Migration and Jobs: Issues for the 21st Century

With an estimated 724 million extreme poor people living in developing countries, and the world's demographics bifurcating into an older North and a younger South, there are substantial economic incentives and benefits for people to migrate. There are also important market and regulatory failures that constrain mobility and reduce the net benefits of migration. This paper reviews the recent literature and proposes a conceptual framework for better integration and coordination of policies that can address the different market and regulatory failures. The paper advances five types of interventions in need of particular attention in design, implementation, and evaluation; namely, (1) active labor market programs that serve local, regional, and foreign markets; (2) remittances and investment subsidies to promote job creation and labor productivity growth; (3) social insurance programs that cover all jobs and facilitate labor mobility; (4) labor taxes to internalize the social costs of migration in receiving regions; and (5) more flexible private sector driven schemes to regulate the flow of migrants and minimize irregular migration.

the old. 6 Climate change, and the resulting increase in environmental degradation and natural disasters, as well as conflict, add further pressures. Much of these will manifest themselves in increased internal movements, including permanent relocations (World Bank, 2017b).
There are, however, market and regulatory failures that affect the (explicit and implicit) cost of migration as well as the potential to reap the benefits at destination. With respect to market failures, migrants often do not have information about specific job opportunities, they lack access to credit and risk management instruments to manage and finance travel and relocation costs, and the skills they have are not necessarily the skills needed in regional or international labor markets. Migration decisions further bring externalities because they affect the broader communities from where the migrants originate and where they arrive, positively and negatively. In receiving communities, migrant inflows can replenish the labor supply and ease fiscal strains on the pension system, but they also increase competition for local jobs and therefore search costs, stretch public services including security and law enforcement, and can create social tensions. 7 In sending communities, on the other hand, an excessive loss of human capital can ("brain-drain") compromise local development and deepen regional disparities. 8 These effects are not considered by migrants or employers in destination countries, nor local investors.
From a regulatory point of view, requirements in terms of entry conditions and documentation to travel across countries (and sometimes regions within countries) can also constrain labor mobility and migrant flows. Together, market and regulatory failures can lead to inappropriate risk-taking, sub-optimal migration flows, and not enough remittances and local investments. Governments have an important role to play in addressing these failures. Provided that migration flows and associated externalities are managed well, the movement of labor can be an important force to increase economic efficiency and shared prosperity-like the movement of capital and goods. Yet, this is not automatic and contrary to what has happened with capital (foreign direct investment) and goods and services (trade), facilitating labor mobility and managing associated externalities has been somewhat of a second-order priority in the agendas of governments and international organizations. 9 This paper reviews the recent literature in the area of migration and proposes a conceptual framework and research agenda to better integrate and coordinate policies for addressing the different market and regulatory failures that affect the main drivers of migration flows: expectations about job opportunities in receiving regions; expectations about job opportunities locally; and the costs (economic and social) of moving to and entering labor markets in another 6 Even in Japan, labor shortages are partly met through immigration. In the 1990s and 2000s, immigrants were mainly descendants of past Japanese emigrants (so called "Nikkei" Brazilians and Peruvians) working for low wages in factories struggling against Chinese competition. More recently immigrants come from Vietnam, China and the Philippines, on student and work experience visas, working in convenience stores and fish factories. Migrants have accounted for roughly 10 to 15 percent of Japan's net job creation since 2013 (Financial Times, 2 August 2017). 7 The arrival of a large pool of new migrants is feared to increase crime, threaten national security, and, especially, erode social identity (World Bank, 2017a). 8 See Artuc et al. (2015) for a discussion. 9 Hickey (2016). While world foreign direct investment (FDI) and world imports as a share of GDP rose by a factor 7.7 and 2.4 respectively, the share of immigrants in the world's population increased only by a factor of 1.4 from 2.4 percent in 1960 to 3.3 percent in 2015. When the partition of India and the dissolution of the Soviet Union are accounted for, migration between developing countries is remarkably stable (World Bank, 2016b). region or country. The paper discusses five types of interventions that will need to receive particular attention in terms of design, implementation and evaluation: 1) active labor market programs that serve local, regional and foreign markets; 2) remittances and investment subsidies to promote job creation and labor productivity growth; 3) social insurance programs that cover all jobs and facilitate labor mobility; 4) labor taxes to internalize the social costs of migration in receiving regions; and 5) more flexible, private sector driven schemes to regulate the flow of migrants and minimize irregular migration. The paper also discusses the needs in terms of data and new research, particularly when it comes to domestic migration (including in response to environmental change).
The paper proceeds as follows. A brief review of the different types of migration (Section 2) is followed by the discussion of a conceptual framework linking the migration and jobs agendas (Section 3). Section 4 reviews the evidence about the impacts of key policy instruments aiming to facilitate migrations. Finally, Section 5 concludes with an agenda for future research and policy analysis.

MIGRATION TYPOLOGIES
There are different "types" of migration, each requiring different types of policy interventions to address market and regulatory failures. Domestic and international migration flows can be characterized by their drivers (forced vs. economic migration), their duration (permanent vs. temporary), their stance vis-à-vis the law (regular vs. irregular), the basic characteristics of the migrants (age, gender, skills), and the geography of the flows (see Table 1). These profiles or forms determine how migration affects job opportunities for migrants as well as the sending and receiving communities. Unfortunately, despite much progress, the data needed to document the migration flows accordingly and inform policymaking are not always available. With much of migration in violation of immigration law or internal mobility restrictions, knowing the total number of migrants, for example, remains a challenge, let alone the disaggregation in its different forms. There are, nonetheless, general patterns that emerge.
We know that most migration is voluntary, driven by economic reasons, and that internal migration is much larger and a more important driver of labor productivity growth and poverty reduction than international migration. While the latter attracts the most attention (in the literature, media and policy making), 10 the stock of internal migrants is estimated to be three times larger. Similarly, the number of internally displaced people (IDP) due to conflict is twice the number of international refugees. We know little, however, about the legal status of most migrants.
