S pe ob G ci al l al M ResearchDigest Fo igr World Bank cu at s ion on VOLUME 14 NUMBER 1 FALL 2019 Migration is Key for Reducing Poverty and Sustaining Economic Growth G lobal migration is a human sto- migrants’ incomes. One suggests that IN THIS ISSUE ry. It may have started as soon national migration policies should as Homo Sapiens began walking focus on easing the costs of short- Migration is Key for Reducing Poverty and on two legs, and it is unlikely to end term dislocations of native-born work- Sustaining Economic Growth … page 1 any time soon. Most of this migration ers and distributing more widely the is voluntary and legal. Indeed, accord- economic benefits generated by labor How Labor Markets Fit Into the Global ing to a 2019 World Bank report, nearly mobility. The second shows that the Migration Debate … page 2 90 percent of the world’s 266 million economic payoff to moving to a richer migrants moved voluntarily for eco- country comes immediately—and per- How Big is the Payoff From Migration? Evidence from a Lottery … page 3 nomic, social, and climate reasons. sists for generations. The third finds Global migration has become an that easing restrictions to migrate Reducing Barriers to Migration Boosts important instrument to lower poverty from China’s rural villages to booming page 4 Welfare in China’s Rural Villages … rates in developing countries and raise cities boosts the welfare of the poor- productivity in developed countries. In est households. Exposure to Immigrant Classmates and 2015, the migrant population gener- The next four articles explore the Natives' Test Scores … page 5 ated about 10 percent of global GDP, troika of aging, education, and mi- even though it accounted for less than gration. A study on test outcomes How Aging, Education, and Migration 4 percent of the global population— indicates that exposure to immigrant Affect Wages … page 6 with nearly 90 percent of the economic classmates does not go hand in hand benefits of migration realized in devel- with lower scores for native children. Highly Skilled Migration: “Brain Drain” or oped nations. Another study suggests that wages are “Brain Gain”? … page 7 Migration is also a great leveler, more influenced by aging and educa- providing better prospects for the tion than migration. A third study Aging, Trade, and Migration … page 8 struggling millions in the developing finds that diasporas of highly skilled world. In 2018 low- and middle-income migrants—such as inventors—offer Global Spillovers of National Migration countries received $530 billion in re- a path toward the internationaliza- Policies … page 9 mittances—which is more than three tion of inventive activity. And a fourth Getting a Better Sense of How Long times the size of official development study indicates that migration offers Refugees Have Been in Exile … page 10 assistance. Migrants held more than perhaps an even better way than trade $500 billion in annual savings, which to cope with the rising share of older Creating a Migrant Rights can be mobilized via diaspora bonds people in the world’s population. Database … page 11 for financing for development projects. The following two articles focus The World Bank’s research on mi- on recent studies that take a new Recent Policy Research Working Papers gration, which is carried out through- approach to determining migration on Global Migration … page 12 out the institution, offers compelling patterns—notably, those of “transit hard evidence on the economic gains. migrants”—and the length of time The goal is to inform the public and refugees spend in exile. A final ar- help policy makers make informed ticle describes a new Migrant Rights decisions within the existing political Database, which will enable country and economic constraints. This issue comparisons as the international of the digest highlights examples of community works toward a global World Bank research on migration. standard for migration governance The first three articles focus on and the protection of human rights. 2 World Bank ResearchDigest How Labor Markets Fit Into the Global Migration Debate Migration policies should help A Pattern of High Concentrations dislocated workers. In addition, the lower short-term dislocation costs One of the report’s key findings is most pressing needs of the labor mar- of native-born workers and more that global migration patterns lead to ket can be met by matching migrant high concentrations of immigrants in workers with employers that need widely distribute economic benefits certain places, industries, and occupa- them the most. T tions. For example, the top 10 destina- •  The creation of a pathway to he rich have many assets; the tion countries account for 60 percent permanent residency for migrants poor have only one—their la- of global immigration. In the United with higher skills and permanent jobs, bor. Because good jobs are slow States, four states host half of all im- which will provide incentives for them to come to the poor, the poor must migrants, and 10 counties host half of to fully integrate in the labor markets move to find productive employ- the immigrants in these four states. and acquire skills specific to the desti- ment. Migrants’ incomes increase These patterns are what would be nation country. three to six times when they move expected from an adjustment mecha- from lower- to higher-income coun- nism that is aimed at productivity tries. The average income gains for a gains, wage increases, and poverty Quotas should be young unskilled worker moving to the reduction. But they pose a problem as replaced with market- United States is estimated to be about market forces rush to fulfill unmet de- $14,000 per year. If the number of im- mand, especially for domestic popula- based price mechanisms, migrants in high-income countries tions who have easily substitutable such as fees, to manage were doubled, by moving 100 million skills and occupations. Furthermore, young people from developing coun- the positive effects and benefits in migration flows tries, the annual income gain would the destination labor markets tend to be $1.4 trillion. This global welfare be more diffuse and slower to realize, The bottom line for the global mi- gain dwarfs the gains from the remov- whereas the costs are more concen- gration debate is that it is a mistake to al of all restrictions on international trated, immediate, and easily attribut- ignore the distributional impact and flows of goods and capital. able to immigration. dislocation such flows would generate, These income gains, however, re- especially in destination countries, as main largely notional because most the efficiency gains are realized. But people cannot move. Only about 3 Effective migration it is also a mistake to deny or ignore percent of the world’s population policies must work