Minimum wages and capital/labor substitution

August 17th, 2015 at 3:43 pm

Two arguments have collided in recent months.

#1: the pace at which robots are placing workers is accelerating;
#2: the historically large minimum wages increases being proposed and implemented in various places will have the unintended consequence of hurting the job prospects of those the increase is meant to help.

Put these together, and you get: increases in the minimum wage, especially large ones, will cause employers to invest in labor-saving technology to avoid the increase in labor costs. This capital/labor substitution will be that much more likely as the pace at which labor-saving technology is entering the workforce is accelerating.

It’s a perfectly logical argument, and Lydia DePillis does a nice job exploring its nooks and crannies in this WaPo article. Focusing on the fast-food industry, she writes:

About 30 percent of the restaurant industry’s costs come from salaries, so burger-flipping robots — or at least super-fast ovens that expedite the process — become that much more cost-competitive if the current federal minimum wage of $7.25 an hour is doubled.

OK, let’s get this out of the way: I salute both cap/labor substitution as a real thing that happens, and that more expensive workers, all else equal, incentivize tapping that shift.

But I don’t find this a particularly compelling argument against minimum wage increases, or any other wage increases, for that matter.

First, one implication of this argument is that we should make sure to keep wages low enough so employers won’t want to bother swapping out workers for machines. That’s a great way to whack productivity growth, not to mention workers’ living standards. I think of it as the let’s-pretend-the-US-is-an-emerging-economy strategy to job preservation. “If we could only pay Mexican-level wages, no employer would ever have a reason to automate!”

Second, argument #1 above is far from substantiated. One look at our quite disturbingly underwhelming productivity-growth statistics should make you think twice about the assertion that the robots are coming ever faster for our jobs.

On the other hand, who are you going to believe, me or your lying eyes? That is, I’ve seen, and DePillis documents, the increase in tablet-based ordering or, in a different part of the food industry, auto-checkouts in supermarkets. But remember, point #1 decidedly does not argue that there’s no cap/labor substitution. It simply says there’s no evidence that it’s speeding up.

Finally, when it comes to unintended consequences and the minimum wage, if you just go with the theory—or with the anecdotes—you’re going to get it wrong in a big way. You’ve got to look at the evidence of job losses associated with minimum wage increases, and here, recent work by minimum wage scholars (Dube, Lester, Reich) is highly instructive.

Tapping the extensive variation of minimum wages across different localities, they ask: when county X has a higher minimum wage than bordering county Y, what happens to restaurant workers in X compared to those in Y? The nice thing about this research is that by comparing employment outcomes across contiguous borders, you’re controlling for economic factors that influence job growth on both sides of the border, and thus better isolating the differential impact of the higher minimum wage.

Dube explains their results:

Comparing across these neighboring counties, we showed that there was no evidence of job losses for high impact sectors such as restaurants and retail. This was true even considering four or more years after the minimum wage hike. In follow up work, we used the same cross-border methodology to study the effect on teens—a high impact demographic group…Again, we found no discernible impact on employment.

So, based on both the evidence and the idea that it is mistake for economies to avoid pay increases in the interest of dis-incentivizing capital investment, I don’t think we should worry too much about the impact of moderate minimum wage increases on labor substitution. Or, perhaps a better way to put it: we should worry a lot about the impact of falling real minimum wages on low-wage workers, many of whom are adults, minorities, parents with kids, and so on.

Large minimum wage increases, like going to $15 in places with low wages, are another story. There, we must consider the local wage and price levels and the length of the phase-in. While we shouldn’t focus on keeping wages low enough to avoid the use of productive technology, we obviously should be mindful of potential tradeoffs when we’re going “out-of-sample” as regards the magnitude of the wage increase.

Print Friendly, PDF & Email

7 comments in reply to "Minimum wages and capital/labor substitution"

  1. mitakeet says:

    Automation is an interesting subject to me because on the surface it seems like a no-brainer. Get rid of the lazy humans who argue with spouses, work drunk and call in sick with perfect robots that never make a mistake. However, having worked at a plant where they did that for one aspect of the process, the robots cost money even if they are turned off (opportunity costs on capital), require very expensive maintenance and are not very flexible if something changes/goes wrong. The company where I worked invested a lot of time and money into their robots, then their sales took a dip and the robots wound up sitting idle 90% of the time. Also, as the normal wear and tear on the support equipment occurred it needed someone who could fix those things. Yes, there was a net decrease in labor costs, but the interesting thing is when the volume goes up or down people can be brought in or have hours reduced (in the short run there are lots of people happy to have an extra day off here and there) and because the work is pretty easy it is easy to train people. Thus I see automation as strongest in places like Japan and Germany where it (literally in some cases) requires a legal act to let people go. As long as businesses can dynamically add/reduce labor I don’t see a lot of automation taking over. In niche locations, like your mentioned tablets to order, I see that more of a matter of dealing with irregular customer interaction (meaning you can’t hire for your peak load else you have excess labor the rest of the time).

    • Jared Bernstein says:

      At least they didn’t try to form a union!

    • spencer says:

      When you replace labor with machinery you are making more and more of your cost to be fixed cost and reducing your variable cost. Labor is generally firms largest variable cost and if they can not reduce that when demand drops they are going to be in financial trouble.

  2. Scott Monje says:

    Looking at the issue from another perspective, if a lot of investment is going into robots and other automation, why is that happening at a time of low labor-force participation and of low and stagnating wages? Shouldn’t there be a link between capital and labor costs in that direction, too?

  3. spencer says:

    But I for one are all for firms employing all the labor saving devises they can

    sorry about that — editing your own writing is the hardest thing in the world.

  4. spencer says:

    In general we have been moving in the direction of reducing investments because labor has become the cheap factor of production. If you look at net real nonresidential capital stock– — you will see that is in a secular long term downtrend. That is the dominant factor behind why productivity has been weak and is likely to remain so. For so many of the conservative economists cheap labor is their answer to everything. But cheap labor is why we have poor productivity and economic stagnation. They have taken the position and are winning the argument that it is a bad thing if higher minimum wages lead to restaurants employing more labor saving devises. But I for one am all for firms employing all the labor savings devises they can because that is what drives our increases in standards of living.

  5. Amateur says:

    The shift in manufacturing to low-wage nations has always been treated as a ‘natural progression’. Many economists look at the past trends and think they’re saying something profound by declaring manufacturing the stuff of developing economies.

    If there are economists in the US that are actually concerned about stagnation, they should be encouraging higher wages at the low end, and they should be happy that this might lead to a further push for mechanization and automation. That could turn ‘burger flippers’ into something else. Wouldn’t it be nice if minimum wage workers had more interesting jobs, perhaps with more skill or more interpersonal time? And if they could afford to eat better to boot?

    It is instructive to look at the US vs. Germany and figure out why the US lost its manufacturing edge but Germany did not, despite similar wage levels. They just weren’t so fast to take the easy route, which was just moving existing jobs overseas and sitting around reaping the profits. No, they actually continued to improve.

    So if America has become lazy, it is because the investors and their CEOs have become lazy. They don’t do anything. And because of that, the rest of the workforce suffers.