Graph and song of the week

August 14th, 2015 at 3:58 pm

The Graph of the Week this week comes from Josh Lehner of the Oregon Office of Economic Analysis (graphs I’ve made myself are not eligible for entry in the GotW contest; if they were, I’d enter the one here).

The figure shows a pretty tight fit between state wage growth and labor market slack, broadly measured (Lehner uses Andy Levin’s slack measure, one I’ve frequently used here at OTE).


Source: Josh Lehner, OOEA

Given that we already knew that slack is negatively correlated with wage growth, what’s so special about this picture? The answer is a point that I’ve often heard economist Josh Bivens make. For a long time, when asked about why wages were largely non-responsive to growth in the economic expansion, we said “weak demand!” which you can take to mean the absence of full employment.

The years went by and we kept saying that, but people got tired of hearing it, and wanted to know why, at low unemployment–the current rate of 5.3% is almost at the rate the Fed says in consistent with full employment (5.1%)–have we not seen much wage growth? And don’t, they said, give me that “weak demand” stuff. What else ya got?

Well, as the figure shows, it’s still a demand story, especially when you measure slack by accounting for not just un- but also underemployment. No one’s claiming that’s all that’s going on–there are always many moving parts to a complex phenomenon like wage growth. But this one variable explains a third of the variance and that’s a lot in this context.

The Song of the Week is Driftin‘, an early jam from one of my all-time favorite jazz pianists, Herbie Hancock, with a great, laid back tenor sax solo from Dexter Gordon. I can’t get this damn song out of my head, so maybe it will help if I put it into yours!

(True story: I recently met the great Herbie and after professing my undying love for him and his work, I showed off my knowledge of his full oeuvre by signing this tune. I even knew what key it’s in and asked him if he remembered. He said D flat, but it’s E flat! So I had to correct him and which point he slowly backed away…)

Print Friendly, PDF & Email

5 comments in reply to "Graph and song of the week"

  1. Tom in MN says:

    So can you use this fit slope to predict what amount of employment gap reduction the country needs for the average wage growth in the country to go from the current 2% to the 3.5% that the fed expects to occur at full employment? Or does it not work that way? The slope looks like it says to go from 2% to 3.5% wage growth needs about 7% change in total employment gap. Is this right? That seems like a lot of slack, but then I’m not sure what the definition of total employment gap is.

    • Jared Bernstein says:

      For that calculation I’d use the national model I ran recently for this post:

      Here, a 1 percentage point decline in the underemployment rate raises nominal wage growth by 0.8%. I think the U6 rate is 10.4% and they say at full employment, it’s 9%. So, if it fell 1.4 ppts, that would boost wage growth a bit more than 1%, so maybe from 2% to 3%. Not quite the 3.5% target but closer!

      • urban legend says:

        Who is “they”? The U-6 got as low as 6.9% in 2000, the last time we had anything that felt to ordinary Americans like “full employment.” We only have a couple of decades experience with the U-6 figure, so I don’t know where 9% would come from.

        There’s no getting around the fact that U-3 applies an arbitrary definition of the labor force — excluding people who themselves would consider themselves unemployed or at best underemployed — and when you think about it, the “active search in the past month” standard puts the thumb on the scale in favor of (1) the government in power at the time, which can crow about low unemployment, and (2) the bankers whose priority for monetary policy is low inflation.

        I would posit that the reluctance of economists of progressive stripe to give U-6 at least equal if not greater billing then U-3 is that all their models are based on U-3. That made sense when there was a steady relationship in the two figures — U-3 as between 55% and 60% of U-6 — but lately it’s been just about 50%.

  2. Robert Goldschmidt says:

    Seems to me that economists have properly weighted the effects of the information technology revolution on employment. I recently estimated the improvement in computer price-performance over the past 50 years to be over a factor of one trillion — and it continues to rise. Never in human history has an amplification of human power of this magnitude taken place. Over this period of time we have gone from manual ledger entries for bank accounts to replacing human personal assistents with hand-held electronics. If we assume that the major impact has been to replace non-managerial workers, then we are about 20% of the way towards replacing humans with technology. The wages of non-managerial workers have fallen from 52% of GDP in 1972 to 42% of GDP today — and the seams are already showing in our record inequality, and the increasing political polarization and radicalism made possible by increased economic pressure on workers.