with, the significant efficiency gains. The live in a country in which they were solutions—that is, the policies—rest not born, a proportion that has not rather than against, in making the pie as large as possible, changed much over six decades of labor market forces and, at the same time, figuring out a otherwise unprecedented global in- way to distribute it more equally. tegration, via trade, investment, and knowledge flows. Distances in space, What policies could be adopted to culture, and language are inherent ensure the benefits of migration are impediments to mobility. But the shared by host and immigrant com- most important barriers are national munities for generations to come? The borders—partly fueled by citizens wor- report highlights the following: ries about what migrants and refugees •  Effective migration policies would do to jobs and wages, welfare must work with, rather than against, programs, crime levels, school quality, labor market forces. For example, and their national identity. where there is large unmet demand A new report authored by Çaglar for seasonal work, temporary migra- Özden and Mathis Wagner seeks to tion programs (like those in Canada or address the tension between compel- Australia) could address labor market ling academic evidence on the eco- shortages while discouraging perma- nomic gains and the stark public op- nent undocumented migration. position to immigration. It suggests •  Quotas should be replaced that what is needed is a labor market- with market-based price mechanisms, oriented, economically motivated such as fees, to manage migration rationale to the political opposition to flows. Such tools can pay for the cost World Bank. 2018. “Moving for Prosperity: Glob- migration. of government assistance to support al Migration and Labor Markets.” Policy Research Report, World Bank. Washington, DC. World Bank Research Digest 3 How Big is the Payoff From Migration? Evidence from a Lottery Outcomes for migrant lottery tries to answer this question by exam- Tonga) were NZ$126, so the gain from winners suggest the economic ining the impacts of migrating from migrating represents a 271 percent payoff to moving to a richer country Tonga to New Zealand via the over- increase in weekly earnings—very sim- subscribed PAC (between 2002-05), ilar to the 263 percent gain estimated comes immediately and persists for where a lottery with in the first year after at least a decade 10 percent odds of Small, typically island migrating (figure 1). T success was used states (mostly in Since the migra- ongo—an archipelago of islands to fill the available tion impact appears in the Pacific about three hours slots. It compares the Pacific or Caribbean) to be stable over north of New Zealand by plane— Tongans who had are the origin countries time, and because has a gross national income (GNI) per successful applica- the authors observed capita of $4,150 (in 2005 PPP $), which tions in the lottery, most affected by no return migration is about one-sixth of the level of that and who immigrated migration after a decade, the of New Zealand. In 2002, New Zealand to New Zealand lifetime income gain introduced a new migration program, almost a decade from winning the the Pacific Access Category (PAC), ago, with unsuccessful applicants in PAC lottery and migrating is straight- that lets a quota of 250 Tongan per- this same lottery. The effort and cost forward to estimate. The typical PAC manently migrate each year. But will of tracking applicants down a decade migrant moves to New Zealand aged migration help boost the income of after applying for migration, coupled around 32. Assuming the individual the Tongans, and if so, how long will with the migration lottery, make this works for 33 years thereafter (the the boost last? study unique in its ability to provide public pension in New Zealand is This matters, and not just for the estimates of the long-term causal im- available from age 65), the lifetime Tongans. The same story is being pacts of migration. earnings gain is NZ$583,440. A fur- played out all over the world, given ther expected benefit in retirement of the massive income gap between rich A Lifetime Win NZ$186,760 accrues because the New and poor nations. In 2014, GDP per The results suggest that the eco- Zealand pension is more generous capita in high-income OECD countries nomic payoff to migrating to a richer than Tongan ones. In net present value was nine times that of middle-income country comes immediately, and then terms, with a 5 percent discount rate, countries, and 68 times that of low- hardly grows. The wages of principal the lifetime gain due to migration is at income countries. Moreover, small, applicants were NZ$340 (US$260) least NZ$315,000 (US$237,000), even typically island states (mostly in the per week higher due to the impact of before allowing for non-monetary ben- Pacific or Caribbean) are the origin migration (controlling for ex ante char- efits (like consumption, durable asset countries most affected by migration. acteristics, the lotteries entered, and ownership, savings, and dietary diver- A recent study by Gibson, the type of survey). The weekly wages sity) and wealth accumulation effects. Mackenzie, Rohorua, and Stillman for the control group (lottery losers in These findings support the idea that cross-country wage differences Figure 1. Wages of Migrants Quickly Jump to a Multiple of that of Non-Migrants, are due to better institutions, higher Before Stabilizing Over the Next Decade quality capital, and other factors in 4 rich countries that serve to raise the productivity of all workers—rather than attributes that are embedded in Ratio of migration to non-migrant incomes native workers (like language skills, higher education, and country-specific 3 knowledge) and take time for migrants to accumulate. 