Syndicate content

Migration as Structural Transformation

Shanta Devarajan's picture
IN024S05 World Bank When a poor person moves from a low-productivity job to a higher-productivity one, we usually celebrate.  The worker is clearly better off; the hiring firm is no worse off; and it’s good for the economy as a whole.  Indeed, development is often described as the process of structural transformation, where low-productivity workers (typically in agriculture) move to higher-productivity jobs in manufacturing or services. 

But when that same worker happens to cross a national border, we call it “migration” and, instead of celebrating, we start investigating the effects on workers, firms and public finances in the new environment; and on those left behind (the so-called “brain drain”).  Instead of promoting structural transformation, we look for policies to manage it.

Even more troubling, many of the statements made about the effects of migration on the receiving and sending countries are based on rhetoric rather than evidence.  Thanks to painstaking data work and rigorous research by Michael Clemens, Lant Pritchett and my colleagues Manjula Luthria, Caglar Ozden, and Dilip Ratha, among others, we now have a fairly robust picture of the effects of migration:

  • Not only do migration patterns follow wage differentials, but those differences are huge and growing over time.  A worker from Haiti moving to the U.S., or from Ethiopia moving to Italy, could increase his or her earnings seven-fold.  No other development intervention can match this 700 percent gain. Nevertheless, the stock of migrants worldwide has grown only slightly faster than world population growth.
  • The gains to the global economy of allowing labor to move to where it is most productive are truly massive—of the order of 100 percent of global GDP.  By contrast, freeing up all restrictions to trade will generate gains of the order of 2-3 percent of global GDP.
  • The effects of migration on the receiving country are positive.  While some native workers (who are easily substitutable for migrants) may see their wages decline, most of the economy will see an increase in wages and employment, as migrants perform jobs—such as childcare services—that permit the economy to be more productive (for instance, by allowing both spouses to work outside the home).
  • Across OECD countries, the fiscal burden of migration is either zero or slightly positive, partly because migrants tend to be younger and hence less of a drain on the health-care system.
  • The effects on the sending country vary depending on whether the migrant is unskilled or highly skilled.  If the former, there is generally a gain, since there is a glut of unskilled labor in poor countries and a shortage in rich ones.  For skilled workers, there is some evidence that wages of the remaining workers goes down, but the effects are too small to justify the hype about the brain drain.  The exception may be very small countries, such as those in the Pacific Islands or some in Africa.  But even here, there is the possibility of migration “chains”, where skilled people from poorer countries fill the jobs vacated by those leaving to OECD countries.  Furthermore, there is growing evidence of a “brain gain,” as the possibility of migrating abroad prompts more people to seek higher education in their own countries.
  • Independent of the empirical magnitudes, there is an argument against skilled migration that goes as follows:  since developing country governments paid for these people’s education, they—and not the receiving country—should get the benefits of their post-graduation services.  This argument is fragile.  First, it applies equally to those who join the private sector in their own countries.  Secondly, if a skilled worker such as a doctor is offered a high-paying position in an OECD country, he or she is part of an internationally mobile workforce.  If the source country wants to keep them, they should pay them internationally comparable salaries.  The reason they don’t is political:  they would have to raise the salaries of all the doctors in the country, including those who are not internationally mobile.  You cannot solve this domestic, political problem by restricting migration.  Finally, at least one study shows that Nigerian doctors in North America remit more than the cost of their education.
In light of all this evidence, why is there so much resistance to migration?  Why do so many people try to control migration while embracing structural transformation?  Politics cannot be the answer, as the small group of people who are likely to lose—typically recent migrants—are not powerful enough to block such a major gain to society.  It cannot be due to racism because migrants (and, increasingly indigenous populations) come from all races.  In his recent book, Paul Collier suggests that migration may upset the social equilibrium in the receiving country, whereby the current residents are willing to pay for the level of social services in their country, given a particular ethnic mix. Migration will change that mix. 

