Though last Friday’s jobs report came in a bit under expectations, many analysts, myself included, noted that (nominal) wage growth has picked up in recent months. That is both expected and precisely as it should be. As the job market tightens, employers should have to bid wage offers up more to get and keep the workers they need.
However, given the weakness in the bargaining power of so many in the workforce, along with some anxiety about price pressures from wage-push inflation, such evidence must be scrutinized. How fast are wages growing? Which wages or compensation series are we talking about? Does wage growth threaten inflationary pressures? Once you adjust for inflation, are workers getting ahead? If so, which workers? The average? The median? Low-wage earners? (Another good question, which I won’t get into, asks about regional variation in wage growth; outside of states hit by the fall in energy prices, moodys.com reports that wages have picked up in most parts of the country.)
In terms of wage trends across the distribution, a strong theme of my own work, often with economist Dean Baker, is that as the unemployment rate falls, the tightening job market disproportionately helps lower-wage and minority workers. I’ve got some new data on that important relationship, too, and I also will use this opportunity to correct a mistake from some of our earlier research in this area.
First, as there are many different wage series, a good way to evaluate evolving wage pressures is to mash them together using principal components analysis (a statistical technique that gives heavier weight to the series with the strongest signal). The figure uses five different series of wages and compensation to make some revealing points (see data note below for details).
Wage growth has mildly and “choppingly” accelerated in recent quarters. The mashup shows that a few years ago, wages were rising at a rate below 2%; now they’re growing a bit above 2%. But that pace is well below historical growth rates, going back to the 1980s, and the increase is extremely gradual compared to, say, the 1990s, when the job market eventually became a lot tighter than it is today.
For a variety of reasons, I do not consider this pace of wage growth to be inflationary:
–Chair Yellen has maintained that wage growth consistent with stable inflation is 3-3.5%, at least a point faster than the current rate (btw, why 3-3.5%? Because it equals trend productivity growth (~1%) + the Fed’s inflation target (2%)).
–Research has found little evidence of wage growth bleeding into price growth in recent years. For example, wage growth has been found to be of limited use in predicting inflation. The figure below shows the evolving coefficient on wage growth in a regression with prices as the dependent variable (see data note). It shows decreasing wage pass-through to prices and declining significance (the standard error widens and crosses zero for the most recent observations).
–As the job market has tightened, the bloated profit share of national income has been coming down a bit relative to labor’s share. This too is a source of non-inflationary wage growth.
–It’s reasonable to worry that rising wage growth amidst slow productivity could create price pressures (though the previous bullet describes a pressure value for that scenario). The variable you want to look to evaluate that possibility is the growth in unit labor costs, which tells you how fast compensation is growing relative to productivity. Since 2015, ULCs are growing at an average rate of 2.1%. That’s an acceleration over their 1.6% pace over the prior two years, but nothing much yet to see their folks…move along, please.
Turning to real wages, paychecks have definitely been growing in real terms, as the tick up in wage growth you see in the mashup has coincided with lower energy costs seriously pushing down inflation. This observation regarding the roles of depressed prices in real wage gains is important, because as price growth normalizes, wages will need to accelerate to grow in real terms. The mashup, the median, and the 25th percentile (lower-wage earnings) have all been growing by about 2% in real terms over the past year. Economist Elise Gould takes a more granular look at recent real wage growth and finds larger gains at the top of the distribution and, in states that raised their minimum wages, at the bottom as well.
History tells us that if tight job markets persist, larger benefits will accrue to lower- and middle-wage workers. Basically, when the economy sniffles, low-wage workers catch pneumonia. The reasoning is simple: the lowest-paid workers have the least bargaining clout, so they depend most on the tight job market to generate competition for their services.
Economists have shown evidence of this relationship both here and in the UK. My own work has often featured the impact of lower unemployment on low-, middle-, and high-wage workers. The table below contains the coefficients on unemployment for three different panel wage regressions, of 50 states (plus DC) over 36 years, so 1,836 observations (controlling for “fixed effects”). The results imply that a 50% fall in unemployment, say from 10% to 5%, would raise low wages by about 5%, middle-wages by about 2%, and high wages, not at all.
[Important note: this corrects earlier figures from these same regressions where we reported much larger elasticities. The key figure has been corrected in the online version of my book, The Reconnection Agenda, Chapter 3, Figure 2.]
So, wages are rising in nominal terms throughout the pay scale, and low inflation means most of these gains are real. Any feedback into prices is dampened by the low level of wage growth, rebalancing of factor shares, and a relatively low correlation that has developed over time re the price-wage relationship. If we can get to and stay at full employment, history shows that the benefits in terms of wage growth will accrue most to those with the least bargaining power. If that doesn’t happen, the extent to which wage inequality remains embedded in our economy and labor market means that these recent gains are likely be short-lived.
Wage mashup: This series is derived from a principal components analysis of annual changes in five series: ECI, wage and compensation, median earnings, production worker wages, and compensation from the productivity report.
The second figure shows the changing coefficient from a Kalman Filter analysis of CPI prices on the wage mashup (quarterly changes, annualized), controlling for energy prices and the value of the dollar in a broad, trade-weighted index. The “smoothed state” of the mashup variable and its standard errors is shown in the figure.
The table of coefficients is from panel regressions of log hourly wages on logged unemployment lagged one year. Data run from 1979-2015. I thank the Economic Policy Institute for the wage data.