A high degree of mobility has long been considered a distinguishing feature of the U.S. labor market.
A commonly-held view is that when a U.S. state is experiencing tough times (relative to other U.S. states), workers quickly leave the state for greener pastures; this keeps the state’s unemployment rate from going up too much and its labor force participation rate from declining too much.
My new work (with Mai Dao and Davide Furceri) offers a less sanguine view of the ability of U.S. workers to shield themselves from the consequences of adverse shocks. We show that, particularly in the short run, the adjustment to tough times occurs more through unemployment rates going up than through people leaving the state. And while migration picks up during recessions, people in the states that are doing very poorly have a difficult time exiting.
Here is a link to the paper and a technical summary of the paper:
Our first key finding is that labor mobility is less important as a cyclical adjustment mechanism, relative to changes in unemployment and participation, than suggested in earlier work. Some of this shift in view comes from the addition of over 20 years of data to the previous work. But the main reason is that, given the availability of official interstate net-migration data starting in 1991 we can also directly look at the behavior of migration, as opposed to backing it out as a residual. We find that it is primarily the relative unemployment rate, not net migration, that is the main adjustment mechanism in the first two years following a relative shock to state labor demand.
Our second set of findings pertains to a newer literature that documents longer-run movements in U.S. mobility, particularly the steady and widespread reduction in gross internal migration rates since the 1980’s. Here we establish several results that reveal important patterns in regional adjustment mechanisms.
- First, in the last two decades or so starting 1990, the response of net migration to given regional shocks in the short run has decreased, as has the response of relative unemployment and participation rates, resulting in less dispersion of employment growth in response to given dispersion in relative labor demand shocks.
- Second, the smaller migration response to shocks is driven entirely by less net out-migration from states that experience adverse labor demand shifts, whereas the net in-migration response to states with favorable labor demand shifts has increased or remained constant (depending on time horizon). This also suggests that in-migrants to the best states do not disproportionately come from the poorest states, a sign of lack of migration directedness and of scope for efficiency gains from an aggregate perspective.
- Third, despite the trend decline in gross migration rates since the early 1990’s, the migration response to a state-relative demand shock increases strongly in recessions, hence potentially playing a larger role as shock absorber during aggregate downturns than in normal times. Importantly, we observe that this counter-cyclical response of migration is driven primarily by a stronger response of positive net migration into states that do relatively better during recessions, while negative net migration from states that do relatively worse only increases by less and the response is delayed, occurring toward the end of the recession. When a state like North Dakota does relatively better than average during a recession thanks to the shale gas boom, it attracts disproportionately more in-migration than for instance Texas during an expansion, when strong demand for oil creates more jobs in Texas than elsewhere. However, the migrants into North Dakota during the recession do not come disproportionately more from states that are doing worse than average, say Michigan, as one would expect.