2 John Gibson, David Mackenzie, Halahingano Rohorua, and Steven Stillman, “The Long-Term Impacts of International Migration: Evidence from 1 a Lottery,” The World Bank Economic Review, 0 2 4 6 8 10 Volume 32, Issue 1, February 2018. World Bank, Years after migration Washington, DC. 4 World Bank ResearchDigest Reducing Barriers to Migration Boosts Welfare in China’s Rural Villages Higher consumption. Out-migration has a positive effect on annual chang- es in log consumption between 1988 A greater ability to migrate from areas may supplement income earned and 2002: out-migration explains be- China’s rural villages to booming locally and directly reduce exposure tween 2 and 2.9 percent of the annual cities boosts goods consumption to poverty. Out-migration may also increase in consumption per capita, have indirect effects on household and or between 65 and 93 percent of the with poorest households key individual welfare within their home annual consumption growth recorded beneficiaries communities—either in the form of in the sample villages between 1988 I increased wages with the depletion of and 2002. n developing countries, barriers the local labor force, or through remit- Poorest villagers benefit most. Whether to the movement of labor are a tances from migrant employment that households participate directly in common institutional feature that are invested in local production. But migration or not, consumption grows may contribute to geographic poverty to date, there is little in the literature more rapidly among households in the traps—regardless of whether these on this topic. poor and middle terciles of the initial constraints are maintained by formal A recent study by de Brauw and average consumption distribution. institutions, cultural or linguistic dif- Giles tackles this issue by examining Higher incomes. Later in the study ferences across regions, or simply high how lower barriers to migration for period (after 1995), incomes also transaction costs associated with find- work contributed to changes in welfare increase among all households, ing migrant employment. This occurs in rural China. The authors use a panel although the increase is faster because the barriers may reinforce household survey (from 88 villages among householders that were an inefficient allocation of resources and 6,305 households) spanning the initially poorer. across regions and influence invest- period from 1988-2002, during which Shifts in labor allocations and invest- ment levels in poor areas, potentially the population of rural migrants work- ments. Out-migration from the village hindering growth. However, when they ing in cities increased from roughly 20 leads to labor reallocation across ac- are removed, the resulting improved million to over 110 million. Over this tivities, and labor allocation changes efficiency of resource allocation may period, institutional changes—notably, differ by wealth tercile. Agricultural boost rural living standards—as oc- a 1988 law allowing migrants to estab- labor days decline among poorer curred in China in the 1990s. lish legal temporary residence and a households, who also provide a rela- What is the impact of out-migra- new national ID card necessary to ob- tively larger increase in days spent tion on the distributional effects of tain legal residence—made it easier to working outside the home township. migration within villages, along with migrate temporarily for work (figure 1). More affluent households, by con- the sectoral distribution of household trast, increase labor supply to local activities across the within-village A Boost for Living Standards non-agricultural activities, potentially wealth distribution? This matters be- The results of the study highlight reflecting general equilibrium effects cause remittances to household or how reducing barriers to migration contributing to increased local op- family members remaining in rural would boost welfare. They include: portunity as migrants remit earnings. As for investment, poorer households Figure 1. Migrants Take Up a Bigger Share of Workforce After National IDs Issued invest more in housing and durable goods, while more affluent house- 0.14 holds invest more in non-agricultural productive assets. Looking ahead, as residential registration is eliminated, and more Share of Migrants in Village Workforce elderly and children join migrant fam- 0.10 ily members in China’s cities, there will be a need for future research on the welfare impacts of family migra- tion—including educational access, the performance of migrant children in 0.05 urban schools, and access to care for the elderly and infirm migrant family members. Local Polynomial Smoother Alan de Brauw and John Giles, “Migrant Labor 95% Confidence Interval Markets and the Welfare of Rural Households in 0 the Developing World: Evidence from China,” The −5 0 5 10 15 World Bank Economic Review, Volume 32, Issue Years Since IDs Issued 1, 2018. World Bank. Washington, DC. World Bank Research Digest 5 Exposure to Immigrant Classmates and Natives’ Test Scores Duration of Stay Matters The key finding is that in primary schools in the Netherlands, the test Immigration debates rage on, but Netherlands, a country that in 2011 scores of native children overall are native children will find new friends had 1.77 million immigrants (about 11 not affected by immigrant children. But in school, not worse learning percent of the population). As in most if there are any spillover effects, the European countries, the majority of magnitude will depend on the duration outcomes these immigrants came from lower-in- of stay of first-generation immigrant A come countries. Early waves involved classmates. For example, immigrant s immigration continues to former Dutch colonies (Indonesia, children who have been in the country dominate political debates, a Suriname, and the Dutch Antilles); for some time do not have any detri- growing number of policymak- later waves involved Morocco, Turkey, mental effect on natives, whereas there ers and citizens are concerned that Iraq, Afghanistan, and Iran. The au- is a small negative impact of the pres- the presence of immigrant children thor uses a rich dataset of students ence of recent immigrants on natives’ in schools may harm native chil- in a panel of primary schools in the learning outcomes in Dutch language dren’s learning outcomes. In Italy, for Netherlands, which includes detailed only—that is, not in mathematics (fig- example, the Ministry of Education information on the background of im- ure 1). These effects tend to be slightly introduced a law in 2010 that caps the migrant and native children—such larger when immigrant children have share of foreign-born students in pub- as the duration of stay of immigrant not been exposed to the Dutch lan- lic school classrooms at 30 percent. classmates in the country. This rare ac- guage prior to migrating, but even Such measures, however, are cess to timeline information enables then, they remain quite small. largely motivated by anecdotal evi- him to separately estimate the impact Yet despite the small magnitude dence of disruption, rather than clear- of foreign-born peers who recently ar- of these estimated effects, the study cut results of rigorous econometric rived in the Netherlands versus those highlights the importance of effective estimations—in fact, the literature who arrived at an earlier age. inclusion and assimilation of immi- on this topic is quite thin. In addi- To estimate the effect of immigrant grant children—like facilitating com- tion, economic theory is inconclusive children in the classroom on natives’ mand of the host country’s language. about whether immigrant concen- test scores, the study uses small It indicates that after only a few years tration in the classroom produces changes in immigrant concentration in the