Scott Sumner  

Is it OK to reason from a (aggregate) wage change?

PRINT
Farewell, James Schneider... Donald Sterling and the Econom...

Even before Mark Carney took over as Governor of the Bank of England, he showed signs of independent thinking. Most famously, he came close to endorsing NGDP targeting at a speech in Toronto during early 2013. More recently, he hinted that aggregate nominal wages are a useful indicator (and target?) of monetary policy:

Has Bank of England Governor Mark Carney already decided to target wage growth as a sign the economy is improving enough to start raising interest rates?

In an interview with the Bristol Post newspaper published Tuesday, Mr. Carney declared that "substantial" wage rises will be needed to ensure the economic recovery that began last year is sustained.

That is an unusual message for a central banker to deliver. Usually, they warn against large wage rises for fear they will lead to a self-perpetuating inflationary spiral. But Mr. Carney's comments reflect the fact that central banks across the developed world are facing an unusual challenge, with inflation rates too low rather than too high.

. . .

Carney's remarks reflect an increasing focus on wages in policy circles. Two former Bank of England officials who are based at U.S. institutions have recently urged the Federal Reserve to focus on evidence of wage gains rather than falling unemployment as a sign the economy is improving enough to start raising interest rates.

Danny Blanchflower, a professor at Dartmouth University, and Adam Posen, president of the Peterson Institute for International Economics, argue in a draft paper that "unlike unemployment, the rate of wage inflation requires less judgment and is subject to less distortion by such factors as inactivity."


Back in 1995 I advocated a policy of using monetary policy to target the price of aggregate nominal hourly wage futures contracts. Later I learned from David Glasner that Earl Thompson had suggested this idea back in the 1980s. The argument goes as follows. Nominal wages tend to be very sticky, but do change gradually when macroeconomic forces alter the equilibrium wage rate. When we observe a change in the path of aggregate nominal wage increases, we can infer that the wages of those workers under existing wage contracts have fallen behind the equilibrium wage rate. Thus a policy that keeps aggregate nominal wages growing at a steady state will minimize the number of workers whose wage is not at equilibrium.

Of course most people favor inflation targeting. But there is a lot of evidence that disequilibrium in the labor market is the key macroeconomic problem. Thus it's nice to see renewed interest in this idea. I later drifted away from nominal wage targeting and ended up favoring NGDP targeting, mostly for pragmatic reasons. But I still believe that a monetary policy that keeps nominal wages growing at a steady rate is likely to lead to a much less volatile business cycle.

So if I am right, it is permissible to reason from a nominal wage change, but only because actual nominal wages are very sticky. Thus sudden changes in the path of nominal wage gains imply labor market disequilibrium.

However it is not permissible to reason from a real wage change, as this could reflect either supply or demand-side factors. It seems to me that the following paragraph comes close to making that error:

Although the U.K.'s unemployment rate has been lower than many economists would have expected given the poor performance of the economy since the financial crisis, wage growth has been very weak, and below the inflation rate for much of the last five years. There are signs that is changing, with real wages likely to rise this year.
I'm troubled by the term 'although,' which leads off the paragraph. It seems to imply that the coexistence of low real wages gains and lower than expected unemployment represents some sort of paradox.

I'm reminded of those NYT headlines that conservatives like to mock; "prison populations rise despite falling crime rate." Perhaps unemployment is lower than expected precisely because wage gains are moderate. Yes, that's not possible in Paul Krugman's brand of Keynesianism, but that doesn't mean it's not possible in the real world.

So by all means go ahead and "reason from a nominal wage change," especially if you are a central banker. But under no circumstances should you ever reason from a real wage change. The forces of both supply and demand explain real wages. And that scissors has two blades.

HT: TravisV


Comments and Sharing





COMMENTS (4 to date)
BH writes:

Interesting thing to note - the basic New Keynesian model results in price level targeting as the optimal monetary policy. However, this is because prices, not wages, are sticky. If the NK model had sticky wages instead, nominal wage level targeting would be optimal policy.

Scott Sumner writes:

BH, Good point.

Rajat Sood writes:

Scott, is there is a risk with nominal wage targeting that unions or wage arbitration tribunals (in countries like Australia that have such things) might be more inclined to seek nominal wage rises at or above the target, just because it's there, unhelpfully pushing AS left? I could easily see our 'Fair Work Commission' pushing all minimum wages (and there are many, applying to lots of different jobs at many different levels) up every year at the target, even when something lower is needed.

Scott Sumner writes:

Rajat, That might be true in some countries. I doubt it would be an issue in the US, but in smaller countries with more centralized wage setting processes it might well be.

Comments for this entry have been closed
Return to top