What's Up, i.e. Down, With the Labor Share?

Even adjusting for measurement bias, its decline is telling us something important. And I don't see it reversing course anytime soon.

A few weeks ago, I noted that labor’s share of national income was doing something a bit unusual and more than a bit undesirable: falling through a period of low unemployment. In most of our business cycles, at least pre-2000, the labor share was cyclical—as unemployment fell and workers had more bargaining power, they could push for higher pay and a bigger slice of the pie.

A number of other recent pieces have looked into this. Greg Ip ties labor share’s decline to people feeling “glum” about the economy, as in they’re contributing to growth but getting a smaller slice. And given the nature of shares, if someone’s getting less, someone else is getting more: “Adjusted for inflation, hourly wages are up 3% since the end of 2019 while profits are up 50%.”

Axios argues that these slice points are “another piece of evidence underscoring the current state of play in the American economy: Surging profits have buoyed stock markets while increasing pressure on an already cranky public.” Ip raises pitchfork concerns: “You can be a red-blooded capitalist and still worry about the political stability of an economy in which ever more output flows toward shareholders instead of employees.”

All of this attention begs for a deeper dive into what we’re talking about re labor or profit shares, and what can and should be done about them. First, we eat our spinach some measurement issues, then a small bit of theory, then a bit of noodling on what to make of it.

The first thing to know is that measuring the labor share is trickier than you might think, in part because a non-trivial share of national income is ambiguous in this regard. Is self-employment income of business owners—called “proprietors’ income” in the national accounts—wages or profits? I can tell you that when I was a self-employed musician, my (meager) earnings felt like earnings, not profits. I didn’t have an employer, but it’s not like I owned a hedge fund or a legal or medical practice. In those cases, distinguishing between wages and profits is harder. However, we can avoid that by focusing on economic sectors that exclude proprietors.

Doing so, however, invokes another problem, well described in this recent post by Owen Zidar and Eric Zwick. They point out that various changes to the tax code over the past few decades made it more profitable for certain businesses to switch their corporate structure from C-corps (think Apple, GM, etc.) to S-corps and partnerships. The main reason for the shift is tax avoidance: these business owners could cut their tax bill by passing their business earnings through to their personal income tax bills vs. the corporate side of the code, thereby reclassifying labor income as profits.

Zidar/Zwick’s figure, pasted in below, shows the sharp drop in the corporate labor share since 2000 (they note in the figure when their first study on the impact of these tax changes came out). They point out that “[m]ost of that missing labor income sits in the same handful of skilled-service industries — professional services, finance, and health care lead the list — where the owners and the workers are often the same person,” and conclude that about a third of labor share’s decline is due to these reclassification dynamics. (BTW, these authors have what sounds like a fascinating new book on the explosion of “pass-through owners” and how they’ve “reshaped the American economy.” I will note here that since my days in Biden tax world, there’s been interest in reducing the tax avoidance introduced by these incentives to pass-through business income. They’re very costly to the budget and such shifting mostly raises after-tax inequality, not economic outcomes.)

Next, a simple equation to keep in mind as we think about the political economy of labor share’s fall:

The change in the labor share = nominal compensation growth - inflation growth - productivity growth

In other words, for labor’s share to fall, and, conversely, for the share of income going to profits (and other stuff, including depreciation, taxes, interest payments, though “capital” or profits dominate) to rise, productivity has to grow faster than inflation-adjusted average pay. All else equal, if inflation goes up relative to comp, meaning real earnings grow more slowly or falls, labor share goes down. If productivity growth beats real earnings growth, same thing.

This is why the slice-of-the-pie framing fits this analysis, and why that break-glass-pitchfork sits atop this post. Economists (correctly) go on all day about the importance of productivity growth, as it is the nation’s key source of income and wealth. But that’s a highly aggregate observation. By tracking labor and profit shares, we get a high-level look into where that growth is being channeled. Is it paychecks (labor) or portfolios (profits)?

“High-level” is an important qualifier. Labor share could easily go up because the highest earners were doing better while low earners fell behind. That’s why you really want to look at the EPI’s work on productivity growth compared to real median pay. Remember, that formula above refers to average compensation, and when inequality is on the rise, the average, unlike the median, gets pulled up by high values. EPI’s figure documents the extent to which mid-wage workers have fallen behind overall growth, a powerful way to understand growing economic inequality.

But the relentless decline in the labor share, even when unemployment is low, is another legitimate way to view a foundational problem with the US economy. And Ip is right to tie this to today’s economy, with AI pumping up equity markets (assets widely held by the wealthy and not at all by the bottom half), at the same time that war-induced inflation is surpassing wage growth.

That’s of course a current-event microcosm of much longer trends, but it’s telling us something about how critically important forces in the American economy—workers’ bargaining power, capital’s political power, the tax code, financial markets—are shaping how the gains from growth are distributed.

My strong sense is that these forces are largely aligned against most working people these days, and that the labor share’s decline, even with the mechanical tax impacts, is indicative of that alignment.¹ Interestingly, Pope Leo XIV agrees: “Income from capital risks replacing income from labor,” he said in his latest encyclical, wherein he worried that labor-displacing AI would sharply exacerbate this trend.

Question #1 is what to do about it. I’ll write up some ideas on that for tomorrow or Thursday. I used to tout full employment as a labor-share booster—refer back to the formula above to see the role of real wage growth—but while that’s clearly necessary, it’s not sufficient. We must take more direct steps, many of which I’ve articulated as part of the affordability agenda, along with my recent minimum wage post, to more fairly distribute productivity’s gains.

Question #2 is whether the folks who are prone to block such solutions—as the status quo serves them well—are comfortable with the pitchfork risk such blocking engenders. Public discontent has a lot of people looking at that case pictured above. We’re not at a break-glass moment yet, but that’s the direction of travel.

1. While the tax impacts documented by Zidar/Zwick bias down the trend, they’re also associated with the upward distribution of incomes. See their figures in the link above (and, I believe—haven’t read it yet—their book) showing where pass-through income falls in the distribution (which is at the top).