Behind the Slow Pace of Wage Growth

Despite continued progress in the labor market, wages have been rising slowly. In 2014, total nonfarm payroll employment rose by 3.1 million and the unemployment rate declined by 1.1 percentage points to 5.6 percent, indicating that the labor market was improving. Meanwhile, average hourly earnings and compensation per hour rose only by 1.8 percent and by 2.5 percent, respectively, a smaller increase than one might expect after 5 years of economic recovery. In this article, we look at some factors behind the slow pace of wage growth, including slow productivity growth and labor’s declining share of income.


One reason wages have been rising slowly is that prices have been rising slowly. Low inflation, however, does not explain the trend in wages completely. Even after subtracting the effect of inflation, wages have been rising slowly. In 2014, real average hourly earnings and real compensation per hour rose, respectively, by only 1.2 percent and 1.3 percent.

In fact, real wages have been rising slowly for several years. Measuring from the end of the Great Recession, real wages have barely risen—real compensation per hour has risen only by 0.5 percent, much less than at this point in past recoveries. The lack of strong wage growth has been one factor that has held down the growth of income, consumer spending, and the recovery.

Some temporary factors may explain, in part, weak real wage growth during the recovery. For instance, Daly and Hobijn (2015) suggest that many firms were not able to reduce wages during the Great Recession, so they compensated by not raising wages as fast during the recovery. This factor, however, became less and less important over time as the recovery continued to progress.

Another factor that may have held down wage growth during the recovery is a change in the composition of jobs and hours—a relative increase in  lower-paid jobs and hours may have depressed the average wage. Data, however, suggests that a change in the composition of occupations did not have a strong effect on the average wage: The employment cost index for total compensation—an index that tracks the cost of labor for a fixed composition of occupations—has risen by 11.5 percent during the recovery, which is similar to the growth in average hourly earnings and compensation per hour, which have risen, respectively, by 11.3 percent and by 11.5 percent. Also, Elvery and Vecchio (2015 , Table 2) find that the effect of the change in the mix of occupations on the change in the average wage between 2010 and 2013 was small (and actually positive). Similarly, Mancuso (2015) finds that shifts in industry composition do not explain much of the weakness in wage growth during the recovery.

Some longer-term changes in the economy have likely played a larger role in depressing real wage growth. The first is the slowdown of labor productivity in the last decade. Productivity growth in the nonfarm business sector has averaged only 1.46 percent since 2004 and 0.85 percent since 2010. As the growth of labor productivity is a key determinant of real wage growth in the long run, the slowdown of productivity has probably helped to depress wage growth.

Other long-term changes in the economy, including the evolution of the technology used to produce goods and services, increased globalization and trade openness, and developments in labor market institutions and policies, have caused labor’s share of income to decline at a faster pace since 2000 than in previous years, and in doing so they have likely held down real wage growth. After declining at an average rate of 0.1 percent per year from 1960 to 2000, the labor share has declined more rapidly since 2000, on average about 0.5 percent per year (see Jacobson and Occhino, 2012). In an accounting sense, the faster decline since 2000 has subtracted about 0.4 percentage points per year from average real wage growth relative to the period before 2000.

Going forward, wage growth will likely pick up in the short run, as inflation rises and labor market conditions strengthen further. In the longer run, whether average real wage growth remains lower than in the past will depend on whether trend productivity growth continues to be low and whether other fundamental economic forces cause further declines in the labor share of income.



Filippo Occhino

Senior Research Economist

Filippo’s primary areas of interest are monetary economics and macroeconomics. His recent research has focused on the interaction between the risk of default in the corporate sector and the business cycle.

See AllFilippo Occhino

Timothy Stehulak

Research Analyst

Timothy Stehulak’s primary interests include applied microeconomics, economics of education, labor economics, and regional economics.



What's been said:

Discussions found on the web:
  1. orsogrigio commented on Jul 16

    Interesting article. Only there is a passing remark “Other long-term changes in the economy, including the evolution of the technology used to produce goods “, put there just for sake of precision (or at least I have felt in this way) I deem is the ONE and thrue origin of the situation. Consider the last graph. It simply states that the weight of cost of labor went from 67 % to 57 % of the worth of production in about 70 years, but look what happened from 1997 … 64 to 57 in 17 years. Let’s see in a different way :-0.08%/year average in 50 years (1947-978), but -0.412%/year average last 17, five plus times steeper. Reading that graph you can only come to the conclusion that work compensation environment of 1947-97 is completely unlike the 1997-2015 one. What’s the change ? computers (and thus robots). If we do not account for this dramatic change, I think we go on walking blindfolded in a forest at night. 15 years ago smelting a ton of steel required about twice as much power as today (or even less): smelting iron & computers ? sure, improve ladle project, know better thermodynamics involved, improve building design, power networking, temperature control … but all these thing can be done ONLY via computer simulation, because with CAD is possible to study configurations at a speed simply unknown before, and so optimize the final one. You reduce the number of draftsmen, the time and cost of project and so on along all the line leading to the ladle because the same holds true for ANY subsistem involved in process. This is a dramatic change. In my work life I have started with A0 size sheets and china ink and moved to CAD and system simulations, I have tasted this on my own skin. I do not know how economist go on placing in their studies data coming from non congruent statistical universes, and never look at the empiric facts under their nose: you can’t study airlines sistems putting together the 4 engines DC4 likes, DC8s, 747s and 777s performances in a cauldron & figure anything out. Work, I mean man hours needed, will drop, output, I mean products ready for sale, skyrocket and in a SHORT time. We are facing a far more violent Industrial Revolution, requiring a huge change in society. Actually I haven’t seen much about …

