Vernazza, Daniel
(2012)
Essays on the causes of migration.
PhD thesis, London School of Economics and Political Science.
Abstract
This thesis consists of three chapters. All three are linked by our desire to better understand the determinants of labour migration; that is, the motivation for a person to change his or her location of residence for a period of at least a year. While immigration receives much public discourse, the economic evidence on how migrants self-select is still lacking. In particular, we have little evidence on
the relative importance of determinants. We focus on three areas that have received substantially less attention in the migration literature: the importance of relative versus absolute income motives for migration; the effect of wealth and intertemporal choice on return migration; and the role of place attachment as an obstacle to labour mobility. Common to all three chapters is an emphasis on counterbalancing forces that tend to offset spatial income differentials in determining migration.
The first chapter examines the extent to which relative income – that is, one’s position in the income distribution – matters in migration choice. Virtually all studies of migration focus on absolute income. This is at odds with the mounting evidence that suggests people care about their relative position in the income distribution. We argue that, in order to test between the absolute income and relative income theories of migration, one needs individual-level panel data on before and after migration outcomes. Indeed, since one has to estimate counterfactual migrant earnings of non-migrants, if migrants are selected on unobservables then cross-sectional estimates will systematically bias the predicted migrant earnings of non-migrants. We estimate the relative importance of the two main theories in explaining interstate migration in the U.S. using a panel of individuals. Relative income is calculated with respect to those persons in the same U.S. state. We find that, although migration leads to a substantial rise in absolute income, the trigger for migration is low relative income and not low
absolute income.
In the second chapter we show analytically that, under some conditions, return migration is optimal. We build a model where consumers choose either to never migrate, permanently migrate or, migrate and subsequently return. To generate an incentive for return migration, the model assumes a nominal income differential between the source and destination and a compensating differential – which exerts a counterbalancing force to the income differential. Examples of compensating differentials may include differences between the source and destination in climate, place attachment, price levels, unemployment and average consumption. We characterise the optimal migration decision space with respect to the three key variables: initial wealth, the income differential and the compensating differential between the source and destination. The marginal utility of consumption is assumed to be location-dependent due to a non-separable nonpecuniary preference for the source. Consequently, when the region with the best economic opportunities is not the source region, there is a trade-off between income maximisation on the one hand and the marginal utility of consumption
on the other. We find that, all else equal, those with low wealth are more likely to migrate and, conditional on migration, those with higher wealth are more likely to return migrate.
The third chapter seeks to estimate a key obstacle to migration: place attachment. Place attachment refers to the emotional bonds a person feels towards the place (or area) he or she resides. We estimate place attachment within a tructural model of spatial job search where migration is a by-product of accepting a job offer from another region. The chapter can broadly be split into two parts. The first takes a standard job search model and adapts it to allow search in many potential destinations. Acceptance of an offer from a destination necessarily involves migration to that destination and its associated costs. We consider two
types of costs: a cost of migration that is related to distance-to-destination and a non-pecuniary cost of leaving the current region. The latter is deemed to be the negative of place attachment. In the second part, we estimate the structural model for a sample of individual durations in a U.S. state. Our estimates suggest that place attachment is steeply increasing in duration for our reduced-form model; however, the opposite is true for our structural model. We also find that for half the population, the dollar values of place attachment are prohibitively large.
Actions (login required)
 |
Record administration - authorised staff only |
Download statistics
Download statistics