Instead of modeling the effects of migration, future research could usefully model the resistance to migration.  This research will probably take us into the area of behavioral economics.  We should build the empirical base to test hypotheses such as the one above.  And we should design policies that are informed by the research on the resistance to migration, rather than just the effects.


Submitted by Mushfiq Mobarak on

Great points, Shanta. I couldn't agree more, and I would actually go a little further. It turns out that the returns to internal, seasonal migration (moves from village to city within a developing country) are also very large. See here:, and shorter policy briefs on this linked here: And from some circles I get the same reaction to this as what you write above: "Migrants cause problems in cities", "we should focus on growing rural areas". Imagine a policy in the U.S. where the goal was "let's put people back in rural areas." That would take us back 200 years.

I think Kathleen Beegle at the World Bank and coauthors have similar numbers on high returns to seasonal migration in Tanzania. And Ana Maria Mayda (Georgetown?) and Dani Rodrik have done empirical research on the question you end with - "how do people feel about migration?"

Submitted by Paul Isenman on

You rightly say "But when that same worker happens to cross a national border, we call it “migration” and, instead of celebrating, we start investigating the effects..." You could also say, "But when that same worker happens to cross a national border, we call it "migration" and, instead of celebrating, we ignore by far its biggest consequence: the improvement in income and life chances for that worker and his/her family (in the new country)." We all tend to accept this intensely nationalistic view of poverty reduction without even thinking about it.

Submitted by Manjula M. Luthria on

Shanta, thanks for this great post! and bravo for pointing out that we don’t need another review of the costs and benefits of migration since it won’t get us any closer to finding the policy levers for making more of it happen safely and productively. On the resistance to migration, I can think of four possible explanations:
1. Not everyone has read the trillion dollar economic case. Let’s get Lant and Michael on the airwaves some more :)
2. Maybe it’s not economics after all? It’s other things that have nothing to do with economics so no matter how large the economic case it won’t matter.
3. It’s high transaction costs: employers don’t know how to sort and select from overseas, don’t know how to value skills acquired overseas etc. so the noise to signal ratio needs to be improved through labour market facilitation. I’ve seen this work when we’ve done pilots. And yet, I wonder if there’s room for a fourth explanation.
4. People can be irrational! They can vote one way and then act another. How else do you explain a flight that could not take off from Brussels a few months ago and had to turn back from the runway with passengers surrounding the cockpit begging the pilot to not take off. Why? A passenger was being deported to Morocco. I recently met someone who was on this flight and found the incident fascinating.

More than most other topics the World Bank analyzes, migration is a political rather than an economic issue. Resistance to migration can therefore only be understood by adding a psychological / sociological point of view ('fear of the unknown')to pure economics. I agree with the general arguments made in this post although they are hardly new and have been advanced repeatedly as part of the GFMD process. Just a couple of points.

1.) Any attempt to look at migration through an economic lens is inherently flawed. The benefits of migration cannot be equated with total gains for the global economy. This type of economic modelling does not take into account social costs to migrants, their families and host populations.
2.) The effects of migration on receiving countries vary. Studies in the UK by Ruhs & Silva (2012) ( show that the effects depend on migrants' skills levels. Moreover, benefits of migration are felt differently by different groups within the host society. They are generally largest for employers and smallest for unskilled native workers, who compete with migrants over low-paid jobs. The care sector in particular is prone to such competition. Such findings need to be taken into account when understanding resistance to migration in receiving countries.
3) The brain gain argument is complex - in the case of many developing countries higher education as a means to facilitate going abroad is part of national culture ('culture of migration'). This is problematic insofar that individuals tie their hopes and dreams to jobs abroad that might not be attainable to them, and also endangers their social relations. Furthermore, long-term gains for sending countries depend on whether their labour markets have the capacity to absorb these highly-skilled individuals. Often this requires a transformation that takes time and political will.
4) Preventing citizens from emigrating is deeply problematic and touches on human rights. However, it can be justified should the education have been paid for by the tax payer, in which case graduates can be required to reimburse the schools. Paying competitive salaries for businesses and institutions in developing countries is unrealistic. Instead, migration becomes a cost-benefit analysis that includes social and economic costs. Finally, whether remittances outweigh the costs sending countries spend on education depends entirely on the external environment. Plenty of studies have shown that remittances are no guarantee for sustainable development and studies have shown that they decrease over time as migrants become more invested in their new home countries. If invested in education they can encourage change at the household level. However, if not absorbed by the local labour market, well-educated individuals will follow their parents' example and go abroad thereby only contributing through further remittances.

