"we are seeing an almost perfect example of the "self-correcting mechanism" after a fall in AD, right out of EC101 textbooks."-Scott Sumner

See It's all in EC101.

"Here is the mainstream media:

WASHINGTON (AP) -- This isn't explained in Econ 101. 

Month after month, U.S. hiring keeps rising, and unemployment keeps falling. Eventually, pay and inflation are supposed to start surging in response.

They're not happening.

Last month, employers added a healthy 252,000 jobs -- ending the best year of hiring since 1999 -- and the unemployment rate sank to 5.6 percent from 5.8 percent. Yet inflation isn't managing to reach even the Federal Reserve's 2 percent target rate. And paychecks are barely budging. In December, average hourly pay actually fell.

Economists are struggling to explain the phenomenon.

"I can't find a plausible empirical or theoretical explanation for why hourly wages would drop when for nine months we've been adding jobs at a robust pace," said Patrick O'Keefe, chief economist at consulting firm CohnReznick.

Perhaps those "struggling economists" should take an intro to macro course, where they might come across this graph:

The severe negative demand shock of 2009 put us at P2. Then wage growth moderates and we are currently moving between P2 and P3. That's EC101.

Now obviously you wouldn't have so many people confused if this occurred every single business cycle. In the past the Fed has often adopted expansionary monetary policies during recoveries. Then the adjustment burden does not entirely fall on lower wages, rather both AS and AD shift to the right.

In that case wage growth often accelerates. But this time the Fed only allowed 4% to 4.5% NGDP growth during the expansion, which is actually less than the long-term average growth of both expansions and recessions. That's a startling level of restraint. So the full burden of adjustment has fallen on aggregate supply. But the key point is this---given the low rate of NGDP growth, we are seeing an almost perfect example of the "self-correcting mechanism" after a fall in AD, right out of EC101 textbooks.

PS. I cheated a bit here because these EC101 graphs tend to deal with levels, and I am following the more modern practice of looking at rates of change. In level terms AD has grown, but because it's grown less that normal, less than expected, it shows up as a negative AD shock. That partly explains the unusually slow recovery. You'd get the same general result with a more realistic dynamic model."