Graphs to Accompany Bob’s Commentary During Contra Krugman Episode 117
[EDITed the section about construction employment.]
For this episode, we covered this blog post by Krugman, in which he says that the unemployment rate has now fallen down to pre-crisis levels, and so–duh–the “demand siders” were right, while the “structuralists” were wrong. The reason so many people were unemployed in 2009 and 2010 is that people weren’t spending enough; the claims about a mismatch between worker skills and job openings have been refuted by the data. After all, if the people claiming “structural unemployment” were right, then we should’ve seen spikes in wage growth, among those people who *did* have the desired skills. Yet wage growth has been sluggish, and is only now just recovering, showing that it was always a problem of demand, stupid.
I won’t spoil my responses, but in my commentary I referred to a bunch of charts. So I’ll collect them all here in one place.
The first chart below shows that even according to Krugman’s own narrative, wage growth right now is surprisingly low. Looking back several decades, anytime wage growth (the red line) was as low as it is now, the unemployment rate (blue line) was much higher than it is now. On the other hand, looking back over the decades, anytime the unemployment rate (blue line) was as low as it is now, you’d see wage growth (red line) much higher than you see now. So, my point is that the Keynesian explanation, where the unemployment rate is driven by “slack” in the labor market due to insufficient demand, doesn’t really match the data either. If Krugman’s basic narrative were right, then wage growth should be a lot higher than it actually is right now. And whatever tweak you use to explain the fact away for Krugman’s story, the structuralists could use as well to rescue their story.
This second chart shows total construction employment. Notice that it is still some 700,000 below the pre-crisis peak. This is exactly what the structuralists were saying: employees laid off from construction needed to go elsewhere. If Krugman is saying the labor market is basically back to normal, then the structuralists were right, and he was wrong. [EDITED to add: I agree with Transformer in the comments, that Krugman’s demand-side story isn’t obliterated by a net exodus of workers from construction. However, I’m saying that this evidence lines up better with the structuralist story than with the demand-side story. Back in late 2008, Krugman was mocking John Cochrane for saying the recession was necessary in order to move workers out of construction in Nevada and into other channels. Now in 2017–nine years later–Krugman is looking at what he believes is a fully employed labor market, with 700,000 fewer people in construction, and he’s saying, ‘Wow those structuralists were morons.’ Uh, this is exactly what Cochrane would’ve expected things to look like.]
The third chart below shows the proportion of prime-age males who are employed. To me, this chart indicates that the official unemployment rate is very misleading. Those claiming that globalization, bad schools (look at the chart on page 13 here), growing regulations (including the employer mandate under ObamaCare) were hampering the labor market, could point to this chart and say, “I told you so.” (Note that another factor is the explosion in male incarceration rates that started during the 1980s.)
Finally, look at the mean duration of unemployment spells. Krugman’s basic story shouldn’t predict that the duration would be so much higher during the Great Recession, should it? But it makes much more sense if the structuralists and also guys like Casey Mulligan (who was blaming the extension of unemployment benefits and other incentives) were right.
Contra Krugman is the highlight of my commute. I particularly liked the pushback against Krugman’s claim that there are no conservative economists with any principles. Yet I must quibble with one thing here:
“Those claiming that globalization, bad schools (look at the chart on page 13 here), growing regulations (including the employer mandate under ObamaCare) were hampering the labor market, could point to this chart and say, “I told you so.”
I’m sympathetic to that view, really I am. But the slow downward trend in prime age male employment rate is just that SLOW. The trend has been going on for 60 years. How much of the decline from 1958 till now is likely to still be involuntary unemployment? A small bit maybe but how much? I would think that for the most part wages would adjust downward over the long time spans we are talking about and that is how the damage from government mandates and poor education manifests itself not in a half century long increase in involuntary unemployment.
Krugman does address the lack of real wage growth. He acknowledges this low growth and adds:
‘we are indeed more or less back to full employment, although those wage numbers suggest that we still have a bit further to go.’
Also, as Krugman probably believes there was a housing bubble (in fact didn’t he famously call for it to be created ?) I doubt he will have much of a problem with the construction workers numbers – he would probably think these can be explained as bubble v post-bubble levels.
Transformer, I agree with you that you’ve correctly stated how Krugman would deal with these two points. But I’m claiming that his attempted escape routes don’t really work.
You wrote: “Krugman does address the lack of real wage growth. He acknowledges this low growth and adds:
‘we are indeed more or less back to full employment, although those wage numbers suggest that we still have a bit further to go.’”
Sure, but look again at my graph. This isn’t a matter of the wage numbers being a tad low. No, for this level of unemployment, we normally would have expected wage growth to be above 4%, instead of what it actually is (slightly above 2%). During the last recovery, when wage growth had recovered to this rate, unemployment was still around 5.5%.
