There has been periodic speculation that one factor in the continuing high unemployment levels is “house lock,” or the reluctance of households to sell their homes in a declining price environment.  This, the theory goes, may create a geographic mismatch between the locations of available workers and available job openings. 

If this is true then it follows that household migration should be relatively greater among renters than owners in the current market and should be higher in areas where home prices have been more stable compared to areas that have suffered large declines in home values.  It also should follow that migration would be higher in households – whether renters or owners – where the head is unemployed.

The Federal Reserve Bank of Chicago recently completed a study that found that none of these conditions apply to the current recession and concluded that “there is little empirical evidence that house lock has been an important driver of the recent high unemployment rate.”

The study, conducted by Daniel Aaronson, the bank’s vice president and economic advisor and Jonathan Davis, an associate economist, used the U.S. Census Bureau’s Survey of Income and Program Participation (SIPP) as the basis for their study.  SIPP is a large representative sample of non-military households which is interviewed every four months (called a wave) for two to four years.  The surveys overlap (one group is interviewed in January, another in February, etc.) which allows for a month-by-month analysis of migration rates, however, when a cohort’s two to four year tenure is ended it is replaced by another group as the next wave begins.  This results in occasional four month gaps in the data.  The SIPP data only allowed for evaluation of state-to-state migration.

In a given year, fewer than 2 percent of all SIPP households cross a state border, but migration is three to four times more common for renter households than for homeowners.  This ratio has held throughout the sample period, therefore over the past 25 years a significant portion of geographic relocation has been among households unencumbered by owning a house.

Aaronson and Davis compared the average four-month state-to-state migration rates during the last three years of the economic expansion (2005-2007) to  the slightly shorter December 2008 to July 2010 period coinciding with the worst of the recession, and found that homeowner migration rates from 0.0025 during the2005-07 period to 0.0019 during the December 2008-July 2010 period, about 0.02 percent.  But renter migration rates dropped as well and did so roughly in tandem.  Furthermore, the results are very similar when the recent period is compared to rates through the entire 2002-2007 economic expansion and when compared to the differences between earlier periods of growth and subsequent recessions In fact, the authors say, “Given the extent of the downturn in 2009-09, the decline in homeowner and renter mobility was rather tame this time around.”

House lock was no more of a factor in those states suffering the most severe housing busts.  In fact, during 2009 and early 2010 homeowner state-to-state mobility rates decreased more for households in states that experienced smaller price declines.  Even in the five states with the largest drop in house prices there was no evidence that homeowners were migrating out at a historically unusual rate.

Finally, there was no evidence found that households where the head was out of work were more likely to move, a fact that held true again for both home owners and renters.

The authors put forth two caveats for their findings.  The study as mentioned earlier was constrained to looking at state to state migration but a limited analysis of in-state migration data, especially in larger states where there may be more than one distinct labor market,  indicates that homeowner in-state migration also fell during the 2009-2010 period  but there are indications that renter migration may have fallen less.  The second caveat is that the study was conducted during a period when unemployment was at 9.5 percent so the lack of jobs may have contributed to the lack of mobility rather than the reverse.  It could be that once the demand for labor picks up any geographic mismatch will become more apparent.