A joint CEPR workshop with the Innocenzo Gasparini Institute for Economic Research (IGIER) on 'Migration, Trade and the Impact on the Sending Country' was held in Milan on 1/2 December 1995. The workshop was organized by Riccardo Faini (Universita degli Studi di Brescia, IGIER and CEPR) and Cormac O'Grada (University College Dublin and CEPR). The workshop was organized within the framework of CEPR's Migration Network, supported by a grant from the European Commission under its Human Capital and Mobility Programme. Additional funding was provided by the Centro Studi Luca d'Agliano.
In 'Wage and Mobility Effects of Trade and Migration', jointly written with Klaus F Zimmermann (SELAPO, Universität München and CEPR), John P Haisken-De New ( SELAPO, Universität München) investigates the relative importance of trade and immigration for earnings and job mobility of German male workers. Using panel data, changes of workplace within the firm as between the firms are separated from occupational changes. The authors find that trade matters more than migration. Wages are affected negatively by a relative increase in imports, and positively by immigration. Trade seems to depress occupational mobility and internal movement, but stimulates inter-firm changes. Immigration affects intra-firm changes negatively, but is largely unrelated with other aspects of labour mobility.
Giorgio Basevi (Universita di Bologna) questioned the choice of a stock variable capturing the impact of immigration: doesn't the immigration flow matter more? Because of the endogeneity of immigration in the model, the choice of appropriate instruments becomes important. Timothy Hatton (University of Essex and CEPR) suggested that these should be chosen following the logic of a migration equation. Riccardo Faini argued that the results on the influence of trade and migration are difficult to reconcile from a theoretical viewpoint. Jaime de Melo (Université de Genève and CEPR) suggested that the influence of trade on wages could be estimated by separate sectoral equations rather than using an aggregate variable (i.e. the trade deficit).
How can the consequences of immigration for economic welfare in the destination country be estimated properly? In 'Modelling the Effects of Migration on the Destination Country: Some Simulations for Switzerland' Tobias Muller (Université de Genève) addresses this issue by using a computable general equilibrium model for Switzerland. He discusses the role of different modelling assumptions concerning factor markets and foreign trade. Welfare effects of immigration are negative only if the price elasticity of export demand is assumed to be relatively small, so that immigration produces an (implausibly) large terms-of-trade deterioration. Conversely, welfare gains are larger if immigrant labour is an imperfect substitute for native labour. However, these gains are substantially weakened with international capital mobility.
Ian Wooton (University of Glasgow and CEPR) pointed out that this paper is an exercise in numerical simulation, in that the empirical model is really being employed to examine the implications of adopting different assumptions about the structure of factor markets. The model is not being used to examine the effects of migration on the Swiss economy. He then gave an explanation of why the results differ crucially between the case where foreign and native labour are differentiated and the homogeneous labour case. With homogeneous labour the gains arise from two sources: as immigration increases, the value marginal product of labour falls yielding infra-marginal gains accruing to the owners of capital and also resulting in a redistribution of income from workers to the owners of capital. To the degree that the Swiss own relatively more capital than the immigrants, this is a redistribution in favour of Swiss households. When labour is heterogeneous there are additional gains possible from the immigration. If Swiss workers are complements in production for immigrant workers, further immigration raises the value marginal productivity of the Swiss workers to their benefit.
In 'Migration as Disaster Relief: Lessons from the Great Irish Famine' Cormac O'Grada and Kevin O'Rourke (University College Dublin and CEPR) re-examine the historical record of the Irish famine. In the 1840s many Irish had the possibility to emigrate, an option not available to today's famine victims. Emigration differs from other types of disaster relief in at least two ways. First, poverty traps imply that it cannot help those who are most in need. Cross-country data clearly point to the existence of such a poverty trap during the Irish famine. Second, food relief is temporary: when the crisis has passed, the aid ceases. By contrast, Irish famine emigrants did not return. The Irish famine thus differed from more recent famines in that it had permanent demographic consequences.