  2. Denis Drew commented on Jul 16

    And what to think of American working hours (deliberately) becoming steadily less productive. I just read in Tom Geoghegan’s book Only One Thing Can Cave Us (around p. 28) that American manufacturers, in particular auto manufacturing, are breaking down jobs into simpler and simpler component parts — in contrast to German firms who, due to the influence of co-determination workers councils, steadily upgrade worker input and innovation.

    The whole world besides the US is straining to learn (or steal) ever more productive techniques to reach maximum possible output some day. In the US which has only a fixed amount of workers at any one time is trying to make those workers less and less individually productive in the mindless pursuit of paying them less and less — leading to ever lowering national productivity. Talk about the fallacy of composition. We are going to be waving to them from the down escalator while they are riding by on the up.

    From Eduardo Portor in the Times: Sizing Up Hillary Clinton’s Plans to Help the Middle Class
    JULY 14, 2015
    “Not only does the American economy suffer from one of the least skilled work forces … *

    “Jobs in the economy’s largest, fastest growing occupations, in retail sales, food preparation and the like, are awful, she said, because ‘companies have created strategies that use people as interchangeable parts.’ ”

    • willid3 commented on Jul 16

      course US automakers also started hiring and paying workers the same wages that some retailers and fast foods restaurants pay. so having to ‘dumb’ down the jobs would seem to have been required.

      course one can look back at GM back in the 60s when it had more than 300,000 workers making a few million cars, to today, where it has less than 100,000 workers and makes about 10 million vehicles.
      CAD has been around for a long time now (since the 60s). and some of the computerization of manufacturing since the 80s.and we can guess that it has accelerated since then (even with US leading the world in manufacturing, we employ fewer workers in manufacturing than we did in the 60s. and that was a smaller country)

      and automation will only accelerate from where it is now. the problem this time around (from the industrial revolution ) is that we dont know what jobs will remain going forward. and without them, is hard to see how businesses will stay in business as if they think they have trouble finding customers (or keeping them) now they havent seen nothing yet

    • orsogrigio commented on Jul 16

      Denis, I simply do not believe that a manufacturer behaves like Geoghegan described, that is one will NEVER reduce production efficiency in order to pay less manpower. A company would, on the contrary push automation up and utterly cut jobs, but NEVER make technically dumb choices to pay less people. Man-hours are a cost, like robot maintenance or gardening at high quarters, and once the economic equilibrium of a line is set they go on. Tooling is enormously more expensive than man hours, so nobody will be so stupid as “breaking down jobs” for saving pennies on workers. The middle paragraph is pure contradiction. Incidentally NO Country has unlimited manpower, it takes 9 months plus about 18 years to make a worker even in China. The last sentence, on the contrary is true, and shows that excess manpower from more sophisticated echelons are drained down to food etc, and THERE the Geoghegan method works, since tooling is minimal and workers plentiful : practically even if you have a PhD in Ancient Oriental Literature and you are pennyless you can carve roasts, but you can’t sit at a CAD station.

  3. reedsch commented on Jul 16

    Even with technological advancements, if it takes fewer workers to produce the same or greater number of goods and services, one would expect the same or even greater pool of productive gains (i.e. profit) to be divided among fewer workers, so no way would this lead to a break-even much less a reduction in worker wages. Unless that is the net plus due to technology is not returned to the pool of gains, in other words is directed primarily (trying to avoid loaded words like diverted or appropriated) to the owners of the technology. It remains a distribution issue, not a production issue.

    But would someone on God’s green earth explain to me why in the age of cheap capital that the owners of capital still get the first and biggest bite of the apple?

  4. reedsch commented on Jul 16

    Is there anyone who still thinks that the economic model based on exchanging human labor for wages is not broken?

  5. reedsch commented on Jul 16

    And I am praying that someone here on God’s green earth can explain to me why, in the age of cheap and abundant capital, why capital still gets the first and biggest bite of the apple?

    • orsogrigio commented on Jul 17

      @reedsh : simple, because what you wrote is a plain statement, not reasoning. From an operative standpoint one can’t put together any comment on a statement, it’s like you said ‘I love apples’, just to be in the flavor.

Posted Under