All of these arguments are well-known but it is important to keep in mind that migration is always a product of a vast variety of different factors and cannot be treated as an economic equation.

Submitted by Shanta on

Thanks for your comment, Johanna. I think we are saying the same things, perhaps with different labels. First, the "economic lens" does not equate the benefits of migration with the total gains to the economy. It explicitly identifies the numerous winners and losers (migrants' families, host populations, etc.) and estimates the costs and benefits to them. These costs and benefits were the ones I cited in my blog post. Second, we agree that the benefits in receiving countries depend on the skill levels. Third, I'm not sure whether what you call the "culture of migration" is different from the general incentive to migrate, rather than anything specific to the brain gain. For instance, low-skilled people pay brokers large sums of money to migrate, which is not that different from people investing in higher education to increase their chances of migration. And you're right that the brain gain will take time to be registered as the better trained graduates are absorbed into the labor market. Here too I think we agree, because I was saying that the salary structure in source-country labor markets often prevents countries from benefiting from migration. Fourth, paying competitive salaries is not unrealistic if, as I suggested, they are paid only to those individuals who are internationally competitive, not to the rest of the profession as well. I realize this may be politically difficult, but it's important to understand what the constraint is (it isn't a purely fiscal constraint). Your final point is actually my final point too: that the economic case for migration is so compelling, that the reasons for its resistance must lie elsewhere, in culture and behavior. I hope we can work together in better understanding these phenomena, so we can design better policies to help migrants, those left behind, and those in receiving countries. Regards, Shanta

Submitted by John F. May on

Dear Shanta:

I read this Blog with great interest, as you bring together in a short space a number of important points.

Two additional topics are worth pursuing. First, the migration thresholds in receiving countries; beyond a certain number of immigrants, their social and even economic integration becomes more problematic. There is a growing body of literature on this.

Second, the circulation migration; a lot of current research looks at migration as one finite step, moving from A to B, whereas migrants can actually circulate (going back and forth several times) between A and B. Actually, this might be one the key policy paths for the future.

The Population Reference Bureau has just published "The Global Challenge of Managing Migration", a new Population Bulletin (see


John F. May
Visiting Scholar
Population Reference Bureau

Submitted by Suhas Ketkar on

The benign wage effects in recipient countries of immigration from (and trade with) low wage countries seem to be rooted in the belief that wages are driven by productivity. But this long-held belief is now disputed by some recent data and research. As a recent piece in the Economist (Labor Pains dated November 2nd 2013)argues: "A falling share of labor implies that productivity gains no longer translate into broad rises in pay." While many factors may account for it, can we categorically rule the role of migration and trade in this phenomenon? The research by Michael Elsby et al. referred to in the Economist article shows that industries in the United States that are more exposed to imports have experienced a greater decline in the share of labor in each industry's output. This research does appear to lend credibility to the view that the competition from cheap labor in poorer countries (via migration and/or imports) has contributed to the decline in the share of income earned by workers.

Like most main stream economists, I too held for a long time that trade and immigration had little negative impact on workers in rich countries. Is the evidence referenced above enough to warrant a second critical look?

Submitted by Shanta on

Suhas, thanks for your question, which is a good one. But I'm not sure that competition from imports is the same as competition from immigrants. I need to check, but I recall seeing some work showing that imported goods are more subsitutable with domestic goods than are immigrants with domestic labor. Worth taking a look, though. Shanta

Submitted by Pierre Jacquet on

Great post, Shanta, thanks.

