When the Curve Failed

I found Paul Krugman’s New York Times column particularly good today. My observation has been that the more closely he hews to economics, “wonkish” as he puts it, the better. In this particular column he muses over why wages have been flat:

From the mid-1990s until the 2008 financial crisis, it looked as if there was a fairly stable Phillips curve – a relationship between unemployment and the rate of wage growth. This wasn’t the “accelerationist” Phillips curve that underlies the concept of the natural rate of unemployment – there was no sign that low unemployment led to accelerating inflation. It was, instead, a neo-paleo-Keynesian relationship between unemployment and wages. You can see how good the relationship looked in Figure 1, in which wage growth is measured on the left axis and the inverse of the unemployment rate on the right axis…

[…]

After the crisis, wage growth dropped – but only about as much as it dropped in the early 2000s slump, despite much higher unemployment. And it has increased only modestly since then, to levels well below pre-crisis growth, despite unemployment rates as low as we’ve seen in a very long time…

He proposes several possible explanations: despite the very low unemployment rate there’s also a lower rate of employment and monopsony for labor, particularly in local markets.

In the Wall Street Journal Federal Reserve Governor Neel Kashkari has an answer of sorts:

Today’s flat Phillips curve and low long-term interest rates are not mysteries. They are products of a bipartisan political consensus that allows the Fed to focus on achieving its inflation and employment objectives rather than short-term political goals. If Americans want to keep this economic expansion going, we all should recognize the important role Fed independence plays in keeping inflation in check.

That’s something of which we should never lose sight. Economics isn’t a “hard” science like physics or chemistry and the Phillips Curve isn’t like plotting the orbit of a planet. It’s an artifact. It’s the result of human behavior and human responses to change.

One more thing. The graph that Dr. Krugman presents in his Figure 1 of “average hourly earnings of production and non-supervisory employees” shows only wage income. It does not reflect total compensation. Although growth in total compensation has been slow relative to the period before 1980, it’s not as slow as you might conclude from his graph. Just as one example you would expect wage income to rise more slowly starting in 2014 and, indeed, the data more recent than Dr. Krugman’s graph suggest that was the case.