Regarding duration, much migration is temporary (internal and international)-driven by seasonal or circular jobs and constituting part of a livelihood system. Furthermore, contrary to common perception, migration is often a journey, with several stops, including return, and not all of it successful in terms of job outcomes (Lucas, 2016;Artuc and Özden, 2017;Ingelaere, et al., 2017). Little remains known about the implications of this cumulative nature of migration for understanding migration processes, as well as the extent and impact of return migration on migrants and migrant communities internally. 11 Yet, to optimize the benefits from migration, all stages and path dependency of migration must be considered.
A third dimension in considering the potential of migration concerns the characteristics of migrants. Male migration (internally and internationally) still dominates, though gender patterns and norms are shifting. Women are not only migrating more than before, 12 they also migrate more for economic reasons (domestic work and care economy), especially in Asia, with Filipino women for example outnumbering men. There appears a stigma against "autonomous" migration of women, with the percent of never married women migrating from rural-to-urban areas on average greater in most regions than the percent of never married men. 13 Overall, a systematic documentation and study of women's mobility and the different challenges female migration brings for all stakeholders involved, remains largely lacking, especially in Sub-Saharan Africa. This lacuna is the result of an enduring paucity of national-level data for the study of sex-specific migration patterns (Camlin, Snow and Hosegood, 2014).
The skills of migrants also affect migration patterns. For instance, migration by semi-skilled and unskilled workers is more often temporary (e.g. seasonal agricultural or construction workers), while migration by skilled people is more likely to be permanent and further away (Özden, 2017). 14 Not surprisingly, the level of skills also affects the destination. Skilled migration rates are particularly high in Africa, with one of every eight Africans with a university education having lived in a country in the OECD (World Bank, 2011). These migration typologies should drive differences in policy objectives as the constraints facing different groups of migrants differ (Cho, 2017). 15 Unfortunately, systematic information on the relative importance of the three different groups is lacking. 11 Hirvonen and Lilleor (2015) are a notable exception. Studying the patterns and determinants of return migration within Tanzania, they find that selection in return migration is associated with unsuccessful migration spells. For men, these are linked to poor job-market outcomes at the migration destination; for women, return was linked to the ending of marriages. After the return, the returnees do not seem to fare better than non-migrants. 12 Domestically, based on the experience in 42 countries, and controlling for age, education, marital status and ethnicity, women are not less likely to move from rural to urban areas (Lucas, 2016). There are also differences across regions. Without controlling for these covariates, female rural-urban migration dominates male rural-urban migration in many Latin American countries by 2-8 percentage points. The opposite holds in Sub-Saharan Africa. 13 Qualitative case study analysis of migration histories in Tanzania points in the same direction. Women there also increasingly migrate for economic reasons, but the journeys are more organized and arranged beforehand than in the case of male migration (Ingelaere et al., 2017). 14 At the individual level, the skilled are likely better placed to handle the greater costs of moving further away. Locations further out are also costlier to return from, making them more permanent. These are average trends however, and in some instances the reverse is observed. In Romania, for instance, it is the skilled that have been leaving the countryside for the cities within Romania, while the unskilled are moving west, looking for employment opportunities in other countries in the European Union (Farole, Goga, Ionescu-Heroiu, 2018). 15 With forced migration, for example, the policy focus is often primarily humanitarian and the timeframe short term. It tends to shift from control to co-development when it concerns regular voluntary and temporary migration (often of unskilled or semi-skilled workers). This is at the core of many managed migration programs with a focus on mutual economic benefits. It is also especially in this space that coordination failures abound and much scope for improvement remains. When it comes to permanent workers, often related to a skilled workforce, the policy focus widens further to fostering co-existence with attention to economic and social cohesion.

Geography
Secondary town/city S-S (38%); S-N (33%); N-N (23%), N-S (6%); Domestically, the focus has been on rural-urban migration, without much distinction between urban areas. Yet, incipient evidence suggests that secondary towns may be more within the reach of the poorer segments of the population than large cities, and that secondary town development may hold more promise to improve job opportunities and for poverty reduction than metropolitanization. This is an area for further investigation with important policy implications for investment (Gibson, et al., 2017;. More broadly, a systematic documentation and incorporation of internal migration patterns in national development strategies (rural-urban, but also urban-rural, urban-urban and rural-rural, and return migration) remains lacking. More systematic use of the census data could help address this lacuna.

ECONOMICS OF MIGRATION AND JOBS
The policy framework that we develop in this section is based on the spatial equilibrium model developed in Rosen (1979) and Roback (1982). The model assumes that workers move in geographic space to equalize "livelihoods" over time. Labor income is a key determinant of these livelihoods and therefore migration decisions are affected by the geographic distribution of job opportunities. 16 At any time, three key factors enter in the individual decision to migrate: 1) their expectations about job opportunities in the area(s) of destination; 2) job opportunities in the area(s) of origin; and 3) the costs associated with moving and realizing the prospects in the destination. If the expected gains in labor income net of migration costs are positive, people are expected to migrate. 17 This continues until a spatial equilibrium is established. 18 This model provides a benchmark to assess the efficiency of the distribution of workers across space and, by extension, the efficiency of the policies and regulations shaping this distribution. 19 In the absence of market or regulatory failures, there would not be a need for government intervention. In practice, however, there are several failures that affect the three factors driving migration decisions and that call for integrated/well-coordinated policy interventions (see Table  2). A first set are classic imperfections that affect labor, capital, and insurance markets and which are the usual focus of migration policy: 1) workers might not have sufficient information about the geographic distribution of job opportunities or the information they have is biased; 2) the distribution of skills across workers might not match the distribution of skills required by 16 Although our discussion puts jobs at the center of the migration decisions, in the New Migration Economics literature (Lall, Selod, and Shalizi, 2006), livelihood prospects are conceived broadly. They depend on economic prospects, such as the likelihood to be employed and earn a good living, build up assets and save for the future, but also on access to social goods and services (education, health), which are typically place-bound, as well as personal security, and psycho-social well-being. 17 Migration costs include the monetary cost of moving (transport), settling in (e.g. rent), and finding a job, but also the psycho-social costs of social adaptation and integration and regulatory barriers. 18 It is important to note though that while migration is often seen as a one-shot event, it is in effect a process of cumulative causation, with people's action space (the actual set of destinations within a person's reach) evolving (expanding/contracting) over time. This induces individuals to regularly reassess their "migration-equation", making return-migration also an important reality (Ingelaere et al., 2017), and an integral, though somewhat neglected, part of migration management. 19 Forced migration, in response to sudden (or slow onset) shocks (natural disaster or conflict), can also be understood within this framework. Such events dramatically reduce the prospects for economic and personal security in the current location, creating a gap, often suddenly, where there might have been a spatial equilibrium before. existing job opportunities; 3) liquidity/credit constraints can affect workers' ability to finance migration costs and move; and 4) workers might not have access to sufficient savings and/or insurance to deal with the uncertainty that surrounds the decision to migrate and associated risks. Standard responses to these market failures include active labor market programs that provide counseling and intermediation services, training, and job search assistance; mobility grants or guaranteed credit programs; and social insurance programs such as pensions, health and unemployment benefits.
