You may already be aware of a number of different ways of looking at the pay-productivity gap, which represents the difference between how much productivity has grown in the United States vs. how much worker pay has grown. This gap concept recently got some social media attention when Bloomberg said "no one has good answers" to why this gap exists, and some enterprising Twitter users brought up Marx as the obvious counter to this. (Here's the Bloomberg article linked by the tweet, titled, "Workers Get Nothing When They Produce More? Wrong: But the trickier question is why they don't get the full benefit of rising output.")
I realized this was a highly simplified debate in need of some nuance. So I decided to to some informal polling and then some research: pay attention as you read to what your emotional response is to each bit of information. What do you want to be true? What do you reject out of hand?
Together we're going to learn what we can about the productivity pay gap.
I asked a few friends what they thought might be driving the gap, and here's what I got, in loose descending order of popularity:
- Rise of automation (higher automation means workers are less in demand for the work, and therefore less valuable)
- Decline of unions (loss of bargaining power)
- Corporate greed and/or fundamental exploitative nature of capitalism
- Globalization (ease of importation means US workers have to compete with lower-priced foreign workers)
- Reaganomics / neoliberalism (changes to US regulatory and tax structure during the 1980s favored corporate wealth)
- The US left the gold standard (meaning wages got eaten up by inflation)
"X Marks the Spot:" on Methodology
One really, really interesting article that highlights a way to approach this question is "X Marks the Spot Where Inequality Took Root: Dig Here." Author Stan Sorscher shows a version of the graph that shows productivity vs. wages of certain goods-producing workers. While the data choices are questionable (for reasons we'll see later) and the second half of the article is not particularly methodologically strict, he does demonstrate an interesting point: there was a change in the 1970s that seems fairly abrupt.
Do Any of These Hypotheses Hold Up?
Interestingly if we're just looking for correlations, here's what we see versus a few hypotheses:
- Automation: The National Bureau of Economic Research concluded that before 1990, automation had a minimal impact on wages, so the correlation is very weak
- Unions: moderate correlation, however non-union workers tend to make 80%+ of what union workers make, and even in 1970 they were only 30% of the workforce
- Corporate Greed: Possibly an argument can be made that before the 1970s corporations / capitalists were not particularly greedy, and afterward they were much more greedy as a whole, but it is difficult to measure
- Globalization: It's tough to draw a strong correlation because "globalization" is tough to measure with a single metric. IMF economists Matthew Slaughter and Phillip Swagel claim, "nearly all research finds only a modest effect of international trade on wages and income inequality." The Economic Policy Institute claims that changes to US trade policy ("unfair trade deals") from the 1990s and early 2000s hurt US worker wages, but this happened well after the 1990s. (Interestingly it seems to be taking Trump's side on trade, and was written in 2015.) The Council on Foreign Relations also claims that trade's impact on US wages picked up in the 1990s. However imports can also affect productivity--if you buy something made in Canada then it wasn't produced in the US--so its impact on the productivity-wage gap may be smaller than its raw impact on wages.
- Reaganomics: Reagan took over as President in 1981. While his economic policies may have impacted the graph in question, the divergence began in the 1970's
- Leaving the Gold Standard: High correlation, as the US left the Bretton-Woods system starting in 1971, causing the "Nixon Shock." But did it cause a substantial jump in inflation? If we use the Consumer Price Index--the way you usually see inflation measured--then we see a substantial spike of inflation in the 1970s from where it was in the 50's and 60's, but from the 80's onward inflation drops to and below 1950's/60's levels.
If inflation was the primary driver for the productivity-pay gap, we'd need to see an inflation rate that remained much higher than the 1950's/1960's period when productivity and pay seemed to track better.
At first glance, none of these answers seem to be a silver bullet. And, as usual, the picture is a lot more complicated than it seems at first glance.
We'll dive into how these two numbers--productivity and pay--are measured, and how they're adjusted for inflation, to learn that it's even more complicated to figure this out than we thought. Plus, we'll look into some of the leading explanations for the gap.
Here's an interesting graph from Heritage to whet your appetite.