country, when immigrant chil- positive or negative effects, if any, across cohorts within the same school. dren gain greater proficiency in the on the performance of natives. That It, therefore, assumes that small varia- local language, they do not have any said, there are reasons to suspect tions in the share of foreign-born stu- adverse effect on natives’ learning that immigrant children from differ- dents across cohorts within the same outcomes. Because of the similarities ent socio-economic backgrounds, school are driven by factors exogenous in the immigration context shared by countries of origin, or with different to native children’s test scores. The the Netherlands and other countries, durations of stay in the host country study also presents a range of balanc- notably the predominance of immi- may have a different impact on native ing and placebo tests, whose results grants from low socio-economic back- children. support the validity of the identifica- grounds, the study’s findings can be of A recent paper by Bossavie ex- tion strategy. relevance beyond the Dutch context. plores what has occurred in the Laurent Bossavie. 2018. “The Effect of Immigra- Figure 1. Immigrant Classmates Don’t Affect Dutch Students’ Test Scores Overall tion on Natives’ School Achievement: Does Length Estimated Effect of a One-Standard Deviation Change in Class Composition on Students‘ Test Scores (in Standard Deviations) of Stay in the Host Country Matter.” Policy −0.18 −0.16 −0.14 −0.12 −0.10 −0.08 −0.06 −0.04 −0.02 −0.00 0.02 Research Working Paper 8492, World Bank. Share of immigrant classmates Washington, DC. Also forthcoming in Journal Math Dutch of Human Resources, 55 (2): https://dx.doi. Share of recent immigrants org/10.3368/jhr.55.3.1017-9151R2. Share of non-recent immigrants Share of immigrants from low-income countries Share of immigrants from high-income countries One additional student in class (Krueger, 1999) Classmates from lower socio-econ. background (Ammermueller and Pischke, 2009) Classmates of lower ability (Duflo at al., 2010) Note: The blue and yellow bars display the effect of a one-standard deviation increase in the share of immigrants in the classroom on the average test score of Dutch students in the classroom (in standard deviations), estimated from Bossavie (2018). An increase by one standard deviation in the share of recent immigrants corresponds to a 5 percentage points increase. The three light blue bars at the bottom display the effect sizes of other changes in classroom composition estimated by other papers in the literature, also in standard deviations. Statistically significant effects at the 5 percent level are outlined in dotted blue. 6 World Bank ResearchDigest How Aging, Education, and Migration Affect Wages higher-skilled people—unlike the old, high-skilled, who suffer a large nega- tive impact (figure 1). Age and edu- Aging and education have the substitutable—and that immigrants cation account for most of the large strongest influence on wages, and natives are imperfect substitutes positive impact, while immigration, favoring the low-skilled, with across age-education cells (based on the villain in much political discourse, U.S. and UK evidence)—the impact on turns out to be a relatively feeble and immigration a minor factor that also wages is derived using a model with actually a positive phenomenon. helps reduce wage gaps imperfect substitutability between The relative impact of these fac- I different types of workers. It also as- tors, however, differs across OECD n industrial countries, the com- sumes no technological changes. countries: position of the labor force is be- •  In most of Western Europe, ing shaped by three forces: aging, Age and Education, Not Japan, and the United States, age- education, and migration. Aging Migration, Dominate ing and changes in the stock of skills, is a powerful, gradual process that The most striking feature is the account for most of the changes in evokes concern but little policy action. decline across all countries in the wages, and migration plays a mar- Immigration is blamed for the predica- share of the young low-skilled work- ginal role. ment of the low-skilled and is provok- ers, which is matched by the increase •  In the Anglo-Saxon countries ing a political backlash. Education is in the share of the skilled (young and (Australia, Canada, New Zealand, the viewed as a panacea to which access is old) in almost all countries. In ad- United Kingdom, and Switzerland), unequally distributed. dition, there has been an increase which have relatively liberal and skill- What is the relative and combined in the share of the old low-skilled in biased migration policies, immigration effects of these labor supply forces on most countries. Further, the number accentuates the egalitarian impact of wages across different skill and age of international migrants grew from ageing on wages. groups? This is a pertinent question as 2.9 percent of the world population in •  In countries that have seen increased wage inequality dominates 2000 to 3.3 percent in 2010. Migrants significant labor outflows—such as the political and academic debate, yet represented 11 percent of the total Ireland and certain Eastern European it remains to be addressed. population in OECD countries. countries (the Czech Republic, In a recent paper, Docquier, Kone, How do these changes affect Slovakia, Hungary, and Poland) that Mattoo, and Özden explore the labor wages? The results show that the now have access to Western European market implications of the changes change in the age and skill structure labor markets—the impact of ageing is in the education and age structure of the workforce is the dominant fac- partially offset by emigration. of the population between 2000 and tor on wage changes. This impact is What happens if the model allows 2010, drawing on the new Database on quite uniform and egalitarian, favor- for sufficiently strong skill-biased tech- Immigrants in OECD countries (DIOC). ing the low-skilled workers by making nological change? The results show Given that workers of different ages them relatively scarce and enabling that this change offsets the egalitarian or education levels are not perfectly them to work with a larger number of effects of higher skill endowments. It also reconciles the narrowing of wage Figure 1. Low-Skilled Workers Benefit Most in Terms of Wage Gains gaps predicted by the model with Wage effects of socio-demographic changes by age-education group the higher skill premium observed in 30 many countries. And it confirms skill- biased technological change as a key contributing factor to wage inequality 20 in past decades. Nevertheless, ab- stracting from technological changes 10 helps to isolate and highlight the relative impact of labor supply forces that are shaping the wage structure in Percent 0 industrial economies—both to inform the political debate and to identify NZL reform priorities. −10 USA IRL LUX SWE NOR CAN AUS BEL FIN DEU DNK JPN CHE FRA NLD CZE −20 GBR AUT GRC SVK TUR Frederic Docquier, Zovanga L. Kone, Aaditya ESP ITA PRT HUN Mattoo, and Çaglar Özden. 