Second, job externalities affect both the efficient distribution of local job opportunities and the net costs of migration and are usually not taken into consideration when designing migration policies. There are two types of job externalities: labor externalities that emerge in the presence of high unemployment and/or underemployment rates when market wages do not reflect the real opportunity cost of labor; and social externalities that appear when jobs bring value to society beyond that captured by workers and employers (Robalino and Walker, 2017). A classic example of the latter is that workers accumulate human capital in their jobs and, like in the case of education, they make other workers more productive. In lagging regions, in the presence of high jobs externalities, there might be investments with high potential for job creation that do not take place because private rates of return are not high enough relative to alternatives (like bank deposits or investments in urban areas with a large concentration of human capital and better infrastructure), yet the social rates of return can be substantial. This would lead to a sub-optimal distribution of local jobs and, potentially, migration flows that are too large. 20 Alternatives to facilitating labor mobility to close the income gap (and avoid negative externalities from migration flows and stocks) are to increase the livelihood prospects in the location of origin, including through foreign aid, trade and capital mobility as well as fostering return migration. 21 Social externalities also affect migration outcomes. Workers deciding to migrate do not take into account the costs they impose on the local economy by reducing the stock of human capital; their current level of human capital and the present value of future accumulations had they taken a local job. 22 Remedies to address the effects of jobs related externalities include subsidies to promote local investments, policies to incentivize and, in some case tax, remittances, and active labor market programs to improve the functioning of local labor markets.
Finally, migrants do not consider the social costs they might impose on other workers in destination regions. These costs include tightening labor markets that can result in lower wages and/or longer unemployment spells, traffic, and congested public services including law enforcement and security. Governments usually attempt to deal with these problems by directly managing flows through visa regulations and border control in destination regions, and/or through different types of bilateral or multi-lateral agreements. But there are other mechanisms that can help internalize the social costs of migration in destination regions such as labor taxes or fees levied through work visas or work permits. 20 When prospect gaps mainly reflect gaps in basic amenities, it is typically more efficient to improve amenities locally (World Bank, 2009). 21 In many (donor) countries, the mindset is shifting from "What (international) migration can do to foster development?" to "What (international) development can do to curb (irregular) migration?" (Carling and Talleraas, 2016), re-emphasizing the role of place-based approaches. 22 For example, when the more able and educated systematically move to urban areas, where the demand for skilled labor could be higher, while the less skilled concentrate in rural areas (Young, 2013), this will tend to increase the prospect gaps. What is the extent of market (and regulatory) failures that affect migration?
Labor mobility. Data on income or wage gaps across regions suggest that, in general, countries are very far from the spatial equilibrium discussed above; there is not enough labor mobility. 23 Gaps are especially large across countries--unskilled workers' wages in rich and poor countries often differ by a factor of 10 to 1. Differences in average incomes between countries explain more than 80 percent of global income inequality (Milanovic, 2012). Such large differences cannot be explained by migration costs alone (transport, jobs search and settlement costs) and suggest large individual gains from migration. Many of the market imperfections discussed above (lack of information liquidity and credit constraints) and regulations (immigration laws, border control) are likely to stand in the way. 24 These large observed income gaps-the perceived gaps are often even larger-and the practical challenge to control labor flows across borders also induce irregular migration, trafficking and labor exploitation.
This brings a potentially important role for global public policy to address market and regulatory failures. There is likely ample room to facilitate labor mobility so that workers can have access to better job opportunities while managing the footprint that migrants leave on their countries and communities of origin and destination (see further below). There will likely be gainers and losers, but the gains seem substantial to enable compensating the losses so that all involved benefit (the migrants themselves, their families as well as the sending and receiving countries and communities).
