An “anchored expectations” model for wage stagnation?

January 22nd, 2018 at 12:32 pm

I found this Bloomberg View piece by Conor Sen to stimulate the old noodle re why wage growth has generally been lagging. That is, given the persistent tautness in the job market, I’d expect a bit more wage acceleration than we’ve seen.

Before I get to Sen’s points, however, a few facts to consider. One, based on still-recovering employment rates for prime-age workers, the job market is still not as tight as the unemployment rate suggests. Two, some series do show faster wage growth. The figure below show solid recent growth in the real earnings of low-wage, African-American workers (25th percentile). The data are very jumpy so I added a trend, with a trend-based forecast at the end of the series. Still, most other series are not this positive. Third, productivity growth, though it has picked up a bit in recent quarters, remains low, and that too is a constraint on wage growth.

Source: BLS

Sen’s argument is that even employers who need new workers are avoiding raising their wage offers because they don’t want to have to raise the pay of their incumbent workforce.

When an employer in the Boston district was asked why the company didn’t raise wages as a way of attracting more workers, it responded that if it had done so, it would have had to pay all the existing workers more, which would be uneconomic. And another contact in the Boston district said that when a worker departs, the replacement typically ends up earning 10 percent more than the departing worker made.

This implies that incumbents could do much better if they left their jobs paying well below what they could get elsewhere given current conditions. That would show up as more labor-market churn, which has, in fact, been down in recent years (the economist Betsey Stevenson has long emphasized this development).

If so, then perhaps incumbent workers are just too used to a labor market market characterized by stagnating pay so they don’t bother looking for a better job elsewhere, under the assumption that, pay-wise, their next job would be no better than their current one. It’s “well-anchored expectations,” applied to a job market long characterized by wage stagnation. And just like with the Fed’s inflation problem–undershooting their inflation target for years on end–it takes a long time for expectations to change.

The problem with the theory is that while the stock of workers is much bigger than the flow, the flow is non-trivial, so I’m not sure how new workers getting jobs in firms that must pay them more in periods like now fit into this explanation. Sen implies that this hasn’t happened yet, though it’s coming:

A company can survive for a while by leaving vacancies unfilled and saddling remaining employees with the work. But eventually key positions open up that absolutely must be filled, and staffing has to be maintained to ensure an adequate level of service. When that time comes, the companies who have been most focused on keeping labor costs low, and have shown industry-leading profit margins to investors, may be in for the most pain.

It’s a bit hard to square that with the creation of millions of jobs per year, so one should still consider shaving with Occam’s Razor here: as the job market tightens further, we’ll see more wage gains both in stocks and flows (i.e., old and new workers). But Sen’s interesting observations do square with my own that US employers have just thoroughly forgotten how to raise pay, and they don’t have too much by way of unions to remind them.

Print Friendly, PDF & Email

8 comments in reply to "An “anchored expectations” model for wage stagnation?"

  1. Prairie Populist says:

    In the plantations of the Antebellum South, there was no unemployment yet wages remained stable at subsistence levels. All power resided with plantation owners who “paid” slaves in kind whatever the owners wanted to pay. Extreme example of market failure to be sure, but it illustrates the propesnsity of power imbalance to override supply and demand in setting wages. Since we eliminated labor unions for all but a few remnants in the labor market, employers – like the plantation owners of yore – use their more powerful position to pay weak disorganized employees whatever employers want to pay. This is not the whole story, but it is a significant factor.


    • Smith says:

      Smaller union membership is practically more a symptom than a cause. If manufacturing is relocated, that tends to cut union membership. The fall in wages and the fall in union membership is a correlation not a cause.
      Also just how much are unions to blame for their own decline? They exactly have a great reputation. They are known for corruption, onerous work rules, job security ending at will employment, seniority rules, dues that take out 1 to 1 1/2 percent of salary. The president of the UAW makes $170,000, the auto companies were able to blame workers when they nearly went bankrupt, the unions allowed a two tier wage system where new workers are paid less than new workers in the past. The teachers union in Wisconsin (or striking teachers in Chicago) didn’t exactly win over the populace either. How are unions not also to blame for their decline?


  2. D. C. Sessions says:

    The “if you want more money go somewhere else and then come back” model was once known as the “Silicon Valley rotation” back in the 70s semiconductor industry. Tracey Kidder’s Soul of a New Machine” covers some of that practice.


  3. Smith says:

    This profit loss seems far fetched. Suppose your average company lost 20 percent of it’s workforce and replaced them with workers earning 10 percent more. For the average company the cost of labor is about 20 percent of their costs. You following? Lets take the 20 percent new workers of the 20 percent costs and add 10 percent. .2 x .2 x .1 = .004 In this pretty much worst case scenario, they are losing a little less than 1/2 percent. Of course productivity in very secret manner that only one or two people know about, myself included, was zero for the whole year of 2016, but a normal increase, even in our pathetic economy, manages to be 1/2 to 1 percent. (historically it’s still 2 to 2 1/2 percent, and 3 percent in the 1950s)


  4. Smith says:

    Employers have forgotten nothing. Economists are flummoxed in the face of evidence that wages aren’t determined by supply and demand. Wages are certainly influenced by those tow factors, but one can’t take the two curves, figure out where they cross, and see an equilibrium. Wages are controlled by the same people who control the means of production. Adam Smith would say wages are not determined by the supply of workers. Ricardo too.
    This notion that there is a free market where employers compete and bid for labor is nuts.
    Since the 1980s the managers have been trained to cut costs, downsize, consolidate, to make do with less workers, automate, outsource, insource, put everyone on salary so there is no overtime pay, easy to do in a service economy of office workers.
    Wages have been stagnating since 1980 according to your graphs, and the charts from Saez and Piketty. I’d expect at least a few quarters of wage gains as recovery continues, but why are you thinking some fundamental change is about to occur in the economy after 37 years of stagnation? Are aware that inequality widened even in the late 1990s the period of real wage growth outside the top income groups?


  5. Person says:

    I believe a big problem is too much information. In the old days of poor information the main mechanism for wage gains was trial and error. If an employer couldn’t attract workers they would try raising the wage. Now they have huge databases of employee wages, and most employers know the so-called going rate quite well. They know that other employers will behave much as they behave, which is to avoid ever offering anything above that prevailing wage.

    Noisy data is better than perfect data for helping markets find their actual equilibrium. There’s a lot of money in favor of perfect data, and no money in favor of noisy data.


  6. Prairie Populist says:

    Of the three classic factors of production – capital, labor, materials – capital has become much more important than the other two in recent decades. There are a couple of reasons. First, labor is the easiest to increase or decrease in response to changes in market demands for product . A capital investment can’t be “laid off” on Friday, a worker can. Most industries can’t control the price of the material they need. Second, cheap money in financial markets means that investment in labor-eliminating technology is now cheaper.

    More problematic is how the elimination of so many millions of human participants from the economy will play out. Right now about 70% of GDP os consumer related. Laid-off workers are bad consumers. Will machines step up and buy each other’s products?


Leave a Reply

Your email address will not be published.