Essays on trade and migration

Essay I. Complementary Links between Bilateral Trade and Migration: Evidence from U.S. Flows, 1970-2002. The theoretical relationship between trade and migration is ambiguous. Using data on bilateral trade and immigrant stock, a series of new studies have provided evidence that migration promotes trade, i.e., trade and migration are complements, in the U.S. and in a few high-income economies. This paper uses a rich, new panel dataset on U.S. bilateral trade covering U.S. trade with up to 116 countries over a 33-year period, representing a significant expansion in country and period coverage relative to existing studies. Trade data were matched with data on migrant stock by country of origin over the same period, a database we constructed using decennial census data and Immigration and Naturalization Service (INS) data. We employ a variety of techniques for reexamining the trade-migration nexus. There is robust empirical evidence that migration promotes trade, and thus U.S. trade and migration flows into the U.S. are complements. An important new finding is the time-varying link between trade and migration flows, i.e., there is strong evidence that trade and migration flows are complements in recent years and some evidence that they may have been substitutes in earlier periods. The results are largely invariant to the choice of estimation procedures, data transformations, sources of immigrant stock data, the use of nominal and real values of trade and economic mass, and the inclusion of controls for tariff and non-tariff barriers. There is also evidence that the "migrant network elasticity of trade" (or the responsiveness of the quantity of U.S. trade to change in the stock of migrants in the U.S.) varies across types of trade commodities; in particular, the highest elasticities are found among manufactured goods. Finally, there is some evidence that immigrant skill is positively related to trade.;Essay II. Individual Preferences for Trade and Migration: Cross-Country Evidence. This paper reexamines the empirical literature on policy preferences for trade and migration, drawing from trade theory to shed light on the individual choices for or against trade and for or against migration. It also offers new insights on an enduring "policy paradox" in the preferences for trade and migration, namely, why some countries are more pro-trade than pro-migration. A major contribution of this paper to the literature is the systematic use of simultaneous equation techniques to model the joint determination of policy preferences for trade and migration. In addition, by recoding the survey responses on preferences to form multiple combinations of trade and migration preference, this paper is able to utilize, for the first time in this literature, a Multinomial Logit model of trade and migration policy preferences. More important, the paper provides an empirical framework for assessing the role of a generous welfare state in the observed asymmetric preferences for trade over migration (Wellisch and Walz, 1998 and 1999; Hatton and Williamson, 2005). Two principal findings emerge from the analysis: First, the paper finds robust evidence in support of trade theoretic predictions on trade and migration preferences and confirm that policy preferences for migration and trade, which are typically modeled independently, are in fact jointly determined. Second, the paper provides compelling evidence suggesting that the existence of a generous welfare state is associated with asymmetric preferences for trade over migration, consistent with existing theoretical predictions that relate the net welfare burden of migration to restrictive policy preferences. This finding is robust to various specification, estimation techniques, and invariant to the inclusion of a known determinant of welfare generosity, the degree of ethnic fragmentation.;Essay III. The Self-Employment Experience of Immigrants: Evidence from the U.S. General Social Survey, 1977-2004. This paper uses biennial data from the U.S. General Social Survey (1977-2004) to reexamine the self-employment experience of immigrants, yielding a large database covering 20 individual years with over 36,000 observations depending on the specification. The regression results strongly suggest that the conditional probability of self-employment is higher among immigrants. Self-employment is also significantly related to age (nonlinearly), educational attainment, race, marital status, occupation, industry of employment, and family background. However these factors are generally less important in explaining self-employment among immigrants. In particular, some variables such as marital status or homeownership (a proxy for access to capital, an important determinant of entrepreneurship) are significantly associated with native self-employment, but not migrant self-employment. We also look at latent entrepreneurship or preferences for self-employment and find no evidence that preferences vary by immigrant status. Taken all together, we speculate that, consistent with the previous literature in economics and sociology, immigrants are self-selected into self-employment, independent of their individual characteristics, family background (father's self-employment), average self-employment in the country of origin, and stated preferences.