Income gaps between rural and urban workers within countries are about a factor of two to three. 25 These are much lower than the international gaps, but the actual costs of migration are also lower. Many policy analysts and academics interpret these gaps as indicative of significant market imperfections (to a lesser extent, regulatory impediments). Thus, there appear appreciable gains from removing barriers to within country labor mobility as well. 26 A number of recent papers suggest, however, that the gaps dramatically decline or quasi disappear when differences in human capital are accounted for and that they mainly reflect selection. 27 This would suggest that one of the main constraints facing workers in lagging regions is their skills. Mapping out spatial inequalities in domestic labor market outcomes (income, wages, employment) and amenities and the reasons behind it (market failures, regulations, selection) is satisfaction with personal living standards (including wealth). The former was also slightly higher than the latter in Asia (38 versus 30 percent), but similar in Latin America (around 38 percent). Differences in area and personal security (conflict and crime) are another motivator, explaining between 10 (Sub-Saharan Africa) and 25 percent (Asia) of the variation in the stated migration intentions among the prospective migrants. For a qualitative analysis of the different factors underpinning (internal) migratory decisions and trajectories, see Ingelaere et al. (2017). 24 Clemens, Montenegro and Pritchett (2016) estimate for example a real wage gap ratio of 6.8 between observationally equivalent migrants from 42 developing countries and the United States. The size of this gap, they show, points to important policy barriers hindering spatial equilibration, as spatially integrated labor markets in the absence of policy barriers rarely sustain real wage gaps ratios above 1.5, even in the presence of important cultural and geographic barriers. 25 Gollin, Lagakos, Waugh (2014); Christiaensen and Kaminski (2014); Christiaensen and Premand (2017);and McCullough (2017). 26 Restuccia, Yang and Zhu (2008); Bryan, Chowdhury and Mobarak (2014); and Munshi and Rosenzweig (2016). Regulations tend to be less binding in determining internal migration, though in some cases they are important (e.g. hukou system in China). 27 Young (2013), and more recently McCullough (2017) and Hicks et al. (2017). Young shows, for example, that rural-born individuals with more years of schooling than average in their sector are more likely to move to urban areas, while urban-born individuals with less schooling tend to move to rural areas. He develops a model which assumes that there is more demand for skilled labor in urban areas. This could generate two-way flows of the kind he documents. He then argues that this can fully explain the observed urban-rural consumption gaps, once sorting is accounted for.
an important area for further inquiry. 28,29 More research is also needed to better understand the costs 30 and benefits to internal migrants (including return migrants), and the role of social networks and secondary towns. 31 When discussing labor mobility, it is also important to consider the gender dimension that affects migration decisions and outcomes. This holds for voluntary migration, as discussed above, but even more so for forced migration (World Bank, 2017a). Forced displacement can transform gender norms. It may create opportunities for education and access to employment opportunities, as for rural Afghan women who are refugees in the Islamic Republic of Iran (Monsutti, 2006). Women's skills may also be better transferable than men's, enabling them to find work more easily, for example as domestic workers (Buscher, 2009). Yet, women and girls are also at greater risk of physical violence (rape, sexual abuse) during the flight and at destination, also within the household, with displaced Afghan women reporting domestic abuse happening more frequently during displacement than before (JIPS, 2012a).
Local Job Opportunities. The negative social externalities in origin regions associated with migration are likely to be more than compensated by the benefits resulting from reduced unemployment, poverty reduction (when the migrants themselves are poor) and, indirectly, through remittances, diaspora investments, skill and technology transfer and increased trade and foreign direct investment (FDI). Remittances (international and internal) are an important way of sharing prosperity between different places and the importance of international remittances (more than three times the size of official development assistance) and their contribution to development has attracted a great deal of attention over the past 10-15 years. 28 Bryan, Chowdhury and Mobarak (2014) show how a small travel grant or credit (randomly distributed) to poor rural laborers in Bangladesh (conditional on seasonal migration) can induce recipients to move to towns and cities for temporary work during the agricultural low season. While indicative of market failures, at the same time, uptake remains moderate (one-fifth of the recipients) (Hicks et al., 2017). The experiment is now replicated and evaluated at a larger scale (https://www.evidenceaction.org/thesolution-no-lean-season/). 29 In a more dynamic perspective, spatial inequality often forms a fertile breeding ground for conflict, especially when it coincides with differences in ethnicity, religion or language (Langer and Stewart, 2015). Conflict in turn reinforces the migratory pressures from spatial inequality. 30 As with international migration, rural-urban transport costs are often only a fraction of the annual income gain that could be obtained from moving . Schwartz (1973) emphasized this early on, showing that the adverse effect of distance on migration follows especially from diminishing-information (i.e. rising jobs search costs), and less likely from actual transport (or even psychic costs). 31 Migration costs typically increase with distance, and thus also the likelihood that liquidity and credit constraints bind. This is one of the reasons why secondary towns may attract more rural migrants, than cities, even though the potential income gains are larger in the latter. Migration costs can be also partly overcome through social networks, by reducing the monetary and informational constraints as well as the psychological costs (Seshan, 2017). This explains the important role of migration networks in understanding migration flows and the choice of destination. Yet, the likelihood of these networks establishing is itself a function of the distance to the destination and the existence of market failures, reinforcing the role of secondary towns in mediating migration (Ingelaere et al., 2017). Alternatives to facilitating labor mobility to close the income gap (and avoid negative externalities from migration flows and stocks) are to increase the livelihood prospects in the location of origin, including through foreign aid, trade and capital mobility as well as fostering return migration. Indications are that moving also "pays" in times of conflict, with those leaving tending to suffer less (physically and economically) than those who stayed behind (Etang Ndip, Hoogeveen andLendorfer, 2015: World Bank, 2017a). But the poorest often do not have the means or social connections to leave.
While the results on the effects on overall job creation and growth remain so far inconclusive, at the micro-level remittances are found to reduce poverty and increase spending on education and health (especially among the better-off households) as well as housing (UNDP, 2009;World Bank, 2011). They also provide livelihood and insurance strategies in the event of sudden shocks and slow-onset disasters, substituting for less efficient financial systems both in terms of ex post responses to disasters and in terms of ex ante risk management strategy (Bettin and Zazzaro, 2017).
Over one-fifth of high-skilled workers from low-income countries, and over 40 percent of those from small island nations have migrated to high-income countries (Artuc, et al., 2015). But remittances, spillovers from return migration (brain circulation) and increased domestic demand for higher education help offset the loss, often resulting in a "brain gain" (Mattoo, Neagu, and Özden, 2008). On balance, the effects for the sending areas are considered quite positive (World Bank, 2016b provides a review).
