If a rising tide is supposed to lift all boats, why have we seen worker productivity rise so much faster than wages?

From the 1940s through much of the 1970s, productivity growth and wages followed a generally similar pattern.

However, since the 1970s, productivity has been increasing faster than compensation, and some use the gap to argue that we need proactive policies to deal with wage stagnation.

It is clearly an important issue, as the key to long-term improvement in living standards is increasing compensation in a non-inflationary manner as a result of gains in productivity. However, many of the simplistic comparisons between productivity and wage growth leave out essential aspects of the issue, thus missing the real underlying challenge.

To back up a step, let’s look at the two key concepts involved: productivity and compensation. Productivity is essentially the amount of output produced per worker. Although the idea is straightforward, complications arise when analyzing it across the economy or looking at changes over time. Productivity varies dramatically by industry (more on that below) and total measures ignore these differences. Also, to look at patterns over time, it is essential to adjust for inflation. The output portion of productivity calculations is usually measured in terms of gross product, and a common practice (and one I use in our forecast publications) is to use real gross product. Real gross product is adjusted for inflation but it’s not a perfect adjustment for this purpose. There are further nuances with the employment portion of the calculation, such as which measure to use and whether and how to adjust for the number of hours worked.

For compensation, many comparisons use a pure wage rate, sometimes converting to an hourly estimate. However, using such a wage rate usually means ignoring benefits such as health insurance and retirement plans, which are a significant portion of total compensation for many workers and have generally increased as a percentage of the total pay package over time. There are also other types of compensation which may be left out such as bonuses, equity awards, profits sharing, and various benefits and perquisites which are common in certain industries.

As I mentioned, productivity for major industry groups varies dramatically. In Texas, total labor productivity (real gross product per wage and salary worker) in 2016 was about $118,500 (according to estimates from my latest forecast). For the services industry group, labor productivity was barely more than half of that ($67,000). Government, construction, and wholesale and retail trade were also less than the overall average. Agriculture was somewhat higher ($142,100), as was transportation, warehousing, and utilities ($151,500). Manufacturing labor productivity was about $226,000, but varied notably between durable manufacturing ($171,700) and nondurable manufacturing ($333,200). Finance, insurance, and real estate came in at $278,100, with $282,800 in the information industry group. Labor productivity in the mining industry group (which is primarily oil and natural gas activity in Texas) is a whopping $661,500.

These variations aren’t surprising when you think about it. A single large oil well can add a substantial amount to output, but it only takes a relatively small number of people and short period of time to do the actual drilling and completion and even fewer to maintain it. Similarly, a handful of employees at a refinery can produce a huge volume of products and high value of output in the nondurable manufacturing category.

Industries vary in the ratio of capital to labor costs, technology utilized, and countless other factors. In addition, there are other layers of differences such as the fact that within each industry, jobs run the spectrum from chief executive officers to newest hires at a minimum wage. Looking at productivity and compensation growth at an economy-wide level ignores these variations.

Researchers at the U.S. Bureau of Labor Statistics recently used information for 183 industries going back to 1975 to look at patterns in productivity and compensation changes at a national level. They found that the average annual percent change in productivity outpaced compensation in 83 percent of 183 sectors studied. However, as they looked at the data by industry, they found notably different patterns. In some cases, the differences were largely explained by how inflation is accounted for. They also looked at the “labor share of income,” which is how much of total revenue is going to labor as opposed to the other components of production such as intermediate purchases, technology, and capital, and found that it declined in 77 percent of the industries studied. In other words, a growing share of income was going to factors of production such as machinery, computer software, intellectual property, or purchases of intermediate goods and services. Interestingly, the 17 percent of industries that saw compensation rise at least as fast as productivity tended to be those with low or even negative productivity growth.

There are any number of valid reasons for the labor share of income to decline. Increased automation is one reason, with technological advances allowing companies to produce more with fewer workers. Such advances are crucial to long-term growth, particularly given the likely slower rates of growth in the workforce which we are likely to see in the future. It is no accident that the decoupling of a simple correlation between wages and productivity went hand-in-glove with the microelectronics revolution and its successors. There are also some not-so-good reasons. High unemployment across the economy or in certain occupations can also suppress wages.

In short, the gap in productivity and wage growth requires a closer look at the reasons for the divergence. In some industries, it is due to good, economically healthy reasons such as the rapid deployment of technological advances. In others, wages are stagnant due to a dwindling need for workers or low productivity growth. Putting it all together, the most important thing that we can do to deal with wage stagnation is to prepare people for the highly-skilled, in-demand jobs of the future. That is the ultimate – and only sustainable – secret to higher standards of living.