Timothy J Hatton welcomed the detailed quantitative analysis of this historical episode. He stressed the three main points of the article: the authors give a more Malthusian interpretation of the famine than past work did, they discuss the issue of 'emigrant quality', and they show clearly that the famine had long-lasting consequences on population growth in Ireland. However, he questioned the authors' conclusion on the existence of a poverty constraint to emigration, since many boat tickets were sent over by relatives who had already emigrated to North America. And why didn't the poor emigrate to England, which would have been a cheaper option?
What is the role of emigration in the transition from a closed communist system to an open market economy? Peter Sanfey (University of Kent) presented 'Job Creation, Job Destruction and Temporary Emigration: the Albanian Experience', written jointly with Ahmet Mancellari (University of Tirana), and Harry Papapanagos (University of Kent). In this paper, the authors analyse the recent experience of Albania, where the fall in production and aggregate employment since 1989 has been accompanied by large scale, and mostly temporary, emigration. They examine the effect of emigration on job flows and unemployment during the transition. The authors demonstrate that the rapid growth in emigration has been associated recently with increased job creation and employment growth in the private sector, suggesting that remittances are playing a key role in the restructuring of the economy.
Alessandra Venturini (University of Bergamo) appreciated this paper because, on the one hand, it provides information on the economy of a very little known European country and, on the other hand, it highlights the positive short run effect of temporary emigration (through remittances and a relaxed labour market). However, it neglects the potential negative impact of emigration in the long run: permanent emigrants diminish their remittances over time and represent a loss of human capital for the country of origin, reducing thereby long run growth. This issue could only be analysed in a dynamic, forward-looking model.
It is generally agreed in the literature that immigrants to the US have assimilated quite well in the twentieth century: receiving initially lower wages than natives, they catch up after a few years. Examining data for the late nineteenth century, some authors have found very different results. Contrary to conventional wisdom, they suggest that pre-1890 immigrants did not assimilate well into the American labour market because their earnings grew relatively slowly and did not converge towards native earnings patterns. In 'The Immigrant Assimilation Puzzle in Late 19th Century America' Timothy Hatton re-examines these results and shows that the nineteenth-century 'assimilation puzzle' disappears when the earnings function is specified correctly. In his estimations, he finds earning profiles that are more consistent with conventional wisdom and with the results for post second World War immigrants.
Kevin O'Rourke observed that this was a convincing attack on the historical literature on immigrant assimilation, demonstrating that the results are crucially sensitive to the assumption of quadratic earnings functions. However, Hatton's own functional specification also seems arbitrary, which makes one wonder how robust his results are. In addition, all historical studies (including this one) suffer from problems not besetting the literature on immigration after 1945: they only incorporate data on particular industries in particular states. This makes the problem associated with inferring temporal patterns from cross-section data more acute. O'Rourke concluded therefore that the recent work by Joe Ferris, who traced individual immigrants through subsequent censuses and found that they did assimilate, is unlikely to be challenged by any cross-section study.
Recently the 'new migration theory' has emphasized that migration can occur as a response to risk attitudes, even in the absence of significant wage or unemployment differentials. In 'Where do Migrants Go ? Risk-Aversion, Mobility Costs and the Locational Choice of Migrants' Francesco Daveri (University of Brescia and IGIER) and Riccardo Faini argue that the link between inter-regional and international migration can provide an interesting testing ground for this theory. Whereas most of the literature analyzes the determinants of migration (whether to migrate), their paper is focused on the choice of destination (where to migrate). If incomes in different locations are imperfectly correlated, the possibility of emigration gives households scope for risk diversification. If the country of origin has a poorly diversified economy, inter-regional migration is ineffective in achieving this purpose and migration flows are expected to be biased towards foreign destinations. The authors test the predictions of their model with data on migration outflows from eight regions in southern Italy. Using measures of labour income riskiness at home and abroad, they conclude that the desire to diversify risk plays a substantial role in the migration decision.
Jean-Marie Grether (Université de Genève) pointed out that the possibility of risk diversification inside the family is simply assumed away: because of the assumption of concave mobility costs, all members of the family will relocate in the same place. The effective diversification only appears at the aggregate level, as households differ in terms of location-specific costs. Efforts to model risk diversification within a single household could therefore be worthwhile. Timothy Hatton suggested that the option value of migration could be introduced in a forward-looking setting.