This does not mean, however, that there are no losers. The welfare and labor market effects of internal migration on the households and communities of origin, and how these could be improved, has received less attention (China being a noteworthy exception). The same holds for the effects on the communities of origin of forced migration in response to conflict. Therefore, it remains important to better understand and address the effects of jobs externalities on local investments and job opportunities. As discussed above, when jobs externalities are not taken into consideration by the private sector local investments, the number and distribution of jobs, can be below socially efficient levels. The loss of human capital (particularly skilled workers) adds to this problem by constraining fiscal expenditures given forgone tax revenues, further reducing investment opportunities, and depressing private rates of return. The solution is not to stop or regulate migration but, instead, to build on the opportunities that migration creates to promote local investments and job creation. Not doing so would not only be a missing opportunity to promote local economic development and improve job opportunities for those who stay behind, 32 it can also lead to an increase in inequality in the sending areas, 33 and in some cases, macro-economic imbalances such as the Dutch disease.
The suggestion therefore is to coordinate/integrate migration policies with local jobs strategies. A critical aspect would be to improve the management of remittances to further reduce the cost of remitting and to enhance their development impact by increasing fiscal revenues (for the production of local public goods), promoting financial innovation, and creating linkages to local investment opportunities and job creation. To this end it would be necessary to revamp local active labor market programs (including training) and to develop "demand-side" interventions such as value chain development and entrepreneurship programs (Datta et al., 2018) that aim to equate private and social rates of return on investments (Robalino, Romero and Walker, 2019). It would also be important to better coordinate skilled outmigration between sending and receiving countries, including for example through co-investment in education (see Box 1). 32 Those staying behind also face psycho-emotional costs that are often not considered in welfare analysis of migration. 33 With better-off households more likely to send migrants abroad (UNDP, 2009), and migrants from better-off households more likely to be successful and send remittances, remittances may further concentrate resources in the hands of the better-off. This might hold, in general, but also in the context of disaster, thereby reproducing the inequalities and vulnerabilities existing within the community of origin (Le De, Gaillard and Friesen, 2013;2015a, b).

Social costs in destination areas.
Internationally, host countries benefit substantially from migration. Migrants replenish the labor supply (at both ends of the skill spectrum), 34 they spur entrepreneurship and innovation and ease fiscal strains on the pension system (UNDP, 2009;World Bank, 2016b). Domestically, to the extent that population density (urbanization) comes along with agglomeration economies, it also fosters economic growth and employment generation, translating into economic density (Duranton, 2014).
In general, the net impact of migration on receiving labor markets seems to be small but this does not preclude that some workers can be negatively affected, at least during an adjustment period. A recent meta-analysis of seven studies, all of high-income countries, shows that increasing the share of immigrants 1 percentage point in a local labor market results in declines in native employment of just 0.011 percent and in native wages of 0.029 percent (Longhi, Nijkamp, and Poot, 2010). Docquier, Özden, and Peri (2014) provide estimates of the impact of immigration on wages and employment in OECD countries in the 1990s. They find positive, but small, impacts across countries on both less-educated and all locals even using pessimistic scenarios. Nonetheless, it is important to recognize that there can be transition costs as labor markets adjust and new investments, innovation, and job opportunities materialize. Behind average effects, there can be groups of workers (e.g., older and less educated workers, or semi-skilled workers in certain occupations) that might lose jobs to competition, face longer unemployment spells, or see a decline in their earnings. Recent work in Malaysia suggests that immigration increases employment overall but may have a negative impact on the employment of less-educated locals (del Carpio et al. 2015). In Thailand, negative wage impacts from doubling the size of the immigrant workforce in a province are isolated to less-skilled native workers. 35 Congestion of public services and fiscal costs of social services are also frequently cited in policy discussions. Net fiscal impacts of migrants tend to be slightly positive, but the effects on social services depend on the rights and access granted. Research for OECD countries finds, not surprisingly, that impacts are more favorable when labor migration dominates, when immigrants have higher employment rates, and when immigrant populations are relatively young (OECD 2013). This is because younger migrants tend to be healthier and use fewer services while employed migrants tend to use fewer services and contribute more in taxes and social security.
The effect of migration on crime depends, in part, on labor market outcomes. In general, fears that immigrants are disproportionately more involved in criminal activities than natives do not seem to be justified by the existing literature; on the contrary immigration has often led to a reduction in crime. For instance, recent research from Malaysia finds that immigrants reduce crime (Özden, Testaverde, and Wagner, 2018). These results are in line with international 34 By way of example, migrants make up 16 percent of the U.S. labor force, but are concentrated in certain sectors. They make up most farm workers and 60 percent of construction related occupations. At the other end of the spectrum, one-third of physical scientists, engineering professors and doctors are also foreign born. 35 And even that is not always easy to establish robustly (Pritchett, 2017). Insights from three country case studies (of Kenya, Tanzania and Uganda) suggest that the local economy often benefits from the influx of refugees, through increased demand for local goods and services and better connectivity following investment in new roads and transport services to reach the camps (Verwimp and Maystadt, 2015). But not everyone benefits. The landless and agricultural workers, whom refugees may compete with on the labor market, and net food buyers suffer, at least in the short run.
evidence showing that immigration has no effect on violent crime and no, or only a small effect, on crime (Adelman 2017;Bell, Fasani, and Machin, 2013;Bianchi, Buonanno, and Pinotti, 2012;Spenkuch, 2013). The instances where immigration contributes to crime tend to be linked to cases where migrants are not able to work. For instance, an increase in the number of asylum seekers who are generally prevented from seeking employment can result in a slight increase in property crime (Bell, Fasini, and Machin, 2013).
Probably issues related to national security and loss of national identity pose the more powerful, but also the more intractable reasons against a liberal migration policy . 36 Indeed, the literature indicates that non-economic factors such as racial prejudice and concerns about crime and cultural impacts are correlated with attitudes towards migration. 37 Direct surveys of attitudes in the United States and Europe show that both low-and high-skilled locals have a preference for high-over low-skilled immigrants, which runs contrary to the predictions of the labor competition model. Rich and poor locals are also found to be opposed to low-skilled immigration (Hainmueller and Hiscox, 2010). Mindful of backlashes against free trade, the concerns and voices against labor mobility in the host countries cannot be ignored. The political economy and real reasons behind them must be properly understood and addressed, and winners and losers duly identified, acknowledged and compensated. This is a recurrent theme in the history of migration (UNDP, 2009;ch2.2).
