Surely one of the most important missing pieces of the current recovery is faster wage growth in response to the tighter job market. Well, last week’s strong jobs report hinted that wage growth could finally be ticking up, at least on average, though it’s too soon to tell for sure. These monthly numbers jump around a bit, but the yearly growth rate of average hourly wages has been slowly picking up speed since June, when it was 2%, through last month, when it was up 2.5%. These are all nominal values, but as inflation has been so unusually low, they imply more purchasing power.
This is not an entirely surprising development for those of us who have long argued that full employment is associated with higher pay, as employers, like it or not, must at some point offer higher compensation if they want to get and keep the workers they need to meet the demands for their goods or services.
That said, we’re not at full employment yet, at least by my preferred measure—the one cooked up by economist Andy Levin, which takes account of unemployment, involuntary part-timers, and those missing in action from the labor force. Each of those variables has contributed labor market slack in recent years, but the second two are left out of the traditional unemployment measure. At its peak in late 2009, the slack variable was 7 percentage points (ppts) above full employment. Now it’s 2 ptts above the target.
In order to quantify some of these musings, I’ve built a simple model of wage growth using our combined 5-series wage mashup (to generate a composite picture of wage growth), and regressed yearly changes of that measure against Levin slack, slack squared (as GS economist David Mericle has uncovered non-linearites is this relationship), and a few (3) lags of the wage-growth variable. I then project that the slack variable will continue along its recent downslope until it hits zero in late 2018.
The figure below shows that the model tracks the wage-growth series well, and it also picks up the recent acceleration, again suggesting that we may be at the cusp of faster average wage growth.
Annual wage growth and forecast
Projecting forward, the model predicts that at full employment nominal wages are growing at a rate of about 3.5% (the reason wage growth flattens at this point is because I don’t allow the slack measure to go below zero*). That 3.5 is an interesting number because it’s the wage target mentioned by Fed chair Yellen and others: it is the sum of underlying productivity growth (1.5%) and the Fed’s inflation target (2%), implying real average wages growing at the rate of productivity, aka steady unit labor costs.
OK, caveats abound: the model is very simple and could of course be wrong; who says we’ll get to full employment, especially if the Fed gets nervous and does more than just tap the breaks?; this will have to be among the longest of expansions if we’re to get to full employment by late 2018; as I’ve stressed, this is the average—in the age of inequality, the average takes off well before the median.
That said, work I’ve done with Dean Baker finds that if we can get to and stay at full employment for a while, the benefits in terms of wage growth will disproportionately go to middle and lower paid workers, including racial minorities (as economist Valerie Wilson has shown). They’re the ones with the least bargaining power in slack labor markets and thus with the most to gain from full employment.
Won’t this drive faster inflation? I don’t expect so. First, the reason 3.5% is the Yellen wage target is because real wage growth at the rate of productivity growth is not inflationary—productivity absorbs the real wage gains. One potentially problematic wrinkle here is that assumed trend productivity growth rate—1.5%–is…um…well above current trend productivity growth.
But more to the point, the evidence for wage-push inflation is just increasingly elusive. There’s now a number of solid, econometric studies questioning the importance of wage-price pass through; it’s actually pretty hard to find it in the data.
So, fear not rising wages. Fear them not at the average. Fear them not at the median or 20th percentile. Fear them not in a box…fear them not with a fox.
Whoops…got away from me a bit there at the end, but you get the picture.
And speaking of the picture above, here are the data if you want to play along at home.
*Zero on the Levin measure means that unemployment is at its “natural rate” according to CBO, and the involuntary unemployed and labor force participation have returned to the their pre-recession trend levels.