بازگشت مهاجرت، انباشت سرمایه انسانی و فرار مغزها

In this paper we present a model that explains migrations as decisions that respond to where human capital can be acquired more efficiently, and where the return to human capital is highest. The basic framework is a dynamic Roy model in which a worker possesses two distinct skills that can be augmented by learning by doing. There are different implicit prices, in different countries and different rates of skill accumulation. Our analysis contributes to the literature on the selection of immigrants and return migrants by offering a richer framework that may help to accommodate selection of emigrants and return migrants that are not immediately compatible with the one-dimensional skill model. Our analysis also has implications for the debate on brain drain and brain gain. In the two skills model presented here, return migration can lead to a mitigation of the brain drain, or even the creation of a “brain gain”, where those who return bring the home country augmented local skills.

Mobility of workers across national borders responds not only to the return to skills, but also to the opportunity and efficiency of skill acquisition. Efficiency considerations suggest that skills should be acquired where the cost is low and applied where the reward is high. This last aspect has been largely overlooked in the literature that analyzes the causes and forms of migration. Thus, individuals may choose to acquire skills abroad that are highly rewarded in their home country and produced cheaply elsewhere. Student migrations are an example with some countries having established themselves as learning centers that provide educational services above those demanded domestically.2 There is evidence that, for migrants who returned to their home country, work experience acquired abroad enhances earnings by more than work experience acquired in the home country. Reinhold and Thom (2009) analyze earnings of Mexican emigrants who returned from the U.S. They find that, for these immigrants, the labor market experience accumulated in the US increases earnings by twice the amount than experience accumulated in Mexico. Papers by Barrett and O'Connell, 2001, Barrett and Goggin, 2010 and Iara, 2006 report similar findings for Ireland and for migrants who returned to Eastern Europe from Western European countries. Co et al. (2000) report a wage premium for having been abroad for female return migrants to Hungary.3 In this paper, we present a model that explains migrations as decisions that respond to where human capital can be acquired more efficiently and where the return to human capital is highest. The basic framework is one in which a worker possesses two distinct skills that can be augmented by learning by doing while acquiring work experience. The two skills command a different implicit price in different countries. The rate of human capital accumulation is also different in different countries. Thus, a person may move to a country where her skills grow fast and then apply these skills in a different country where these skills have a high price. In this regard, there is an important difference between human and physical or financial assets. Human capital cannot be separated from its owner and he/she must move in order to exploit differences in returns in different locations. An early paper that discusses higher return in the home country to skills acquired in the host country as a motive that triggers return migration is Dustmann, 1994 and Dustmann, 1995. Other papers that analyze this motive are Borjas and Bratsberg, 1996, Santos and Postel-Vinay, 2003, De Coulon and Piracha, 2005 and Mayr and Peri, 2008. These models assume that individual skills are one-dimensional. In the single skill model, individuals move based on the prices of this skill in the two countries. If the price is higher in the receiving country some highly skilled workers will move. If the possibility to learn abroad is added, some of those who moved will return but those will be the least skilled among the emigrants. Conversely, if the price of the single skill is lower abroad, low skilled workers will emigrate and among these immigrants the most skilled will return.4 Considering two skills and allowing comparative advantage to play a role, we obtain “non-hierarchical” migration and remigration patterns with movements that are neither positively nor negatively selected. Among the stayers in the home country, there are some who are more able (in the sense of having a larger endowment of both skills) than some of the movers. At the same time, there may be some movers who are more skilled than some of the stayers. In both comparisons, those who stay have a relatively high component of the skill that is more highly valued in the home country and those who move have a relatively high component of the skill that is more highly valued in the host country. By the same logic, the selection of return migrants may exacerbate or alleviate the impact of migrant selection for the initial out-migration for both emigration- and immigration country. In these regards, the multi-dimensional skill distribution yields a richer set of testable implications than the one skill model of Borjas and Bratsberg (1996). Our model has important implications for the debate on brain drain and brain gain. In an early paper, Kwok and Leland (1982) describe brain drain as a (permanent) outflow of skilled workers. The model discussed by Borjas and Bratsberg (1996) adds an additional dimension to this: A brain drain issue arises when the price of skills is higher abroad, and may be amplified by those who return being the less able among those who left. In the two skills model presented here, the brain drain is mitigated because those who return come with augmented local skills that are more applicable in the home country. If the proportion of those who return is large enough, aggregate output and even output per capita may increase, implying a brain gain. We also show that by imposing entry standards based on skills that are tailored to the host country the potential brain gain is reduced, because some of those who would return with augmented local skills are barred from skill acquisition abroad.5 We discuss these issues in the context of a dynamic Roy model, in which skills vary over time. In contrast to the static Roy (1951) model, in which alternatives are characterized by the prices of skills only, our model specifies each alternative in terms of its price and learning opportunities. Such a model can generate planned mobility even under conditions of certainty.6 This richer framework is suited to explain migration and remigration between countries with different technologies and a different industrial structure. In particular, it can explain immigration patterns between developed and developing countries that are incompatible with the one skill model that is often applied empirically.7 We conduct the analysis in three steps, working backwards. We first examine the return decision of immigrants who are already in the receiving country and investigate who shall return to the home country and when. Based on the results of this last stage, we examine the timing of emigration from the home country. Based on these two considerations, we finally discuss who shall emigrate. We show that these decisions depend crucially on the extent of transferability of work experience acquired abroad to the home country. Specifically, if one can augment the skills that are highly valued in the home country more efficiently abroad, it motivates both emigration and return migration. We then discuss the potential brain gain associated with return migration.

We have presented a tractable model that focuses on the incentives to return immigration based on investment considerations.24 It was shown that, under some conditions, the model can generate a brain gain. The basic idea is that some countries are learning centers where one can learn skills more effectively, including skills that are applicable to the home country. Therefore, some individuals who emigrate will return to apply their acquired skills in the home country. Moreover, those who return have a relatively high endowment of the skill that is more valued in the home country. We have extended the work of Borjas (1987) by adding learning considerations to the Roy model. This richer framework has the potential to explain migration and remigration patterns between developed and less developed countries that differ in technology and industrial structure. An empirical application of such a model requires measures of the various dimensions of skills. Schooling alone is insufficient for this purpose but one can add the “task content” of occupations as in Autor and Handel (2009). We discussed here only learning by doing on the job. However, the basic ideas also apply to learning at school. A policy issue that applies to both cases is how to allocate the gains from immigration among individuals and in between countries. Tuition policies in the context of student migration are discussed in Rosenzweig, 2008 and Kennan, 2009. This problem is somewhat more complicated in the case of learning on the job, where the opportunity costs are not directly observable. Further issues arise if one can move to a learning center in order to acquire skills that are applicable in a third country rather than the home country, resulting in chain migration. In this case, the natural solution would be to require a payment from the immigrant for the “general” human capital that she acquired. Finally, an important issue that we did not discuss is the potential externalities if the skills of different workers in a given economy are complements. Hence, brain drain or gain can have magnified consequences through the impacts of emigration (immigration) on the workers who stay in the home country and also further consequences for the workers in the receiving country.