In terms of domestic migration, the types of concerns are similar. Fast urban population growth challenges cities to keep up their capital base. While Africa's fast urban population growth (4.9 percent during 1960-2010) follows mainly from its rapid urban natural increase (2.9 percent), it is compounded by rural-urban migration. 38 As a result, cities struggle to capture the economies of agglomeration from density (Lall, Henderson and Venables, 2017). High rates of unemployment, strained infrastructure and environmental degradation are widespread instead. It is, however, the natural increase of the urban population, and not rural-urban migration, that is found to be associated with indicators of congestion (Jedwab, Christiaensen and Gindelsky, 2017). 39 How best to manage the speed of urban population expansion is an important policy challenge, with the speed and destination of rural-urban migration (cities versus towns) providing one of the policy entry points. Similarly, signs are emerging that more fertile rural areas in Africa tend to be poorer (Wantchekon and Stanig, 2016), especially when poorly serviced by road infrastructure. With fertile land attracting migrants, overcrowding may result when not accompanied with the necessary investment in infrastructure and market access, an important hypothesis for further investigation. 36 The concerns and evidence are similar when it comes to the impact of protracted refugee situations on hosting communities (Verwimp and Maystadt, 2015

POLICY INSTRUMENTS AND THEIR EFFECTIVENESS
There are a host of policy instruments that have been considered to improve migration that operate largely disconnected and outside a coherent framework to address the market and regulatory failures discussed above. Migration policies are usually organized around the migration cycle: 1) predeparture; 2) during migration, and 3) possible return. 40 Some are aimed at increasing the benefits or mitigating the costs to migrants and their families (e.g. remittances), while others focus on increasing the gains and reducing the losses in the sending or receiving communities (or both) (Table 3). A classification based on the "cycle," which suggests that different policies are required at different points in time, might preclude the emergence of more efficient jobs and migration strategies where, from the "start," different interventions are coordinated to maximize the welfare of migrants and those in origin and destination regions (see Table 2). Still, better integration/coordination requires an understanding of the effectiveness of different instruments. In this section we provide an overview of the main insights from the few studies that have tried to assess policy impacts.
In, general, we know more about international migration. 41 Bilateral migration agreements for countries whose workers have few other migration options (e.g. New-Zealand-Pacific Islands) have been found to substantially increase migration and the income of migrants, while financial education of migrants and their families (pre-departure or at destination) raises their savings and investments. New savings and remittance products that allow migrants more control over how their money is used also increase savings. These studies do not, however, examine the broader, second round effects of these interventions on the sending and receiving communities. -Unilateral actions Hindering migration: Countries with restrictions on rights of women to migrate or requiring all citizens to get government permission to travel have 5 to 6 percent less migrants per capita than countries with similar income, population and governance levels without these restrictions. Higher passport costs are associated with less migration (with the cost of a passport exceeding 10 per cent of per capita income in some countries) Fostering migration: Little effect of unilateral actions to assist people by reducing the information barriers, the frictions in jobs search or documentation barriers.
There is also suggestive evidence about the impact of policies that affect remittances. Lowering the costs of remittances appears to induce relatively large increases in remitting. 42 Costs of remitting have come down a lot already, but there is scope for further reduction (Ratha, 2017). It remains unclear, however, how much of the past reduction in the cost of remitting has been due to an increase in competition in money transmission markets and technological improvements versus policy efforts to increase information for migrants about the costs of different transmission services. 43 There is also evidence that matching programs can increase the development effect of remittances when conditioned on specific objectives (e.g., education of children, investment in local public goods). In particular, there appears to be demand for matching grants for education (see Box 1). Little is known, however, about the potential to link matching grants to local investments and job creation as discussed in Robalino, Romero and Walker (2019).

Box 1: A global remittance agenda
To enhance the development impact of remittances, Ratha (2017) articulates a Global Remittance Agenda around 4 pillars (monitoring, analysis and projection; improving the remittance retail payment system; increasing financial access for households and SMEs; and improved capital market access for countries and companies, including the link between remittances and sovereign credit rating, bonds backed by future remittances, and diaspora bonds).
Overall, remittance costs have come down a lot already, but scope for further reduction remains (Ratha, 2017). Matching programs are further explored to increase the development effect of remittances by affecting the spending pattern, though there is very little evidence. There appears demand for such matching when it comes to education. In an experiment, Ambler, Aycinena and Yang (2015) offered Salvadoran migrants in the United States the opportunity to direct their remittances toward the education of a student of their choice, with the program providing matching funds. Take-up increased from zero without matching funds to 18.5 percent among migrants offered the 1:3 package, i.e. a $3 top up by the program for every $1 remitted to the program. Moreover, beneficiaries increased their educational expenditures by $3.72 for every educational $ they received from the program, pointing to a substantial crowd-in effect. It also increased the likelihood of attending private school and lowered labor supply among the target students.
Yet, not all matching programs are equally successful. The Mexican "Tres por Uno" (Three for One) program encourages Mexican migrants abroad to invest in their communities of origin, with each dollar invested matched by $3 from the Mexican government. Household access to sanitation, water and drainage in participating villages increases (Duquette-Rury, 2014), but these positive effects are somewhat offset by a decline in family remittances and the program can be somewhat regressive, since very poor municipalities have little migration. The Mi Comunidad program on the other hand, using migrant funding to start maquilas, proved a failure (McKenzie and Yang, 2015).
While little attention has been given to understanding the role in origin regions of active labor market programs (e.g., training, pre-departure orientation) to facilitate migration, there are several studies exploring how to improve integration of regular migrants (voluntary and forced) in the receiving countries. OECD, ILO, World Bank, and IMF (2017) provide a review of the experience. Early integration in the labor market proves critical. This requires legal access to the labor market, especially language skills and occupational skill matching, as well as coordination among stakeholders and coherent implementation. Access to housing proves also a challenge, especially for recognized refugees. When it comes to policies enforcing strong economic rights for migrants in the receiving countries, such as high minimum wages, the research suggests the existence of a trade-off, with higher benefits resulting in a decline in the number of migrants benefiting.