There are persuasive instances of brain gain over the long run, as the exemple of research capacity building in economics in Latin America over the last decades seems to show. However, I don't think one should systematically expect brain gain, as suggested by work by Frédéric Docquier and others. Another caveat to your story is that labor mobility within countries is not always a cause for celebration, and the pace of rural exodus toward cities in the developing world while there is a lack of provision of basic services in the cities raises many concerns: actual labor mobility does not always imply moves from low productivity to high productivity jobs, and neither does migration.

In fact, you might make almost similar arguments for trade in goods and services (domestic vs. international), even though resistance to migration is probably higher and deeper than resistance to trade, notably because the mobility of persons is of a different nature. However, there is the same analytical slip there: benefits from trade are (presumably) more easily understood and accepted for domestic trade than for international trade. I was always intrigued by the fact that trade theory only addresses international trade (thus taking politics as given). I guess that beyond the factor mobility argument (which in fact may also applies domestically, for example when there are patent infrastructure gap like in many African countries show), much has to do with the financial and social “sustainability” of domestic mechanisms needed to mitigate the costs of transformation, as you suggest in your post. We might need to study not only the resistance to migration but also, more generally, the nature and multifaceted dimension (fiscal, economic, cultural, political, administrative, logistical, technical…) of the resistance to things foreign, i.e. the separation between “domestic” and “international⍊ and to what extent, or whether, this resistance is endogenous or should be taken as exogenous. Which echoes the work done on why there are countries and whether it is efficient…, but I don’t think this work has gone far enough, and it should involve many disciplines. Yet, it also implies discussing the effectiveness of the "international market mechanism": is the international reallocation of resources, including labor mobility, related to any market failure or to an efficient functioning of markets? Is there an optimal human resource allocation across countries, and is there room for an international “politique d’aménagement du territoire”, to echo one of the French public policy concerns across the French territory, as regions may fill up while others become empty?

Submitted by Shanta on

Pierre: Thanks for the great comments. Very thought-provoking. I hadn't made the link with the literature on why there are countries and whether they are efficient. We should follow up.

Submitted by Dilip on

This point - about the rationale for the nation state - is key to understanding the resistance to international migration. We need to address head-on the trade-off between diversity and national identity. Sen's book Identity and Violence has an interesting discussion of a solitarist approach to identity (national identity) vs a pluralist approach where an individual has many identities, and John Gray, in his review of Sen's book (see, mentions examples of how nations are built after much bloodshed and suppression of some forms of identities. This discussion then leads to the definition and efficiency of national borders: the border is not the same for goods as it is for services, and it is not the same for migration, and it is not the same for you as for me. And we are definitely getting more comfortable with different borders for different things and even "less" border for all things than, say, before 1990. There is a regression to erecting borders arising from the recent financial crisis-induced paranoia, but I hope this paranoia is temporary. Time has come to open the case (or lack of it) for the nation state.

Submitted by Luke on

Thank you for this.

On resistance, if I am a moderately stupid and lazy person in a rich country (and as we are not all above average, that's 49%), do I gain more from the general increase in wealth than I lose from the increased job competition? I am thinking not just of cash, but also positional goods such as status, well-located property, attractive mates etc? Or is it a rational fear that I might?

And are men or women more opposed to immigration? I ask because (historically at leastt) been most migrants have been men.

Submitted by Shanta on

Luke, good point, but the evidence is even more puzzling here. The opposition to migration in Europe seems to come from older people, who face the least competition from migrants (and have the most to gain). Young people, who would be the most threatened by competition from migrants are generally more supportive of migration! This is why I think it's necessary to look beyond the rational actor model to understand the resistance to migration.

Submitted by Counterpoint on

I think the work of the growth economist Garrett Jones on the externalities of human capital needs to be mentioned and understood to see the most plausible economic case against complete labor market liberalization. He finds very large externalities in macro regressions, and there are plausible micro-foundations: higher-IQ individuals cooperate more in Prisoner's Dilemmas, save and invest more, and have policy views that more closely resemble those of economists. Similar results have been found by other economists, such as Hanushek and Rindermann, using PISA and other international assessments of academic performance.

