A Sticky Situation

There have been a few posts on sticky wages recently, and I'm sure there will be more.

Tyler Cowen starts it off by pointing out that the highest unemployment rates are in low-wage, low-education areas where we've been hearing for years that wages have been stagnant or falling. Cowen asks: "Doesn't that mean...um...their wages aren't so sticky downwards? And thus Keynesian economics is not the final story?"

He can't even manage to convince his own colleague. Bryan Caplan calls him "completely wrong". He raises two points - first, that the wages for this subset of the labor force took decades to erode, and "even staunch Keynesians" think the economy could right itself in that time (somebody please tell Wayne William Anderson, who thinks Keynesians believe economies will never recover without the government). A fair enough point by Caplan. Caplan's second point is better, though - the downward pressure on wages in this subset of the labor force was primarily due to the slow erosion of inflation, not flexible nominal wages. It is nominal wage rigidity that Keynesians who cite rigid wages generally refer to. Then he gets into (well, links to lecture notes on) - you guessed it - unions and the minimum wage. Clearly the scourge of the American economy right now, right? I'm not particularly sympathetic to unions, but give me a break. When they were strong we were prosperous, they got weak and you're still citing them? Unions can do major damage, but it's barely worth mentioning as a contributor to what's happening. Everyone has their pet peeve, I suppose.

Anyway - I would add one things to Caplan before moving on to another Keynesian response. What makes Tyler and Bryan even think rigid wages are necessary in the first place? I refer readers to Chapter 19 of the General Theory, which discusses nominal wage adjustments. Keynes says quite clearly that sometimes nominal wage adjustments will clear the market, and sometimes they won't. Nominal wage rigidities as an explanation for unemployment preceded Keynes (A.C. Pigou wrote about them in 1926 in his book Industrial Fluctuations), and they've only become identified with Keynes since the New Keynesians. It's a useful thing to point out, but it's not what Keynes said drives depressions.

Nick Rowe agrees with Caplan that Cowen's post comes up lacking. Rowe points out that in a recession caused by sticky wages where the decline in demand for labor is carried entirely by quantity adjustments, firms make offers to high quality labor and unemployment hits low quality labor the hardest. So low quality labor with (as Caplan points out) rigid nominal wages, and falling real wages due to inflation has the highest unemployment rates by virtue of its lower quality. In other words - what we see is precisely what you would expect from sticky wages anyway.

I'd also add the point that our unemployment rate is self-reported and not necessarily consistent with the economic concept of "unemployment". If you think about a supply and demand model where demand shifts to the left and adjustment of wages is perfectly smooth, you'll move from one equilibrium with no unemployment (Q1, below), to another equilibrium with no unemployment (Q2, below).

Now, if the market operates as the simple model suggests that it should (and I'm not saying that it does in practice), then the Q1 minus Q2 workers are (1.) not employed, and (2.) will not accept employment at the new wage level. The problem is, the Current Population Survey (where we get our unemployment rates) doesn't tell them what the new wage level is (because the CPS staff doesn't know that!), they don't ask if they've been offered a job at that wage level and what there response was, etc. etc. They ask if they're employed or not, if they're looking for work, and if they are ready, willing, and able to work. That's all. The people sitting between Q1 and Q2 might be holding out for jobs at the old wage, not knowing that there's a new wage that economic theory says they should be refusing! Simple theory acts like the labor market is an auction. It's not - it's a search process. Some people may give up the search if someone told them what the new equilibrium wage was, but since they don't know that they keep searching. Or some people may work at the equilibrium wage if they knew they weren't going to find a better one. In the table above, there is no unemployment, but that doesn't mean people between Q1 and Q2 won't register as unemployed in the Currrent Population Survey! In other words, no unemployment as it is theoretically defined and lot's of unemployment in the CPS numbers can very easily coexist! The only thing the simple model tells us that we can reliably empirically verify is that even if there is no unemployment as economic theory defines unemployment, there will be less people with jobs. There comes a point where it doesn't matter if there is no unemployment according to theory - there is joblessness in actuality that we don't want (nor, I might add, that we are powerless to do something about).