The new jobs figures and macro theory
By Scott Sumner
Each January the government revises its payroll jobs estimates, to provide more accurate figures. The recent revisions are not dramatic, but do slightly strengthen two arguments that I’ve been making over the past few years. Let’s look at the previous and revised estimates for jobs growth over the past for years:
Year — Previous — Revised — Wage growth
2011 — 2.083m — 2.080m — 2.0%
2012 — 2.236m — 2.257m — 2.2%
2013 — 2.331m — 2.388m — 1.9%
2014 — 2.952m — 3.116m — 1.9%
The 2013 figures still show no evidence of fiscal austerity slowing growth, indeed the speed up in jobs creation (over 2011 and 2012) was revised slightly upward (albeit not significantly.) Paul Krugman’s prediction that growth would slow is still wrong. And Krugman was also wrong about 2014. He expected that the ending of extended unemployment insurance would not lead to more employment, but the new figures suggest that employment growth in 2014 was significantly higher than in 2013 (728,000 more.) That’s 0.5% of total employment; consistent with my earlier claim that the extended UI program may have boosted the unemployment rate by roughly 0.5%. Brad DeLong made a similar argument in 2008. Both AD and AS matter, even in a recession.
The usual caveats apply—no single data point is definitive, and of course these tests were proposed by Krugman.
One other point, the extended UI test is further clouded by the ambiguity surrounding the date of its demise. The program ended at the beginning of 2014, but it was not until midyear that it became clear that Congress would not extend the program retroactively. So its effects may have played out gradually during 2014.
The wage growth figures are consistent with the standard AS/AD model. After a sudden drop in AD (during 2008-09) unemployment rises sharply. Then wage growth moderates and the SRAS curve gradually shifts to the right, bringing the unemployment rate back to its natural rate (which itself varies over time, but by much less than the actual unemployment rate varies.) The people who predicted that unemployment would get stuck at high levels were wrong.
I see no sign of “hysteresis,” the theory that a demand shock can permanently increase the unemployment rate. Many Keynesians proposed that theory to explain why European unemployment rates stuck at high levels (in both booms and recessions) after increasing sharply in the 1970s and 1980s. I don’t think they were willing to face up to the fact that the European welfare state and/or labor market regulations were causing the high unemployment. You don’t see that tendency in the US, Switzerland, and East Asian developed countries. In fairness, even welfare states may be able to avoid high natural rates of unemployment with flexible labor markets and/or sensible wage subsidy programs (as we’ve seen in Denmark, and more recently in Germany.)
The AS/AD model is also the best way to explain the situation in the eurozone. The overall rise in unemployment is due to sharply lower growth in AD (NGDP.) But the variation between countries is also explained by supply-side factors, especially the structural problems in the Mediterranean countries. Greece faces two problems, too little AD and too little AS. Given their debt problems, the only way Greece can significantly address the AD problem is by leaving the eurozone. The only way it can address the AS problem is by cutting minimum wages, deregulating, and privatizing. The new Syriza government has proposed exactly the opposite—less AD and less AS.