2018. “Labor Market POL Effects of Demographic Shifts and Migration in −30 OECD Countries.” Policy Research Working Pa- Young, LS Old, LS Young, HS Old, HS per 8676, World Bank, Washington, DC. World Bank Research Digest 7 Highly Skilled Migration: “Brain Drain” or “Brain Gain”? technology-leading European coun- tries (like Germany or France) lag way behind compared to the United States. Diasporas of highly skilled migrants, as capital or the expatriates’ network And the share of inventors from devel- such as inventors, offer a path of colleagues and acquaintances. oping countries is considerable—and toward the internationalization of Against this backdrop, diaspora it is on the rise in recent years. networks are being increasingly stud- inventive activity ied in the context of trade, foreign Diaspora Networks Could Offer A direction investment, international “Brain Gains” s migration flows across na- diffusion of ideas, and firms’ inter- The results show a robust effect of tional borders intensify, the nationalization strategies. A recent high-skilled diasporas on the interna- oft-quoted broad figures hide an study by Ernest Miguelez contributes tionalization of inventive activity be- important variation: unlike in the past, to this literature by investigating how tween developed, receiving countries highly skilled individuals represent a high-skilled emigrants activate their and developing, sending economies: substantial and increasing share of diaspora networks, overcome interna- a 10 percent increase in the inventor international mobility flows. During tional barriers, and foster the interna- diaspora abroad is associated with the 2000s, the number of tertiary tionalization of knowledge production. a 2.0–2.2 percent increase in inter- educated immigrants living in OECD Specifically, it looks at inventor dias- national patent collaborations. The countries increased by 70 percent poras and the production of patents evidence found survives the inclusion compared with only 10 percent for in international teams for a sample of of a large number of controls, fixed- the low-educated ones. And migra- developed, receiving countries and a effects, robustness checks, and iden- tion rates for the tertiary educated are group of developing, sending econo- tification issues. Moreover, the effect higher than those for the rest of the mies from 1990 to 2010. The use of in- is stronger for inventor-to-inventor population and generally increase with ventor data offers two main advantag- collaborations (co-inventorship) than further education. es. Patent data (together with inventor for applicant-to-inventor co-patents Moreover, bilateral flows are be- information) are registered and so can (R&D offshoring), which suggests that coming skewed. OECD countries be organized on a yearly basis—unlike diaspora effects specifically medi- (especially the Ango-Saxon ones) are census data, which are collected only ate interpersonal relations between the larger receptors of global talent. every 10 years. Also, the level of edu- co-workers. Plus high-skilled migrants originate cation attained may differ markedly These findings do not suffice to in a larger number of countries, in- among tertiary educated workers. conclude that a “brain gain” exists that cluding emerging and middle-income Among this group, Canada, makes up for the loss of high-skilled ones. This trend has brought back old Australia, and, notably, the United human capital of sending economies. concerns about “brain drain” and the States stand out as being the primary Note, however, that boosting inter- depletion of skills resources in ori- receiving countries, when compared national co-inventorship and team gin countries. However, high-skilled to their resident stock of inventors formation is only one of the multiple emigrants can also contribute to their from developing countries, while brain gain effects of emigrant inven- home country’s development in a Japan is, and has been over the years, tors—which may eventually include number of ways—referred to as “brain one of the developed countries with the international diffusion of knowl- gain”—by sharing embedded knowl- a smaller share of inventor immigrant edge or the accumulation of human edge and accessible resources, such population (figure 1). Meanwhile, capital in sending economies. These findings also support the Figure 1. Some Receiving Countries Have a Much Higher Share of Immigrant Inventors idea that exploiting high-skilled dias- from Developing Countries than Others pora networks in technology frontier Immigration rates of inventors, 2001–2010, receiving countries economies might be an instrumental way of engaging in international in- Immigration rate of inventors novation networks. This subject mat- Immigration rate from developing countries ter ranks high among policymakers in these countries, as witnessed by the Immigration rate 2015 visit of the Indian Prime Minister to Silicon Valley. Ernest Miguelez. 2016. “Inventor Diasporas and the Internationalization of Technology.” Policy Ire d d um US nd s ia UK a da k ay d en ain Ge ce y ly n a nd ar an Research Working Paper 7619, World Bank, lan lan ali lan pa re Ita str rw n ala na ed lgi nm Sp Ko rla Fra rm str Ja er Fin Au Ca No Sw Be Ze e De itz Au of th Washington, DC. Sw w R. Ne Ne 8 World Bank ResearchDigest Aging, Trade, and Migration Dynamic sectors may relocate to expenditures on health care (apart of view, erecting barriers at the border younger markets when population from what is provided by national may appear to make sense economi- ages. Immigration can prevent this insurance) comprised 7.6 percent cally. However, this logic is shortsight- of goods and services spending for ed: when one country erects barriers, relocation more effectively than households headed its partners do the tariffs by a senior aged at The problem is that same in response. In B least 65 years old industries like those the end, a trade war y 2050, the percentage of the and only 2.9 percent may be triggered, world population over 60 years for a household producing smart-phones which would hurt of age is expected to reach 21 headed by someone are often more innovative the ageing country percent—a dramatic increase from 12 under 30. In addi- more than the part- percent in 2013, and just 8 percent in tion to health care, than non-tradable ner. Yes, some firms 1950, according to the United Nations. seniors would be services would come back What are the economic implications of more likely to pur- home, but the losses