The solutions also critically depend on the initial economic conditions of the hosting communities. In developing country settings, lessons from experience suggest addressing longstanding developing issues, which the presence of the forcibly displaced exacerbates, supporting those who have been hurt within host communities, strengthening and expanding social service delivery, and encouraging granting the forcibly displaced the freedom of movement and the right to work (World Bank, 2017a). Overall, rigorous empirical evidence on the effectiveness of different approaches, in both developed and developing host country settings, remains largely lacking, though with many opportunities to learn more systematically, for example today in Europe and the Middle East.
Finally, the role of return migration in managing migration and increasing development outcomes is also increasingly attracting attention. This includes policies to remove regulatory, bureaucratic and informational barriers (e.g. pension portability, go and see visits for refugees), policies changing financial and other incentives to return (assisted voluntary return, permanent return of qualified migrants), as well as policies to make return migrants more productive and adjust reintegration easier (recognition of skills abroad). Except for greater mobility following dual citizenship, evidence on the effectiveness of most of such measures is lacking.
Return migration remains especially an important policy focus of countries/communities hosting forcefully displaced persons. But the costs of ensuring favorable conditions for repatriation of refugees have often been underestimated, 44 while integration in the local labor markets has often been neglected as a solution. This reluctance to integrate refugees locally has often also precluded the opportunity to test local integration approaches. Irrespectively, whether repatriation or local integration is promoted, it is especially the protracted nature of forced displacement that is pernicious. It holds the displaced hostage in short planning horizons. The more rapid the situation can be normalized, the smaller the welfare losses. 45 Early involvement of development agencies, along with humanitarian agencies, rather than a sequential approach, with an initial humanitarian effort response followed by a development effort when the situation becomes protracted holds some promise in this regard (World Bank, 2017a). 46 44 Land tenure and property issues often pose a real challenge for example, when returnees find their land or houses occupied by others (World Bank, 2017a). As Verwimp and Muñoz-Mora (2017) show in their study of returning refugees in Burundi, the plight of (former) refugees does not end upon return. 45 Studying children in Burundi Verwimp and Van Bavel (2013) find that for every additional year children passed in a displacement camp before returning, they were 2 percentage points less likely to finish primary school. Similarly, Verwimp and Muñoz-Mora (2017) show that duration of displacement has long lasting negative effects. 46 World Bank (2017b) and Verwimp and Maystadt (2015) review some of the lessons from the experience with forced migration.
There is little rigorous evidence on the effectiveness of instruments to increase the development benefits of internal migration. 47 The effectiveness of poor area development programs, i.e. interventions which try to improve the conditions in the areas of origin (and thus reduce migration) have been studied (e.g. Chen and Ravallion, 2005;Chen, Mu, and Ravallion (2009) for China), though typically only on their own cost-benefit terms; whether they generated a net benefit, not in comparison with other potential interventions such as those fostering migration.

JOBS AND MIGRATION: A WAY FORWARD
Jobs are at the center of the conceptual framework developed in Section 3. In general, workers migrate within and across countries to improve job opportunities and standards of living, but several market and regulatory failures can affect their ability to move and, when they are successful, the net benefits for origin and destination regions. These net benefits also depend on jobs outcomes. In origin regions, through remittances and incentives to invest in human capital, migration can create a virtuous cycle to promote local development. But in the absence of enough local public and private investments that improve job opportunities, migration can instead depress the economies of origin regions, accentuate income disparities, and lead to social instability. In receiving regions, migrants can replenish the labor supply (at both ends of the skill spectrum), spur entrepreneurship and innovation and ease fiscal strains on social insurance systems. But they can also tighten labor markets with potentially negative impacts on unemployment and/or wages, at least for some groups of workers. Impacts on public services, fiscal budgets, and crime are also mediated by jobs outcomes and migrants' ability to pay taxes and social security contributions.
The main implication of this conceptual framework is the need to integrate and coordinate policies for addressing the different market and regulatory failures that affect the main drivers of migration flows: expectations about job opportunities in receiving regions; expectations about job opportunities locally; and the costs (economic and social) of moving to and entering labor markets in another region or country (see Table 2). 48 We conclude by briefly discussing five types of policies that will need to receive particular attention in terms of design, implementation and evaluation: 1) active labor market programs that serve local, regional and foreign markets; 2) remittances and investments subsidies to promote job creation and labor productivity growth locally; 3) social insurance programs that cover all jobs and facilitate labor mobility; 4) migration related taxes to internalize the social costs of migration in receiving regions; and 5) more flexible, private sector driven, schemes to regulate the flow of migrants and minimize irregular migration. We also discuss the needs in terms of data and new research, particularly when it comes to domestic migration (including in response to environmental change).
Active Labor Market Programs that serve local, regional, and foreign markets. Programs that provide information about jobs in foreign markets, support intermediation and jobs search assistance, transfer the necessary skills, and help reduce the costs of migration and provide funding should not have to be created from scratch. Most countries offer these active labor market programs through their public employment services (PES). Granted, existing programs are not up to the task; they are, often, not able to serve local labor markets and expecting them to respond to the needs of potential migrants would not be realistic. But the same types of reforms that are needed to improve the effectiveness of the programs locally can allow them to expand and serve regional or foreign local markets. The results from recent impact evaluations and meta-analysis suggest five types of reforms: 1) adopting modern identification and registration systems; 2) adopting statistical profiling systems to help assess the main constraints facing migrants to access different labor markets; 3) outsourcing the provision of services to, local and international, private/public providers (for profit and non-profit) on the basis of contracts that pay by results; 4) rethinking financing mechanisms to be able to expand coverage, relying more on contributions from job seekers and employers (possibly including small dismissal taxes 49 ); and 5) implementing rigorous evaluation and monitoring systems. Reformed programs can then serve both migrants and those who stay behind. The latter group is critical. Essentially, services should be available to all workers/job seekers, some of whom will be able to find jobs outside the local market and might then be eligible for additional support such as mobility grants.
