Foreign Aid Can Dampen Migration If It Improves Public Services

Introduction

With the refugee crisis and the arrival of thousands of migrants on the southern European coasts, pressure has been mounting on the European Commission and the most affected EU member states to find ways to effectively manage (and stem) migration flows. Many see foreign aid as an essential part of a long-term strategy to address the root causes of migration through the creation of job opportunities, quality education, and better public services. Indeed, pledges to scale up aid to developing countries are now routinely accompanied by statements arguing that helping countries to develop gives their people an incentive to stay at home. In June 2015, for instance, the U.K. defence secretary declared that ‘‘Britain needs to spend more of its budget on helping stabilise countries so that it doesn’t have to ‘fish’ migrants out of the Mediterranean”.¹ Mason, R. “Fallon: UK aid budget should be used to discourage mass migration from Africa.” Guardian, 21 June 2015. www.theguardian.com/politics/2015/jun/21/michael-fallon-uk-aid-budget-discourage-mass-migration-africa But have donors followed up on their promises by actually increasing aid budgets? And, if so, can additional foreign aid help reduce migration flows?

Trends in aid allocation

At initial glance, the answer to the first question is a clear ‘yes’. As shown in figure 1, net official development assistance (ODA) by OECD donor countries has risen markedly in the years since 2012. Yet, the increasing aid volumes were mainly driven by humanitarian aid and in particular aid spent within donor countries, of which a large part is accounted for by in-donor refugee costs. The amount of long-term development aid that actually reaches recipient countries barely changed between 2010 and 2016, which appears to be at odds with the objective of tackling the root causes of emigration from developing countries.

While a general trend for long-term development aid is thus hardly discernible, donors have adjusted their aid allocations in a specific manner to rising refugee flows. Previous empirical research by Czaika and Mayer (2011) on the 1990s and early 2000s showed that the predominant donor response to refugee movements was to provide additional humanitarian assistance. More recently, however, donors have increasingly recognized the importance of long-term development aid that aims to reduce incentives for emigration. This is exemplified by the recent EU agreements with Lebanon, Jordan, and Turkey, the main countries of first asylum for Syrian refugees. In all three countries, long-term assistance—in particular, for social services including education—accounts for more than two-thirds of donor commitments (Lanati and Thiele 2018c). Education projects supported by the EU include, for instance, training for Syrian doctors and nurses in Turkey and the integration of Syrian refugee children in Jordanian public schools.

Our own empirical research (Lanati and Thiele 2018c) confirms for a broader sample of recipient countries that aid donors have not only changed their rhetoric, but also their behavior: since the early 2000s, higher numbers of internally displaced persons (IDPs) as well as refugees in countries of first asylum have been associated (on average) with higher allocations of long-term development aid (figure 2). The descriptive evidence shown in figure 2 is corroborated by a regression analysis that includes standard determinants of aid allocation such as bilateral trade links and recipient country governance along with additional control variables. Even so, the estimated donor response to refugee movements is very modest: a doubling in the number of hosted refugees and IDPs increases long-term development aid by about 2 percent for each category of displaced individuals.

Disentangling the aid channels: Higher incomes vs. better public services

Even if recent pledges to scale up development aid materialize, the effect on emigration to Europe will be much less clear-cut than some policymakers seem to think. Rather, whether aid increases or reduces migration depends on the channels through which aid mainly affects migration decisions. Broadly speaking, as long as foreign aid is not completely wasted, it may raise private incomes or improve non-monetary dimensions of well-being such as public services, or both.

If aid raises private incomes, the impact on migration is expected to exhibit an inverted U-shaped pattern (e.g. Hatton and Williamson 2002). The logic of this so-called migration hump is as follows: at low levels of development, additional income is likely to enable a larger share of the population in the countries of origin to finance migration costs, raising the number of people who leave. At higher development levels, such budgetary constraints that prevent people from emigrating become less binding, whereas rising household incomes provide an incentive to stay at home because the income gap relative to potential destination countries declines.

The threshold at which the income-migration relationship turns negative has been estimated at between US$8,000 and 10,000 (purchasing-power-parity adjusted; Clemens and Postel, forthcoming). Since income levels in most aid-receiving countries are far lower, Clemens and Postel argue that more foreign aid will almost inevitably lead to more people leaving developing countries, rather than discourage emigration.

Still, their reasoning is unduly focused on the private income dimension of development. After the international community agreed on the Millennium Development Goals, foreign aid has been reoriented toward non-monetary objectives, particularly in the area of basic health and primary education. At the same time, Dustmann and Okatenko (2014) have shown that the quality of local amenities such as health facilities, schools, and the quality of a country’s institutions is an even more important factor shaping migration decisions than household wealth. Foreign aid may therefore reduce emigration from poorer countries through improved public services.