this trend for aging economies? chase personal and from paying more for A recent study by Chisik, Onder, household maintenance services than imports and earning less from exports and Qirjo tackles this question by ask- the young. are much greater than these gains. ing, in theory, how demographic aging Therefore, the higher the share Third, in comparison, allowing in could shift trade policy preferences in of old people in the population, the young immigrants provides a better mecha- an economy. It features several techni- higher the demand for services (which nism to mitigate the firm delocation effect. cal innovations. A key one is that it are more difficult to import than man- Immigration can reduce the demand- focuses on demand changes, unlike ufactured goods) and the lower the de- driven effects of an aging society—as most studies on demographic changes mand for goods that can be imported. long as it is not too tilted toward and trade, which focus on factor mar- As a result, more firms find it lucrative older immigrants. If the immigrants ket changes. The authors do this by to get into the services business when are primarily young, the population is breaking away from a convenient, but demand pushes the price of services initially small, purchased services are empirically challenged, assumption: up. Some firms may cease producing a small percentage of total income, or the marginal (and average) propensity goods and become service provid- the difference between young and old to consume a good are not dependent ers, and others move overseas where consumption patterns is small, then on income. Instead, their model com- demand for goods like smartphones a country’s share of manufacturing bines Stone-Geary form preferences, remains large. Technically speaking, will increase along with increased im- where the demand for non-tradable aging has a firm delocation effect. By migration. This study’s results on the services rises disproportionately when itself, such delocation is not necessar- age of immigrants suggests why coun- income increases, with a monopolistic ily a bad thing: the economy produces tries like Australia, Canada, and New competition model that helps show more services because it consumes Zealand may have stressed a point- how trade costs can interact with more of them. The problem is that system (which rewards youth among market size in forming firms’ location industries like those producing smart- other things) over family reunification decisions. phones are often more innovative than (which can attract elderly parents). non-tradable services. Thus, the aging A Case for Immigration and Trade country would potentially lose income The results suggest that both im- in the long-term, as its workers would migration (especially of younger miss an opportunity to become more people) and trade offer policy offsets productive. to the negative outcomes from demo- Second, with demographic aging, more graphic aging. firms move overseas if trade barriers are low. First, in aging societies, production When firms can freely ship back their may shift toward sectors that are neither products after moving overseas, mov- easily tradable nor prone to rapid productivity ing closer to a larger overseas market growth. To see this, note that the old becomes more appealing, as the home consume more services like health- market is not lost. But if the ageing care or long-term care and the young country imposes egregiously high tar- consume more goods like smart- iffs on imports, smartphone producers Richard Chisik, Harun Onder, and Dhimitri Qir- phones. For example, a 2013 survey might rethink their relocation deci- jo. 2016. “Aging, Trade, and Migration.” Policy of Canadian households found that sions. Thus, from a protectionist point Research Working Paper 7740, World Bank. Washington, DC. World Bank Research Digest 9 Global Spillovers of National Migration Policies Current and potential migrants common language and similar in- However, the distance between the constantly weigh location-specific stitutions. The cost of moving from current location and a new destination utility levels, migration costs, and the United Kingdom to California brings almost no additional cost for is probably lower than the cost of high-skilled transit migrants. (expected) future opportunities moving from the United Kingdom to What happens when the moving T Japan, although both destinations are cost from Canada to the United States he migration paths of interna- about 6 thousand miles away from rises to a level such that migration lev- tional migrants often include the United Kingdom. Hence, when a els decrease by about 99 percent? The multiple destinations and transit computer scientist from a develop- results show that migration from other routes. Many people—so-called tran- ing country decides to migrate to the countries to Canada falls between 3 sit migrants—leave their birth coun- United Kingdom, she is probably also percent and 18 percent, even though tries and live in different locations be- accounting for job prospects in Silicon there are no changes to migration fore settling permanently in a foreign Valley. Thus, when one takes transit costs or benefits regarding Canada. country or returning home. Between migration into account, mitigation The impact is especially high for Latin 2001 and 2012, around 9 percent policy is no longer a bilateral eco- American and Caribbean migrants. of the people who migrated to the nomic issue. Another potential effect is the impact United States were living in a coun- Yet transitory migration patterns on other critical transit countries. But try other than their birthplace prior to are not explored in depth in the the findings show that there is almost their arrival. This pattern is even more international migration literature. no impact on migration from most common for migrants with tertiary Canonical global migration models are countries to the United Kingdom when education: nearly 14 percent of them static, not allowing for repeated mi- the Canada–U.S. border is closed— did not come directly from their birth gration decisions or transit routes. In although there is higher migration countries. a recent paper, Artuc and Özden con- from Canada to the United Kingdom, High-income OECD countries struct a novel dynamic model of global as the Canadians are also unable to are key transit stops for immigrants. migration that is aimed at incorporat- move to the United States. Among those living in Australia, ing and explaining transit migration How about when the moving cost Canada, or the United Kingdom just patterns; it uses the World Bank global to the United States rises to a pro- before their arrival in the United migration database and the American hibitive level for direct migrants from States, over 30 percent were born in Community