2.6 Summary
3.2.7 Distortions
Migration may not respond to economic growth if immigration is controlled. Wilkin- son (1970) analyzed European emigration to the United States in 1870-1914 econo- metrically using the levels of output in the U.S. and the origin country and the di↵erence in real wages as explanatory variables. Output was used over employment as the latter was considered to be endogenous. He augmented this basic model with labor supply variables namely lagged emigration rates, changes in the U.S. labor force, and U.S. immigration rates. He used autoregressive distributed lag models on top of the OLS. He found that U.S. output had a positive e↵ect on emigration from Denmark, Sweden, Russia and Italy, but with a one-year lag for the first two, and no e↵ect on emigration in Germany and the United Kingdom. However, con- trolling for labor supply variables, the e↵ect of U.S. output disappeared except for Sweden. The absence of e↵ect may be due to holding U.S. immigration rates fixed or to an inelastic labor supply. He concluded that pull factors seem to have little e↵ect. On the other hand, origin country income had a significant negative e↵ect on emigration for Denmark, Germany, Russia, Sweden and Italy even when controlling for labor supply variables. Where data were available, the wage di↵erential had a significant positive e↵ect on emigration. Lagged emigration rate had a significant, but surprisingly, negative e↵ect on current emigration rates for Denmark, United Kingdom and Sweden.
Migration may also be constrained by minimum wages. Puzzled with the ab- sence of either a pull or push e↵ect for the UK in Wilkinson (1970), Gallaway and Vedder(1971) modelled migration from UK to US (in di↵erences) on unemployment in the UK, economic conditions in the US as a proxy for unemployment, wages in the UK and US (in di↵erences), emigrations to alternative destinations Australia, South Africa and Canada, and dummies for cyclical shocks. They found unemployment in the UK and economic conditions in the US to be significant, wages in the UK to be weakly significant, and wages in the US not to be significant. The insignificance of wages may be due to market distortions such as minimum wages. Emigrations to other destinations were not significant but economic shocks were significant. Con- sidering only the significant variables in a step-wise regression, they found that US economic conditions had the strongest influence, followed by UK unemployment, UK wages and cyclical shocks. They argued that ‘pull’ factors were more impor- tant, accounting for 40 percent of the variation in emigration from the UK to the US, but that ‘push’ factors were significant as well.
Migration can decrease with greater labor supply in the origin country if employment increases. Modeling labor demands in the U.S. and origin countries (Sweden, United Kingdom, Denmark and Germany) as functions of their respective wages, output, and labor supplies (native labor force less (plus) net migration in the origin (destination) country), Williamson (1974) derived a reduced form equa- tion relating migration on the labor force and outputs in the origin and destination countries. Interestingly, the reduced form equation did not include wage variables as including these apparently misspecified the model, especially when lagged popula- tion growth was included. Wage would be capturing the e↵ect of lagged population growth on migration. A higher coefficient on U.S. output relative to that of the origin country meant stronger pull e↵ects. This can be due to relatively lower wage elasticities in the U.S., higher output elasticities in the U.S. and higher capacity growth in the U.S.
The theoretical model was modified in the econometric estimation to account for immigration to the US from other countries and the e↵ect of changes in em- ployment conditions. The significance and direction of pull factors are consistent across four countries. Despite low explanatory power and significance, Williamson found that migration in all four countries was directly related to U.S. output and negatively related to U.S. labor force, as expected. The direction of push factors is less consistent. Migration was negatively related to outputs in Denmark and Sweden and positively related to labor force in these countries, also as expected. However,
the coefficients for output and labor force in UK and Germany had the wrong signs. This can be explained by the fact that an increase in origin labor supply holding income fixed would increase employment, thereby decreasing migration. The under- lying coefficients confirm a positive e↵ect of U.S. output on U.S. labor demand and a positive e↵ect of wage di↵erentials on migration. However, contrary to expecta- tion, labor demand was positively related to wages in the UK and Germany, and negatively related to output in Germany.
Minimum wages would decrease migration. McKenzie et al. (2014) found that Filipino migration to various destinations is significantly positively related to desti- nation incomes. However, migrant wages are not significantly related to destination incomes. They attribute this to binding minimum wages as the Philippine govern- ment requires that overseas Filipino workers’ contracts guarantee wages not lower than the minimum in the Philippines, host country or that set in bilateral agree- ments or international conventions. This has increased Filipino wages in destinations but decreased Filipino migration to low wage destinations. For domestic helpers, migration to destinations with minimum wages decreased by 55 percent compared to those with no minimum wages. Compared to other occupations, the migration of domestic helpers decreased by 57 percent.
Restrictive polices reduce the e↵ects of push and pull factors. Mayda (2010) analyzed migration to 14 OECD countries and found emigration rate rose by 20 percent for every 10 percent increase in (log) GDP per worker in destination coun- tries. However, the income level in the home country did not a↵ect the emigration rate. Distance between the destination and origin countries was highly negatively related with emigration rate. On the other hand, having a common land border, a common language, and colonial link did not a↵ect emigration rate. The share of the young in the population positively a↵ected emigration rate. The average education in the origin country positively a↵ected emigration rate, while that in the desti- nation country negatively a↵ected emigration rate. The foregoing push and pull factors were in turn a↵ected by migration policies in destination countries. Less re- strictive policies reinforced the expected e↵ects of push and pull factors, making the income level in the home country significant. Moreover, Mayda (2010) found that a rise in the relative inequality between origin and destination countries increased emigration. On the other hand, unemployment rates in the destination and origin countries were insignificant.