The quantitative relevance of these two counteracting channels depends on whether foreign aid effectively raises household incomes or improves public services. While the empirical literature is inconclusive on whether there is a positive and significant aid-growth relationship (e.g. Qian 2015), even studies that find a positive impact on growth suggest that this effect is modest. For example, Clemens et al. (2012) estimate that raising growth by 1 percentage point per year in the average recipient country would require additional aid in the order of 10 percent of GDP, which is arguably far more than what donors would be willing to spend to reduce migration.

Regarding local amenities, several studies point to substantial positive effects of aid on social infrastructure. This is especially true in education, where development cooperation has been shown to contribute to rising primary school enrollment and lower drop-out rates (e.g. Birchler and Michaelowa 2016; Dreher et al. 2008). There is also limited evidence that foreign aid specifically targeted at governance issues—inter alia, support for government capacity to formulate and implement policies and activities to promote human rights and prevent conflict—may help improve local political institutions (Jones and Tarp 2016).

The evidence so far

The link between aid flows and migration has hardly been explored empirically. Until recently, there was a consensus among development researchers that foreign aid tends to encourage emigration from developing countries by enabling poor people to incur the cost of migrating. The most influential empirical study that supports this assertion is Berthélemy et al. (2009); they find a positive link between the aggregate aid received and migration for a large cross-section of developing countries.

Yet, their approach has methodological weaknesses. Most notably, due to lack of data, they use the stock of migrants born in country i and living in country j as the dependent variable in their regression analysis. The development aid that i receives from all donor countries j is the explanatory variable that is meant to capture the effect of aid. This approach may be misleading because migrants from country i may have lived in donor country j long before development aid was ever given. Hence, the stock variable cannot effectively capture the causal effect of foreign aid on migration decisions over time.

In Lanati and Thiele (2018a), we revisit the aid-migration link for a large sample of donor and recipient countries over the period 1995-2015.² Including a time dimension rather than performing a pure cross-section analysis as in Berthélemy et al. (2009) allows us to investigate how migration from country i to country j responds to changes in the overall foreign aid received by country i. Furthermore, it attenuates econometric problems concerning the identification of causal effects. Unlike Berthélemy et al. (2009), we employ annual emigration rates from recipient country i to donor country j as the dependent variable. After controlling for a number of factors that have previously been shown to affect migration decisions—including the size of migrant networks in destination countries as well as the cultural and geographical proximity between donor and recipient countries—we obtain robust evidence for a negative aid-migration relationship. This is in clear contrast to the previous consensus view.³ When we replace emigration rates by stocks of immigrants, the effect of aid on migration turns positive. Hence, we can interpret the discrepancies between our results and those obtained by Berthélemy et al. (2009) as driven mainly by the choice of different dependent variables.

Given that the predicted effects of foreign aid on migration that run through higher incomes and better public services point in different directions, it is important to disentangle their respective empirical relevance. In Lanati and Thiele (2018b) we do so by departing from the assumption that all aid has the same impact. Specifically, we use the distinction between early-impact aid and late-impact aid proposed by Clemens et al. (2012). Early-impact aid may generate income growth in the short to medium term, say after 2–3 years. It includes, for example, support for agricultural production such as improved seeds and fertilizer. Late-impact aid, in contrast, immediately affects non-monetary dimensions of well-being but may lead to higher incomes only in the very long run, say after 10 years. Assistance for social infrastructure (schools, clinics, etc.) is a typical example. We find that a rise in late-impact aid is associated with falling emigration rates, providing support for the relevance of the public services channel. In quantitative terms, the effect is not large but still non-negligible: on average, a 10 percent increase in late-impact aid lowers the emigration rate by 1–1.5 per cent. The negative association between late-impact aid and migration applies to a wide range of service improvements, from better schools to cleaner air and more reliable state institutions.

Perhaps more surprisingly, for early-impact aid the effect on migration is also significantly negative, if very small. This may be due to the fact that aid categories that are supposed to foster income growth often also improve amenities. Aid for infrastructure, for example, may have an income effect through lowering transport costs but at the same time establishes collective goods such as roads and railways.

Two studies by Gamso and Yuldashev (2018; forthcoming) similarly provide a differentiated account of the aid-migration relationship. Gamso and Yuldashev (2018) compare the effects of rural and urban development aid on international migration. They analyze aid from OECD donors to more than 100 low-and middle-income countries over 25 years (1985–2010) and find that countries that receive larger amounts of rural development aid have lower emigration rates. Results from survey data suggest that emigration from developing countries is reduced mainly through investments in capacity building in the agricultural sector. No significant link could be detected between aid to urban areas and migration.

Gamso and Yuldashev (forthcoming) also consider aid disaggregated by sector for the same panel of donor and recipient countries and investigate the relationships between emigration rates and inflows of three types of aid: aid targeting economic development, aid targeting governance and aid targeting social development. They find that emigration rates are lower when aid for governance is higher, whereas aid for economic or social development does not affect emigration rates. Again, there is no indication of any substantial income effects that might loosen the budget constraints of prospective migrants and enable them to leave.