Survey. In the model, mi- developing (non-OECD) origin coun- a different country. Although Canada, grants decide to stay in their current tries? The results show that Canada the United Kingdom, and Australia location or move to another one every and the United Kingdom become are far apart geographically, they are period, taking instantaneous utility more attractive destinations, because important transit routes for each other payoffs and bilateral mobility barriers they are high-income economies and for the United States, owing to into account. with similar income opportunities for migrants—and also provide a path- “All Roads Lead way to the United States. Migration Table 1. Migration to U.S. from Developing Countries is Blocked to America” to Canada and the United Kingdom to U.S. via Directly to to U.S. via Directly to The study increases by between 25 percent and Birth Country Canada Canada UK UK begins by asking 45 percent for many Central American Canada 0.0 0.0 3.5 0.0 whether language and Caribbean countries, and by a and distance staggering 165 percent for the Other Mexico 6.6 9.4 6.6 9.4 matter for transit Caribbean region (includes Haiti El Salvador 29.2 43.9 28.0 43.9 migrants when all and other smaller island countries) Guatemala 16.1 22.0 15.9 22.0 paths are open. (table 1). Other Central America 17.9 25.8 17.6 25.8 The results show With these findings in mind, we that common shall adapt the famous medieval prov- Cuba 19.3 25.9 18.6 25.9 language lowers erb to modern times: “All roads lead to Dominican Republic 21.6 29.8 21.0 29.8 transit migration America,” even when some roads are Jamaica 30.5 46.8 28.3 46.7 costs significant- closed. ly, with migra- Other Caribbean 94.6 165.4 87.6 165.3 tion increasing Brazil 2.8 3.9 2.8 3.9 60 percent for Erhan Artuc and Çaglar Özden. 2016. “Transit Colombia 6.3 6.8 6.0 5.8 unskilled and Migration: All Roads Lead to America.” 2016. 95 percent for Policy Research Working Paper 7880, World Other South America 4.9 5.5 4.5 5.5 skilled migrants. Bank, Washington DC. 10World Bank ResearchDigest Getting a Better Sense of How Long Refugees Have Been in Exile A rigorous review of UNHCR data Commissioner for Refugees as of end- bulk of the current Somali refugees finds that more than half of global 2015, a recent study by Devictor and (0.4 million), and people fleeing from refugees have spent 4 years or less Do proposes a new approach to esti- Colombia (0.3 million) and Myanmar mating the average and median dura- (0.2 million). in exile, a historical low tions of exile, and their variations over •  About 2 million people have P times. The authors re-calculate the been in exile between 10 and 34 years. olicy makers—from East Africa earliest date at which various cohorts This cohort includes years during to South Asia, from the Mashreq of refugees (including asylum seekers which numbers are relatively low, and to the European Union—are and people in refugee-like situations) two episodes where they are higher— struggling to find an adequate re- could have arrived in each situation, around 14 years ago, with the arrival sponse to a global refugee crisis, and then aggregate all situations into of about 0.2 million Sudanese refu- which has deep and potentially sig- a single ”global refugee population.” gees, and around 24 to 25 years ago, nificant political, economic, and social with the arrival of about 0.1 million consequences. Developing countries A Snapshot of the “Global Refugee Somalis and 0.1 million Eritreans. host the bulk of refugees. Yet, recent Population” •  Another 2.2 million refu- events demonstrate that without an What are the results? First, the gees have been in exile for 35 to 37 adequate international response, the study shows that among the “global years—mainly Afghans, but also crisis cannot be easily contained to refugee population”– currently at about 0.3 million ethnic Chinese who a few host countries; some refugees 15.5 million—there have been dis- fled into China during the 1979 war engage in secondary movements tinct episodes of displacement since with Vietnam. and increasingly try to reach OECD 1991(figure 1). •  There is also a tiny cohort countries. •  There is a large cohort of about (mainly Western Saharans) that has From an operational viewpoint, the 8.9 million “recent refugees,” who ar- been in exile for up to 55 years. type of support needed in protracted rived over the last 4 years. It includes Second, as of end-2015, the median situations, as opposed to relatively about 4.8 million Syrians, as well as duration of exile is 4 years—meaning short crises, differs. Where the crisis people fleeing from South Sudan (0.7 half of the refugees worldwide have is short, humanitarian aid may suf- million), Afghanistan (0.3 million), spent 4 years or less in exile. The me- fice; when it lasts, development in- Ukraine (0.3 million), the Central dian has fluctuated widely since 1991, terventions are required. This means African Republic (0.3 million), and between 4 and 14 years, and is now at that the very nature of the needed Pakistan (0.2 million). a historical low. By contrast, the aver- response to a refugee crisis is deter- •  Another large cohort, of about age duration is 10.3 years, and has mined by the (expected) duration of 2.2 million, has spent between 5 and been relatively stable since the late the crisis. 9 years in exile. It includes refugees 1990s (between 10 and 15 years). Using data from the UN High from Afghanistan (0.5 million), the Third, while the number of refugees in exile for over 20 years has been re- Figure 1. More than Half of Global Refugees Have Spent 4 Years or Less in Exile markably stable since 1991, at 5 to 7 Number of refugees by year of exile million throughout most of the period, the average duration of exile increases 3 over time—largely because of the lack of a solution for Afghan refugees. Number of refugees (million) 2 1 Xavier Devictor and Quy-Toan Do, 2016. “How 0 Many Years Have Refugees Been in Exile.” Policy 0 20 40 60 Research Working Paper 7810, World Bank, Duration of refugee status—reference year 2015 Washington, DC. World Bank Research Digest11 Creating a Migrant Rights Database related to family, education, expul- New Migrant Rights Database an objective system of benchmark- sion, asylum, and non-refoulement. enables country contrasts as global ing—all contained in an efficient cod- On the other hand, Russia and Turkey community crafts a global standard ing instrument that can be applied score lowest at 65.7 percent and 66.4 cross-nationally and over time. In percent, respectively. for migration governance and human this pilot study by Gest and Wong, Interestingly, disaggregating the rights protection the authors apply the database’s 65 scores by specific categories of rights W binary indicators to five important reveals greater