The National Employment Law Project has a new report out called “The Good Jobs Deficit,” in which they note that the terrible job market is even worse than people realize. Not only are few jobs being created, but those that are being created are predominantly low-wage jobs, worse than the ones they are replacing. Thus the wages of American workers are stagnant or even falling in some cases.

This isn’t really surprising, as we’ve known about the problem of low-wage job growth for a while. But the report made me think about something else: Tyler Cowen’s recent book, The Great Stagnation. In that book, Cowen took note of the stagnation of incomes for the broad majority, but he interpreted it as a symptom of a deeper problem:

Median income is the single best measure of how much we are producing new ideas that benefit most of the American population. Yet the picture is depressing . . . You can see the rate of growth of per capita median income slows down around 1973, which I take as the end of the era of low-hanging fruit. As an approximation, if median income had continued to grow at its earlier postwar rate, the median family income today would be over $90,000.

Cowen goes on to say that “The American left has pointed out and indeed stressed measures of stagnant median income, but it usually blames politics, insufficient redistribution, or poor educational opportunities rather than considering the idea of a technological plateau.” So for Cowen, the causal story is that technological stagnation leads to stagnating income. He treats the innovation slowdown as basically exogenous, the result of a lack of “low hanging fruit,” easily discovered and exploited technologies that can increase our standard of living. So at the end of the book, Cowen’s recommendation is essentially that we should try to make science a higher-prestige occupation, and then just wait around and accept stagnation until somebody finds some more low-hanging fruit.

But I think Cowen gives insufficient attention to the reverse causal story: one cause of technological stagnation is that labor is too cheap. As Daron Acemoglu explains in this paper, you can use the tools of mainstream economics to construct a model in which the development of labor-saving technology is more rapid when there is scarcity of labor. Economic historians have suggested that one of the reasons that technological progress in the nineteenth century was faster in the United States than in Britain was that labor was scarce in the US.

The reasoning here is pretty straightforward. A rational manager will only adopt labor-saving technology if it is cheaper than the labor it replaces. And when labor is scarce, wages rise as employers compete against each other for workers, making it more attractive to save on labor by using machines instead. For instance, suppose a self checkout machine for a grocery store ends up costing $10 per hour over its lifetime, when you account for purchase and maintenance costs. If your cashiers make $8 per hour, there’s no reason to use the machines. But if they make $12, you have an incentive to replace cashiers with machines, and manufacturers have more incentive to come up with this kind of labor saving technology.

This isn’t great for the cashiers who lose their jobs, obviously. But in the larger scheme of things, working at a supermarket checkout isn’t the kind of fulfilling and valuable work we really want to preserve, and this kind of technological change is necessary if we want to improve our overall standard of living and move in the direction of a post-scarcity society. That’s one of the reasons I argued that preserving and creating jobs shouldn’t be the Left’s main preoccupation. Instead, we need to ease the pain of unemployment for those who are displaced.

But in addition, we need to raise wages. So how do we make labor more expensive? One way is to raise the minimum wage and increase rates of unionization, which are both good ideas. And rising wages in China will hopefullly start to improve the situation on a global scale. But in the United States, the most important thing is to get back to full employment — i.e., create labor scarcity throughout the labor market. Just keep in mind that we don’t necessarily need to do it by creating a ton of jobs.