Would more migration in Europe be a good thing? Among European policy makers the answer seems to be both yes and no at the same time. Many observers believe that, within the EU, there is too little migration between member countries for a monetary union to work smoothly, and to allow adjustment to region-specific shocks. But at the same time, Europe's political leaders are increasingly concerned about an influx of migrants from Eastern Europe and North Africa.

In both cases, the favoured policy response is 'convergence', but of quite different sorts. Within the Union, it is hoped that harmonizing labour regulations, social security procedures and transport networks will make cross-border migration easier. Convergence means something else outside the EU: one motive behind the Union's policies towards Central and Eastern Europe is to speed income convergence, in order to stem the flow of migrant workers into the EU.

Conventional models of migration suggest that boosting growth in the sending countries is the correct strategy for stemming the tide of migrants into the EU. These conventional models, based on work by Harris and Todaro, assume that the choice to migrate depends on a comparison between income at home and in the host country, weighted by the probability of finding a job in either. So to the extent that aid and trade polices boost incomes and job prospects in countries outside the EU, they should encourage more people to stay at home.

The problem with using the Harris-Todaro model as the basis for policy-making is that it does such a poor job of explaining recent European migration. The model predicts, for example, that migration will come predominantly from very poor countries, but Southern Europe has been a much more important source of migratory flows to Northern Europe than the much poorer countries of North Africa.

Even within Europe, migration between European countries has remained surprisingly low despite substantial and persistent gaps in per capita income between south and north. Indeed, migration flows from south to north have fallen dramatically since the first oil shock in 1973, even though there has been very little convergence of per capita incomes between the two regions.

One possible explanation for this decline in migration is, of course, the fall in labour demand in the receiving countries following the oil shock. But unemployment rose sharply in south European countries as well. And even during the 1980s, when the economies of northern Europe recovered markedly, migrations from southern Europe did not return to their previous levels.

What is the conventional model missing? According to recent research by Riccardo Faini and Alessandra Venturini, the model ignores one simple but essential factor: people prefer to live in their own countries for social, cultural or psychological reasons. In very poor countries, the prospect of riches elsewhere will probably outweigh this 'home-bias'. But the richer the country, the more its citizens will be locked into more domestic consumption – both of goods and of culture – and so less inclined to move.

Faini and Venturini find the home-bias effect complicates the relationship between growth and migration. Home-bias means that rising income in the home country will discourage migration, even if wage or income differentials with respect to other countries remain unchanged – precisely what has happened in southern Europe. But if the home country is relatively poor, an increase in income may have the opposite effect on migration. Potential migrants may be unable to move abroad because of monetary constraints, but if their incomes rise, such constraints would become less important. Higher incomes may thus encourage migration flows in poor countries.

If this analysis is correct, the relationship between migration and income is likely to be hump-shaped – as income grows, migration first rises as the financial constraint is overcome but then falls as the home-bias effect takes over. And this is precisely what the authors find in their empirical analysis. They estimate the relationship between migration and income in Greece, Spain, Portugal and Turkey over the period 1962-88. The link between migration and income is indeed hump-shaped, with the turning point at incomes of around $4000 (in 1985 prices).

Europe's policy-makers will find little comfort in this analysis. Within Europe, Faini and Venturini argue, the model suggests that there is little prospect of an increase in cross-country migration. As long as income continues to grow and national, cultural and social differences persist, the home-bias effect will discourage migration. Faini and Venturini's analysis suggests that if more internal migration were needed for a monetary union to work effectively, then policy-makers need to focus on non-economic notions of 'convergence' such as language and schooling rather than on inflation or even growth rates. There is a better chance of stemming the tide of external migration into the Union by boosting incomes in the sending countries through trade and aid policies, at least in the long-run. But in the meantime, EU policies may well have the opposite effect if the home countries are still on the rising section of the hump. This is quite possible since most East European and North African countries have incomes per head well below the $4000 turning point. Not that aid and development policies should be discouraged, the authors are quick to stress. But trying to sell these policies on the basis of their short-run impact on migration does not seem to fit the facts.