nuance and variation. ith the signing of the Global destination states: Germany, Mexico, For example, Turkey, one of the lowest Compact for Migration, inter- the Russian Federation, South Africa, scoring countries, features the most national policy makers and and Turkey. complete protection of migrant rights activists are pursuing a global stan- related to labor, while South Africa dard for migration governance and the Rights Provisions Vary Across and Mexico, the two highest scoring protection of migrant rights. However, Countries countries, are below the international there remains a shortage of data about The results reveal variations within baseline for rights associated with vic- the baseline of international legal and across the selected countries tims of crime and due process rights. standards and the extent to which this (figure 1). Among the countries ana- The greatest variation appears in cat- baseline is reflected in the laws of dif- lyzed, South Africa and Mexico score egories related to the rights of legal ferent destination states. the highest at 73.4 percent and 72.3. personhood, vulnerable migrants, The Migrant Rights Database seeks percent, respectively (out of a total family, expulsion, asylum, and non- to address this gap. It applies a novel possible score of 100 percent) (table refoulement. None of the selected instrument to create an objective, 1). South Africa offers strong protec- countries protect migrant rights to cross-national accounting of the laws tions of rights relating to vulnerable civic and political participation. protecting migrant rights in national migrants; life; nationality; and free- The pilot database demonstrates legal frameworks. The database per- dom of thought, opinion, and assem- its use beyond comparative analysis. mits aggregation, disaggregation, and bly. Mexico offers strong protections It is also a means of uncovering and highlighting where derogation takes Figure 1. Migrant Rights Database place and for which specific rights. Comparative and global mean spider plot Further, the results reveal the areas in which there is a near cross-national Vulnerable Migrants consensus on the protection of mi- 100% Equal Protection Life grant rights. From the initial coding, 90% promising areas of law include rights 80% Liberty Education associated with nationality; freedom 70% of thought, conscience, and religious 60% belief; freedom of opinion and expres- 50% Health Legal Personhood sion; labor; and equal protection. 40% With an eye toward strengthen- 30% ing rights protections, future work 20% has the potential to shed new light Labor 10% Due Process on heretofore undiscovered relation- ships between migration flows, mi- grant rights protections, and positive integration outcomes that not only Civil and Political Victims Of Crime improve the lives of migrants but also Life improve societies more generally. This research works to fulfil Sustainable Development Goal 10.7—“to facilitate Peaceful Assembly Expulsion, Asylum, and Association and Non-refoulement orderly, safe, and responsible migra- tion and mobility of people, including Opinion And Expression Nationality through implementation of planned and well-managed migration poli- Thought, Conscience, and Religious Belief Family cies”—and it responds to calls for the Germany Mexico Russian Federation collection of better cross-national South Africa Turkey Global data. (continued on page 12) 12World Bank ResearchDigest (continued from page 11) Recent Policy Research Working Papers on Global Migration 7554 Why Are the Elderly More Averse to Table 1. Overall Measures of Migrant Rights Protection Immigration When They Are More Likely to In percent Benefit? Evidence across Countries Simone Schotte and Hernan Winkler Category Russian South 7619 Inventor Diasporas and the Internationalization of Technology (number of indicators) Germany Mexico Federation Africa Turkey Ernest Miguélez Vulnerable migrants (3) 67 100 33 100 100 7740 Aging, Trade, and Migration Richard Chisik, Harun Onder, and Dhimitri Qirjo Life (3) 33 100 33 100 33 7752 Should I Stay or Should I Go? Internal Migration and Household Welfare in Ghana Liberty (6) 67 67 83 67 67 Vasco Molini, Dan Pavelesku, and Marco Ranzani Legal personhood (4) 75 50 88 88 75 7852 Global Talent Flows Sari Pekkala Kerr, William Kerr, Çaglar Özden, Due process (4) 75 50 75 75 75 and Christopher Parsons 7875 Global Migration of Talent and Tax Incentives Victims of crime (2) 100 100 50 50 100 Evidence from Malaysia’s Returning Expert Program Expulsion, asylum, and non- 81 93 57 79 71 Ximena V. Del Carpio, Çaglar Özden, Mauro refoulement (8) Testaverde, and Mathis Wagner 7880 Transit Migration: All Roads Lead to America Nationality (5) 90 80 90 90 70 Erhan Artuc and Çaglar Özden 8014 Can Parental Migration Reduce Petty Family (4) 63 75 0 63 38 Corruption in Education? Lisa Sofie Höckel, Manuel Santos Silva, and Freedom of thought, conscience, 100 100 100 100 100 Tobias Stöhr and religious belief (1) 8034 Migration and Cross-Border Financial Flows Maurice Kugler, Oren Levintal, and Hillel Freedom of opinion and 100 100 100 100 100 Rapoport expression (2) 8081 A Review of the Anthropological Literature on the Civil Service Freedom of peaceful assembly 0 50 50 100 50 Colin Hoag and Matthew Hull and association (2) 8492 The Effect of Immigrant Concentration at School on Natives’ Achievement: Does Civil and political life (2) 0 0 0 0 0 Length of Stay in the Host Country Matter? Laurent Bossavie Labor (12) 73 63 73 71 79 8817 How Mass Immigration Affects Countries with Weak Economic Institutions: A Natural Health (4) 50 50 50 0 0 Experiment in Jordan Alex Nowrasteh, Andrew C. Forrester, and Education (2) 50 100 50 0 0 Cole Blondin Equal protection (1) 100 100 100 100 0 Aggregate mean (65 indicators) 69 72 66 73 66 Justin Gest and Tom K. Wong. 2018. “Migrants Rights Database.” KNOMAD Working Paper 34. World Bank, Washington, DC. The World Bank Research Digest is a quarterly publica- The Research Digest is financed by the World Bank’s Editorial Committee: Shanta Devarajan, David tion disseminating findings of World Bank research. Research Committee and managed by DECSO, the Rosenblatt, Aslı Demirgüç-Kunt, Roberta Gatti, The views and interpretations in the articles are those Strategy and Operations unit of the Development Hiau Looi Kee, Chorching Goh, and Shiva S. Makki of the authors and do not necessarily represent the Economics Senior Vice Presidency (DEC). The Research (managing editor). Editor: Laura Wallace; production: views of the World Bank, its Executive Directors, or the Digest is not copyrighted and may be reproduced with Roula Yazigi. countries they represent. appropriate source attribution. http://www.worldbank.org/researchdigest. The World Bank 1818 H Street, NW Washington, DC 20433, USA Printed on Recycled Paper