Unemployment, mobility and foreclosures
Kyle posted a very interesting picture not a long time ago (see above) according to which there is a clear increase in the variance of the variance across states in 2008-2011 that coincided with the the Great Recessions. So why don’t the unemployed workers move from the most troublesome states such as California, Nevada, Arizona to the East, where job prospectives seem to look more promising? It seems like moving to a different state is very costly for the workers that the variance in the unemployment rate does not converge to its 2001-2008 level. Presumably a decrease happened somewhere between 2005 and 2008, because pre-2005 spike and after-2005 levels of the variance are essentially the same. The question is that whether foreclosure delay explain this lack of mobility?
There are two main points that I wanted to make here. One is one is my own that I have not seen mentioned anywhere (yet?) and the other one is a summary of the existing evidence on the problem.
In the ongoing debate about whether the unemployment rate for the most is structural (supply side caused, skill/education mismatch) or cyclical (demand side caused, lack of aggregate demand) no consensus has been reached, but many people tend to believe that the demand explanation is more likely (see below).
However, suppose for a second that supply side does explain a portion of the unemployment rate. Then it is probably those construction workers in California, Nevada etc. that do not want to move to the East are causing the large spread in the unemployment rate across states- most of the “structural” unemployment is usually attributed to the lost jobs in construction industry and financial services. I say “want” instead of “cannot” because if “structural unemployment” theory does explain high unemployment rate, it would be equally hard for those workers to find a job in the East. SO they just stay where they had their last job – in the states that experienced the housing boom and where the unemployment rate is the highest in the country. In this case the graph above might have nothing to do with interstate mobility, but just reflect the difference in “structural unemployment” rates.
Moving on to the second part of my post, Mike Konczal posted the following picture on mobility that comes from Census:
As Mike points out
The quick read was that everyone, post bubble-popping, was abandoning their properties and moving across the town to live with their parents and friends while looking for a new local job to save their house, a house they couldn’t sell because they were underwater.
However, as he later points out, the paper by Greg Kaplan from Minneapolis Fed Interstate Migration Has Fallen Less Than You Think: Consequences of Hot Deck Imputation in the Current Population Survey says that
…much of the recent reported decrease in interstate migration is a statistical artifact. Before 2006, the Census Bureau’s imputation procedure for dealing with missing data inflated the estimated interstate migration rate. An undocumented change in the procedure corrected the problem starting in 2006, thus reducing the estimated migration rate. The change in imputation procedures explains 90 percent of the reported decrease in interstate migration between 2005 and 2006, and 42 percent of the decrease between 2000 (the recent high-water mark) and 2010. After we remove the effect of the change in procedures, we find that the annual interstate migration rate follows a smooth downward trend from 1996 to 2010. Contrary to popular belief, the 2007–2009 recession is not associated with any additional decrease in interstate migration relative to trend.
Which probably means that housing boom and subsequent number of foreclosures has nothing to do with the decline in workers interstate mobility. Another piece of evidence is the paper by Colleen Donovan and Calvin Schnure “Locked in the House: Do Underwater Mortgages Reduce Labor Market Mobility?” and its summary that originally belongs to Kash Mansori. The main question was whether the fall in house prices since 2007 in the US created a lock-in effect that depressed labor mobility. The paper finds that even though
the evidence presented in the paper indicates that the fall in house prices has indeed caused a “lock-in” effect, but has not significantly impacted labor market efficiency.
and even more
The lock-in, however, results almost entirely from a decline in within-county moves. As local moves are generally within the same geographic job market, this decline is not likely to affect labor market matching. In contrast, moves out-of-state, which are more likely to be in response to new employment opportunities, show no decline, and in fact are higher in counties with greater house price declines
which together with the previous evidence reasserts the fact that the foreclosure rates did not significantly affect labor mobility.
[Update: Aug 20 1:30pm]
Kyle posted a reply to my post. However, there is no difference between Bob Hall’s graph (Figure 1) and the one I posted – the drop on both graphs happened in 2005 and is the size is the same – approximately 35%! I think there was no drop in labor market mobility due to the foreclosure issues. There was no drop at all! It is just an artifact in calculating procedure as I mentioned in the post and the paper I mentioned.
Clarification about my point about construction workers: I did not mean only construction workers, but all workers who are considered to be “structurally” unemployed: construction, financial services, etc. Even if those people do not contribute to the national unemployment rate, they might contribute to the variance of the unemployment rates across states – just because there are more construction workers in the states where was housing boom.