Many economists who subscribe to some version of the Phillips Curve have been puzzled by the tepid nominal wage growth in recent years, even though the official unemployment rate has collapsed and we should be in “overheating” territory at this point. (This same, falsified framework led Krugman back in 2010 to warn of impending price deflation—ending with “Japan, here we come”—though curiously, he hasn’t since apologized for using a model that gave the wrong predictions about price movements in the wake of the financial crisis.) In this context, Scott Sumner offers a much more straightforward explanation, based on his Market Monetarist framework [Ed.: see here], that attributes nominal wage growth to Fed policy, rather than any “micro” considerations. Yet as I’ll demonstrate, Sumner’s breezy explanation is seductively simple, and smuggles in a lot of his theoretical assumptions in the guise of macro tautologies.

To set the stage, let’s first quote from Sumner’s reaction to the economists who are tying themselves in knots over the apparent “puzzle” of sluggish (nominal) wage growth in the face of an apparently tight labor market:

Monetary policy drives NGDP [nominal GDP—RPM] growth, and NGDP growth (per worker) is by far the most important determinant of nominal wage growth. (The other determinant is labor share of GDP.)

During the past 4 years, NGDP growth has been running at 4.05%/year, well below the historical norms. So why is wage growth running at only about 2.5%/year? The answer is simple; payroll employment has been growing at 1.78% over that same period. The predicted growth in average hourly earnings is 2.27%, whereas actual wage growth has been running at 2.47%. The recent puzzle is why is wage growth so high, not low.


So why are so many people puzzled over the lack of wage growth? They make the fundamental error of confusing micro and macroeconomics, unfortunately a common mistake among Keynesians. Factors like international trade, tight labor markets, etc., may affect relative wages in particular fields. That’s microeconomics. But at the macro level, nominal wages are mostly driven by monetary policy. (Of course real wages are a different story, heavily influenced by productivity.) [Scott Sumner, italics in original, bold added.]

I readily concede that if we independently thought that Sumner’s framework were the best way to think about the Fed policy, then its ability to effortlessly “solve” the unemployment/wage puzzle would be gravy. Yet as I have pointed out earlier, Sumner’s whole framework relies on a definitional trick.

Specifically, Sumner defines the stance of monetary policy in terms of NGDP growth. Consider the following graph taken from the St. Louis Fed, which plots the level of nominal GDP (blue line) against the total assets held by the Federal Reserve (red line):

2018.05 Problems With Market Monetarist Explanation graph 1.png

According to Sumner’s framework, the Fed foolishly tightened monetary policy in the recession period (denoted by the middle gray bar). We “know” this because the blue line not only stopped growing, but actually fell for a bit. Sumner would describe this as Fed tightening, even though the Fed funds interest rate fell to virtually zero percent, and even though the Fed more than doubled its balance sheet in a few months’ time (the sharp spike in the red line).

Now consider a graph of the 12-month percentage growth rate of nominal wages:

2018.05 Problems With Market Monetarist Explanation graph 2.png

In the above graph, we see that the growth rate of nominal private wages bounced around 4 percent in the late 1990s before falling sharply once the dot-com bubble burst. After bottoming out in 2004, wage growth picked up steam, again reaching the 4 percent range by 2007. Then again it collapsed when the financial crisis struck.

Now according to Sumner’s reasoning, it would be naïve and “micro” to say the wage movements had anything to do with the fact that recessions occurred in the early 2000s and in 2008-2009. Instead, the way to explain the above pattern in wage growth is that the Fed tightened monetary policy in the early 2000s and from 2008 onward. This is the case even though these periods were associated with rapid reductions in the Fed’s policy interest rate and with large increases in Fed asset purchases. (Here’s how the Fed “stimulated” the economy in the early 2000s, which—I argue—helped fuel the housing bubble.)

Does anyone really want to bite the bullet and say that? To be clear, Sumner’s approach doesn’t “throw out” the valuable information about the state of the economy if he were trying to predict the movement of nominal wages. Rather, his metric of NGDP growth captures it, in the sense that NGDP growth collapses during major recessions. But it’s the first step in his argument—when Sumner claims that monetary policy drives NGDP growth—that is (in his hands) virtually tautologous and nonsensical, as it forces us to say things like, “[M]onetary policy during the 2008-13 [period] was the tightest since Herbert Hoover was President.

Finally, let me point out that if Sumner were correct, then the contrast between his approach and that of (say) Krugman doesn’t really have anything to do with macro versus micro, as Sumner seems to believe. I could just as well “explain” the wages of a particular individual using Sumner’s framework, too.

To see this, suppose a father sees his 25-year-old son lounging on the couch watching reruns. The father exclaims, “Jimmy, this has got to stop! All through your teenage years you kept getting raises and better jobs. But after your messy breakup last year, you’ve fallen apart. Your mother and I shouldn’t have let you move back in. You get a job within two weeks or we’re kicking you out!”

Then Jimmy, who has spent some of his copious free time reading the work of Scott Sumner, responds to his father in this fashion:

“Dad, dad, stop thinking like a micro economist. You seem to believe that my effort determines the pay I earn. But actually, monetary policy drives NGDP growth, and NGDP growth (per worker) is by far the most important determinant of my wage growth. (The other determinant is my income as a share of GDP.)”

Obviously Jimmy’s flippant response is not as ludicrous as Scott Sumner’s explanation of sluggish wage growth, but if you can spot the problems with Jimmy’s response, then you can see the analogous shortcomings in Sumner’s.

This article is republished with permission from