Remittances and investment subsidies for job creation and labor productivity growth. There are already important initiatives underway to improve the management of remittances and, in particular, to reduce the cost of remitting (see Box 1). In the presence of jobs externalities, however, just reducing the cost of remitting and increasing flows might not be enough. There is an important policy agenda around how to mobilize remittances to promote investments, job creation, and/or improvements in labor productivity and the quality of existing jobs. Some of the interventions involve subsidizing private investments, in various forms, to align private and social rates of return. 50 This can be done through programs that support value chain development, SMEs, and entrepreneurship. The evidence about potential impacts is still scant but the operations of some of the existing initiatives seem promising. 51 What is missing is to identify the type of financial instruments that can be used to mobilize remittances to fund the private part of the investments. In some cases, those receiving remittances might invest directly. But in many cases, those receiving remittances and those investing are different individuals. Interventions that could be explored and evaluated involve having local financial institutions issue bonds to capture remittances and then fund part of the programs that support investments and job creation (the other part would be funded by public resources).
Social insurance programs that cover all jobs and facilitate labor mobility. Two fundamental problems with existing, Bismarckian, social insurance programs (pensions, health insurance, and unemployment benefits) are that they rely on formal wage employment, and that contributions made by plan members are delinked from the expected cost of the benefit provided. Domestically, the reliance on labor contracts implies that migrant workers transiting between jobs and regions are unlikely to be covered, thus reducing incentives to move or exposing workers and families to undue risks. Many migrants, in fact, are likely to be selfemployed workers who are not covered by the social security. Internationally, the lack of a link between contribution and benefits makes it difficult to transfer acquired rights between social security systems and, in addition, can generate implicit tax/subsidies when workers move from one country to another. For instance, low-income migrants enrolling in a defined-benefit pension 49 These taxes could replace severance pay. When workers are dismissed, employers would make payments to a fund that contributes to finance active labor market programs to re-connect workers to jobs (see Ribe et al., 2010;and Robalino, 2018). 50 See Robalino and Walker (2017) and Robalino, Romero and Walker (2019). 51 For a review of the programs, see Buba and Aterido (2019), Datta et al., (2018), and Packard et al., . system in the receiving country are likely to receive implicit subsidies and, therefore, increase unfunded liabilities (or actuarial deficit of the system). The same with health insurance benefits where, very often, contributions are based on earnings and not on the expected cost of the health package (which depends on the health status of the individual and family and pre-existing conditions/risk factors). An important part of the migration/jobs agenda, therefore, involves improving the design of social insurance programs with the following objectives: 1) expanding coverage to all workers irrespective of work status and type of job; 2) linking contributions to benefits and eliminating implicit redistribution; and 3) funding explicit transfers to vulnerable groups through general revenues (see below). There are today specific proposals to achieve these objectives, including estimates of fiscal costs, that should be considered when discussing and designing migration and jobs policies. 52 Taxes to internalize the social cost of migration in receiving regions. We have seen that the social costs of migration in receiving regions/countries -labor market tightening, congestion of public services, and fiscal costs --are likely to be small and localized but they are not zero. This is particularly the case when migrants have access to social insurance programs that do not link contributions to benefits (see above). As explained in Section 3, migrants do not take these costs into consideration when they decide to migrate. Yet, ignoring these costs can increase the risk of societal rejection of migrants in hosting countries and communities. Managing the political economy of migration might then involve having in place policies that internalize social costs and compensate the potential losers. Like in the case of the dismissal tax discussed above, one option to consider is the introduction of time-bound labor taxes on migrants and their employers that contribute to fund public services, including local active labor market programs and unemployment insurance. Some countries have introduced labor taxes, but their design is often ad-hoc and could be having adverse effects (forcing migrants into the illegal circuit). 53 Alternatively, migration permits could be taxed to price them better to reflect these externalities. Being a fixed tax, as opposed to a tax on variable cost, this has the added advantage of being less distortionary.
More flexible schemes to manage migration flows and reduce irregular migration. From the host country perspective, an important research/policy question is how to improve the management of migration flows to encourage the migration of workers that help close skills gaps while minimizing irregular migration flows. Outside of regional labor markets like the EU, most countries have adopted "command-and-control" methods to regulate migration, setting explicit or implicit quotas for different categories of migrants. This presumes having enough information about the demand and supply of different types of skills and potential future dynamics; something that, in general, public administrations are unlikely to master. Options to consider include giving a more prominent role to the private sector; both employers and local intermediation agencies, that have incentives to understand the local and global pull of labor. The main problem with this approach would be the existence of negative social externalities that would not be taken into consideration by employers and migrants. But as discussed above, these social externalities could be addressed with appropriate tax policies.
Clearly, none of the policies and programs discussed here is easy to implement and we still have scant evidence about their potential impacts. But given the potentially large gains in terms of jobs outcomes and economic development, it seems imperative to continue to work on the design, implementation, and rigorous evaluations of comprehensive migration/jobs strategies along the lines described in this section. Not all evaluations will be able to rely on randomized control trials; some of the interventions are simply too complex or would take too long to deliver results and inform policy. But there are cheaper alternatives able to provide real-time feedback on the effectiveness of different policies/programs that are being actively discussed in the evaluation literature. 54 Effective migration policies will also require investments in data and diagnostics tools, a new research agenda for operational solutions, technical and financial assistance to countries, and continued global dialogue and coordination among the different actors. It is critical to expand and improve migration data 55 to monitor global trends in movements and remittances, to properly document the spatial gradients in livelihood prospects, to identify the main drivers and constraints affecting migration decisions (e.g. economic opportunities, access to amenities, information and aspirations), and to develop the analytical understanding of the effects of climate and environmental change on migration.
Globally, the need for coordination and knowledge exchange among the different national and international players is more needed than ever, with the Global Compact on Migration approved in 2018, an important new milestone.