Policy implications

There is now robust evidence that foreign aid with a significant public services component tends to reduce emigration. Policymakers in donor countries thus correctly perceive a scaling-up of development aid as a possible instrument to curb immigration. At the same time, our results also suggest that an unrealistically large increase in aid would be required to reduce emigration rates in a sizeable manner. If we take our point estimates at face value, a doubling of late-impact aid would be required to lower emigration rates by a still fairly modest 10–15 percent.

We may expect a somewhat stronger response if (contrary to recent trends) additional aid actually reaches recipient countries, rather than mainly financing refugee-related expenditures in donor countries. Such ‘in-donor’ aid does little to create local development opportunities and will therefore hardly affect migration decisions in countries of origin.

More generally, the primary objective of foreign aid should be to foster development and reduce poverty in recipient countries—regardless of any indirect impact on emigration. If, say, income-generating projects such as providing improved seeds to small-scale farmers were desirable from a development perspective, it would be problematic if such projects were not realized because a small fraction of beneficiaries might emigrate as a result. This is not to deny that in many cases there will be no trade-off between development needs and migration objectives in the first place. However, more effective instruments exist to manage migration that act directly on incentives to migrate (MEDAM 2018). Hence, we caution against misusing foreign aid in a (most likely futile) attempt to manage migration.

¹ Mason, R. “Fallon: UK aid budget should be used to discourage mass migration from Africa.” Guardian, 21 June 2015. www.theguardian.com/politics/2015/jun/21/michael-fallon-uk-aid-budget-discourage-mass-migration-africa

² Including a time dimension rather than performing a pure cross-section analysis as in Berthélemy et al. (2009) allows us to investigate how migration from country i to country j responds to changes in the overall foreign aid received by country i. Furthermore, it attenuates econometric problems concerning the identification of causal effects.

³ When we replace emigration rates by stocks of immigrants, the effect of aid on migration turns positive. Hence, we can interpret the discrepancies between our results and those obtained by Berthélemy et al. (2009) as driven mainly by the choice of different dependent variables.

References

Berthélemy, J.C., M. Beuran, and M. Maurel. 2009. “Aid and migration: Substitutes or complements?” World Development 37 (10): 1589–1599.

Birchler, K. and K. Michaelowa. 2016. “Making aid work for education in developing countries: An analysis of aid effectiveness for primary education coverage and quality.” International Journal of Educational Development 48: 37–52.

Czaika, M. and A. Mayer. 2011. “Refugee Movements and Aid Responsiveness of Bilateral Donors.” Journal of Development Studies 47 (3): 455–74.

Clemens, M.A., S. Radelet, R. Bhavnani, and S. Bazzi. 2012. “Counting chickens when they hatch: Timing and the effects of aid on growth.” Economic Journal 122 (561): 590–617.

Clemens, M.A. and H.M. Postel. Forthcoming. “Deterring Emigration with Foreign Aid: An Overview of Evidence from Low-Income Countries.” Population and Development Review.

Dreher, A., P. Nunnenkamp, and R. Thiele. 2008. “Does Aid for Education Educate Children? Evidence from Panel Data.” World Bank Economic Review 22 (2): 291–314.

Dustmann, C. and A. Okatenko. 2014. “Out-migration, wealth constraints, and the quality of local amenities.” Journal of Development Economics 110: 52–63.

Gamso, J. and F. Yuldashev. 2018. “Does rural development aid reduce international migration?” World Development 110: 268–282.

———. Forthcoming. “Targeted foreign aid and international migration: Is development-promotion an effective immigration policy?” International Studies Quarterly.

Hatton, T.J. and J.G. Williamson. 2002. “What fundamentals drive world migration?” NBER Working Paper 9159, , National Bureau of Economic Research, Cambridge, MA.

Jones, S. and F. Tarp. 2016. “Does Foreign Aid Harm Political Institutions?” Journal of Development Economics 118: 266–281.

Lanati, M. and R. Thiele. 2018a. “The impact of foreign aid on migration revisited.” World Development 111: 59–74.

———. 2018b. “Foreign assistance and migration choices: Disentangling the channels.” Economics Letters 172: 148–151.

———. 2018c. “How Donors Respond to Refugee Movements.” In 2018 MEDAM Assessment Report on Asylum and Migration Policies in Europe, MEDAM (Mercator Dialogue on Asylum and Migration), 109–115. Kiel: Kiel Institute for the World Economy.

MEDAM (Mercator Dialogue on Asylum and Migration). 2018. 2018 MEDAM Assessment Report on Asylum and Migration Policies in Europe, Kiel: Kiel Institute for the World Economy.

Qian, N. 2015. “Making Progress on Foreign Aid.” Annual Review of Economics 7 (1): 277–308.