In other words, Krugman is acting like the unemployment and wage growth numbers are lining up pretty great for his story, when in fact they are out of whack.
Transformer wrote: “Also, as Krugman probably believes there was a housing bubble (in fact didn’t he famously call for it to be created ?) I doubt he will have much of a problem with the construction workers numbers – he would probably think these can be explained as bubble v post-bubble levels.”
Again, I agree with you that that’s what Krugman now has to say, but does it line up with the way he pooh-poohed the housing bubble having anything to do with unemployment, back when he thought the numbers were on his side? At that time, he agreed that a connection between unemployment and the housing bubble would be a feather in the cap of the structuralists, and since (at that time) he doubted the connection, he said it proved the demand-siders were right. So he can’t very well now flip and try to say that a massive net exodus from construction employment that took years to accomplish, is evidence against the structuralists.
Transformer I edited the main post to reflect your point about construction.
This just goes to show that the worst outcomes happen on things that you and Krugman agree on.
The real problem with the US economy is that those construction workers should never have lost their jobs and there was no reason for them to have to shift to other sectors.
Kevin, do you have any objection to me quoting your two recent comments–on average duration of unemployment and this one, on construction workers–in a new blog post, and then responding? To warn you, I think you’re totally wrong, but I will be civil if that affects your answer.
Sure. I’m surprised. I figured you would agree about emergency UEI and durations. It looks to me like median durations under 45 years are pretty close to previous cyclical bottoms while older medians and averages for all ages remain high.
On construction employment, I have a pair of books coming out in a few months that makes the case. It may be hard for me to respond with the full case in the form of blog comments.
I have a ton of novel evidence regarding construction. One simple piece of evidence is that housing units per adult was lower in 2007 than it had been in the 1990s and is much lower now. That’s just something you can easily check from census data, but the full case is a bit complicated.
In a way, I agree with you that there is a structural problem in American labor markets, and the largest problem is that we have reversed natural migration patterns. Because of urban housing restrictions, American households are moving out of cities with high incomes to cities with lower incomes. This is perverse.
This is related to my claim that there were never too many construction workers. Even in 2005, our problem was that we didn’t have enough houses.
But, here, supply side and demand side play a dual role. Since there were many misunderstandings about what caused the housing bubble, we added a demand-side shock to a supply-side problem, and the demand-side shock has worsened the supply side problem because the severe tightening of lending markets has blocked the natural conduit for new housing. Rents have risen, and an increasing portion of national income is flowing to real estate owners. This is why real wages are not as high as they should be. It’s all going to the landlord.
“Because of urban housing restrictions, American households are moving out of cities with high incomes to cities with lower incomes. This is perverse.”
I agree with structural issues begin significant, but how do you account for cost of living differences when you make this claim? I know people who moved from ‘high income’ cities to ‘low income’ cities and saw their standard of living jump way up. My sister and her husband moved from a D.C suburb to a rust belt city and were able to afford a much nicer house, have better commutes and are probably in a better school district. We should be, in theory, perfectly happy with people moving to lower earning areas with other benefits, or low cost of living. Nothing perverse about that.
No.
If we put a 100% tariff on foreign cars and Americans choose to buy domestic cars as a substitute, their decision is reasonable. The perversion would be that the foreign cars have been made arbitrarily expensive. The fact that they have been induced into a suboptimal decision is the perversion.
Hi Kevin,
I can’t get to this for a bit; I’m about to travel. You’re right, we do agree on a lot of stuff, but definitely not on the median duration thing. I looked quickly at the housing stats and I admit it is not as slam-dunk a case on that score as I had thought.
Bob,
Jeffrey Snider has been beating this drum for quite a while now. Is there something “wrong” with the labor market? Yes, there is, and it shows up most clearly here, which is the only place it really matters.
One more graph that could bolster your case is one that shows immigrant labor over this span. IIRC there is ~ 1 million worker shortage over what you would have expected extrapolating immigration patterns from the 90s and early 00s, which obviously won’t show up in UE statistics, but nonetheless aren’t in the US pumping out GDP in GDP factories.
Fails to challenge Krugman’s ethical judgement that American wage growth is a good thing. Maybe many other people throughout the world would be better off if Americans couldn’t afford so many bombs and drones and nukes and various items made by slaves!
Also makes a startling revelation regarding Walter Block’s views on the topic of neighborly behavior. Given Walter Block’s apparent lack of regard for other people’s cats, it is not surprising that at least one libertarian (not Walter Block) committed an act of rape to punish someone for trespassing. Libertarian depravity knows no limits.