Citizens of the poorest developing countries where surveys show the largest desire to migrate tend to have very low levels of human capital using multiple measures including educational attainment, PISA scores, and IQ compared to developed countries (China has strong test performance, but only a smaller portion of the population expresses desire to migrate, the population of migratory age is a small portion of the world's potential migrants, and it is already rapidly developing). So this is not just a theoretical issue: full labor mobility would drastically lower the average level of education and human capital in immigrant-receiving countries.

Clemens' work is an excellent contribution to the debate, but it assumes that differences in productivity between countries are exogenous to the population inhabiting those countries. If the human capital, political views, and other characteristics of migrants affect the level of productivity in countries through externalities, such as on institutions, then this assumption breaks down. The analysis would then need to weigh the gains from moving people to better institutions against any deleterious (or positive) effects on the quality of those institutions.

This could still easily yield a huge net gain, but the case is not as simple as Clemens' presentation.

Submitted by Shanta on

If this argument is valid, why doesn't it also apply to movements within countries? When someone moves from a village to the city, why don't we observe average productivity going down in both places? As I said in the opening of my post, which several commentators confirmed, there is a large body of evidence from development economics that shows that migration (typically from farm to cities) enhances overall productivity in the economy. If this works for a rural Thai farmer moving to Bangkok, why doesn't it also work for an Ethiopian moving to Milan?

Submitted by Jaime Alvarez on

Albricias, albricias, Shanta! This is a very good piece and I am encouraged by the Bank's research on migrations, long overdue. I also noticed that the 2015 WDR will be on Mind and Culture (looking forward to their take on the effect of migrations). I participated in early activities on the subject, and recall Aaditya Mattoo’s initial moves. At the time I expressed him that this topic ought to be studied more, it’s gone a long way already.

We are in agreement that positive structural reform usually comes with migration. Apparently economists, with few exceptions, have a hard time seeing the obvious. Invasions, conquests, and dramatic forms of displacements have given us EVIDENCE that (by most part) migrations bring structural transformation. For good or bad.

I agree with you that many of the "statements made about the effects of migration on the receiving and sending countries are based on rhetoric rather than evidence". However, the robust evidence is not only found on recent research, as you imply, even if one decides that reality is optional. For example, in addition to hundreds of history books that put migration in context, Sowell's Conquest and Cultures, and Migration and Cultures, shed light into the effect of mobility around the world.

Now, I have to point-out to what I perceive as a shortcoming on your piece. And I have to be fair, and give to you that you did not have 20 pages to elaborate. The problem with research is that "evidence" is often what is available for research or what is sought (and desired) for the purpose of that research. This is applicable not only to social sciences but even biological and physical sciences. For example, you state:

“The effects of migration on the receiving country are positive. While some native workers (who are easily substitutable for migrants) may see their wages decline, most of the economy will see an increase in wages and employment, as migrants perform jobs—such as childcare services—that permit the economy to be more productive (for instance, by allowing both spouses to work outside the home).”

I wonder whether this statement suffers from the same lack of evidence you mention in your piece. What you said in the above paragraph may be correct, yet it is very incomplete. In a “welfare state” one wonders whether the effects of migration are positive for the receiving country. And let me state for the record that as a libertarian I tend to agree with the proposition that mobility enhances the general welfare. However, I have no evidence that this is so in a “welfare state” and may have perceive that the opposite is the case.

If you go to the slums of the Bronx, Los Angeles, or any other city in the USA you may be able to elaborate a hypothesis to challenge the conclusion that the recipient country is better off. Unlike the migrations from the early 1900s, that were basically workers going into the USA free market in exchange for a worker’s compensation the reality today could be (in many instances) completely different. Just recently the EU expressed its concerns about welfare transfers to immigrant from its member countries.

