change in wages has on employment.

A one-sided debate?

The question the impact of a change in the minimum wage has on aggregate employment tends to be framed almost exclusively in terms of the impact higher wages will have on employment and generally avoids discussing the Impact lower wages have on aggregate employment or the impact fewer hours of labor might have on wages. Why is this?

The literature on the relationship between wages and employment seems to offer no opinion whether the relationship between wages and employment, if such a relation exists at all, is commutative or non-commutative. This might suggest either there is no real relationship between wages and aggregate employment that can be determined by the empirical evidence, or the participants to the debate are deliberately trying to avoid discussing it.

Indeed evidence that wages and employment are not correlated is offered by a paper done for National Employment Law Project: “Raise Wages, Kill Jobs?” In that paper the authors argue the data shows no such correlation:

“[Basic] economic indicators show no correlation between federal minimum-wage increases and lower employment levels, even in the industries that are most impacted by higher minimum wages.”

However, in an article for Forbes, William Dunklberg challenges the validity of the conclusion a higher minimum wages has no impact or even a positive impact on employment:

“[A higher minimum wage], it is alleged, will provide more income to support spending and stimulate the economy.  If it works that well, why not make the minimum $50?  This would provide someone working 2,000 hours a year an income of $100,000, eliminating poverty and stimulating the economy.  Obviously, $50/hour would be detrimental to employment as is $7/hour, it’s just a matter of degree.

Not high enough?

Dunklberg here inadvertently raises another equally important question: If a higher minimum wage has no or only a positive impact on employment, why can’t the minimum wage just be raised to any level? In other words, is the impact of higher minimum wage discontinuous?

Discontinuous change in the state of a relation is well known in scientific literature. To give a good example, we can gradually cool water from 100 degrees to 33 degrees without producing any change in the state of the liquid. It would appear that the temperature of water has no impact on its state as a liquid. But reduce water from 33 degrees to 32 degrees and there is a profound alteration in its state.

The difference of only a single degree determines whether water appears as a liquid or a solid.

It is equally possible a change in the minimum wage level also may have a discontinuous impact on aggregate employment. Aggregate employment may be subject to a sudden alteration only if the change in the minimum wage is large enough? If the existing empirical evidence shows no correlation between a change in the minimum wage and the level of aggregate employment this just may be because no increase in the minimum wage level has ever been of sufficient magnitude to produce a significant effect on aggregate employment.

The grounds for believing this latter interpretation of the empirical evidence might be true is suggested by the fact that, according to official government statistics, the minimum wage today only has about half the real purchasing power of the minimum wage in the 1960s.

(NOTE: In gold terms the collapse of purchasing power has been even more dramatic.)

While the minimum wage is certainly higher in nominal currency terms than it was in the 1960s, the real purchasing power of the minimum wage actually has been declining for almost 50 years. This raises an extremely relevant question about the current discussion because at the same time the real purchasing power of the minimum wage has been declining in real terms, aggregate employment has been increasing.

Has a declining real minimum wage increased employment?

It is just possible aggregate employment has been increasing for the last 50 years in large part because the real minimum wage (the nominal minimum wage adjusted for inflation) has been declining?

Surprisingly, not one of the economists I have read made the effort to nail that question down. In the empirical data, a serious case can be made for the proposition the real, purchasing power adjusted, minimum wage has been historically correlated with the level of aggregate employment, but we have just been looking in the wrong place for evidence of this.

To put this in terms that are easily understood: It is possible higher real wages do reduce employment as some economists insist. And it is possible the evidence for this has been staring at us in the empirical data for the last fifty years in its inverse form. A lower real (inflation adjusted) minimum wage can be used by the state to artificially increase aggregate employment.

Not surprising, this is idea was first suggested by the Marxist economist, Henryk Grossman, almost 80 years ago.

If the state can continuously lower wages by steadily reducing the purchasing power of nominal wages, it can force more people into the labor force to work more hours in a vain attempt to escape poverty.

The evidence for the impact the minimum wage has on aggregate employment may be concealed beneath empirical data that does not take into account the declining purchasing power of the nominal minimum wage rate. Moreover, since the data employs unadjusted currency wages, it appears statistically as if both wages and employment have been rising for 50 years — suggesting there is no correlation between wages and employment.

The evidence for a relation between the minimum wage may be buried beneath data that relies on unadjusted nominal values.