There are anecdotal evidence of people working for a salary in the informality, paying no taxes while collecting unemployment benefits, child support, spousal support and many other forms of welfare benefits that are likely to be a drain rather than a gain for the receiving country. In other words, the effects of free migration has to consider benefits (not only the benefits that allow people we know to have child care and have two incomes) but to the economy as a whole. To my knowledge, this research has not been undertaken completely, and the parts that have been done may have been recently modified because of political expediency?
Further, you talk about the net gains emanating from a developing economy sending a doctor overseas. I ask you, what are the costs assigned to the death of a child who is not served by the doctor that migrates to an industrialized nation and later sends money back home? And I am not yet getting into the crime and related matters brought about in the informality and an underclass. But let me refocus and go back to my point on immigration in a welfare state.

In a "rent seeking" global world, it is hard to find-out where the benefits and costs are really located. And when you find them they are hard to measure. Some well-connected corporations or industries may be interested in using cheap labor to be competitive in the short term, leaving a segment of the population unemployed and even as an underclass. Interesting to see how many economies are now showing more income disparities than before, revealing that the rent seekers at the top are doing pretty well. While some immigrants in these economies may be improving their lives at the expense of displaced workers that become an underclass, or simply takers, is that a desirable outcome? Too bad for those former workers as they were “the suckers” that contribute to the welfare that is feeding the newcomers? Is this outcome either efficient or fair?

In explaining the net impact of migration you went into the non burden nature that new populations have into medical systems of recipient countries. If reads as if medical systems are the only welfare feature in those systems. About education, transport, water and sanitation. But you are also limiting yourselves to one or two examples. What about populations that come as immigrants and malnutrition, in addition to other illnesses that may not put a burden on hospitals but may have public health implications for the society as a whole? If one include the negative externalities that contagious diseases and other things could bring, maybe the picture is not that clear. Among my friends we still discuss whether The Discovery was positive or negative. I will speculate that for some countries it was the transition from the Stone Age to the Renaissance, including the introduction of the concept human rights in the New Continent despite the dismal record in applying it in the New Colonies (measure this). Detractors of my view counter me by pointing at the negative externalities LAC received from The Discovery. In your example of migration from the farm to the cities, what did the urban taxpayer got in exchange for the rural migrant? Higher taxes, the slums, congestion, pollution, and crime and violence?

I know that my take is not based on research "evidence" and that my quest for answers may sound as if I disagree with free flows of workers. One thing I have clear: Workers and firms respond to market and non market rents. I am afraid that the data is not isolating what the mind is able to perceive, and the human action is very quick to capture. But maybe my feelings are based on "evidence", after all. Circa 1986 I had a discussion with a researcher at a think tank. I recall that individual debunking my argument for free migration by stating that in the welfare society, free migration was non-viable. I recently heard that same institution favoring migration. Why their change when the logic remains? I will be inclined to follow the money.

Once more, I welcome the Bank’s push for looking into the impact of migrations in economic development. Very promising.

Submitted by Shanta on

Jaime, thanks for that thoughtful comment. Your point about migration in a welfare state was exactly the analysis done by the OECD (most of whose members fall into this category). They looked at the net contribution of migrants to the fiscal balance (taking into account more than just medical services--that was just an illustration) and found it to be zero or positive. On the costs of doctors' migrating, I agree that the cost of a child dying is unspeakably high. But most of the evidence points to the fact that these children are dying all the time, and the number who die because a doctor migrates is no higher. In rural Tanzania, for instance, public doctors spend an average of 29 minutes a day seeing patients. This tells me that there is a shortage of patients, not doctors. (In fact, these doctors are also working in private practice, and try to encourage the patients to come to their private clinics and pay fees). If this is the state of medical services in poor countries, the migration of doctors is the least of their problems.

Submitted by Anonymous on

Very good and timely discussion. Looking forward to more discussions/comments.

You highlighted some very legit concerns, it would be good to have balanced in migrations / out migrations.. better yet, be able to live, work, play in the same locale.

Submitted by Pascal Jaupart on

A sad example of practical resistance to migration:

Submitted by senbolly on

Great piece of writing, I really liked the way you highlighted some really important and significant points. Thanks so much, I appreciate your work.

Add new comment