Comforting immigration critics? Public opinion toward development aid as a tool to reduce refugee inflows to Germany

Development aid as a response to the European ‘refugee crisis’

With the arrival of more than 3 million people from the Middle East and Africa since 2015 (UNHCR, 2017), public concerns about immigration have risen. As a response, policy makers in Europe have almost univocally supported fighting the root causes of migration in refugees’ countries of origin. They have promoted cooperation with and financial support of developing or conflict-ridden states to curb migration flows to Europe by fostering socio-economic development in the Middle East and Africa (see, e.g., Dreher et al., 2018).

Development aid–or more precisely official development assistance (ODA)–¹ In this policy brief, we use both terms interchangeably.has also been vigorously debated in Germany, where the public has become increasingly polarized about immigration (see Ademmer and Stöhr, 2018). Germany accommodated the largest share of asylum seekers in Europe due to the German government’s decision to bypass the so-called ‘Dublin agreement’ in September 2015. Opposing the welcoming policy toward asylum seekers, the right-wing populist party Alternative for Germany (AfD) campaigned on anti-immigration issues and rose to electoral success (see, e.g., Mushaben 2017; Wiesendahl 2016). Against the backdrop of increasing societal polarization about migration, policy makers have publicly debated potential response strategies, among which development aid has figured prominently (see BMZ 2017).

The expansion of development aid may indeed have the potential to ease the increasing tension between immigration proponents and opponents: immigration supporters may dislike the fact that immigration remains restrictive, but are likely to support a policy that is designed to help potential migrants. At the same time, immigration opponents may dislike the fact that resources are spent on potential migrants in their countries of origin, but they are also likely to support a policy that helps to reduce immigration. The expansion of development aid can thus represent a policy compromise that respects principles that both sides hold dear, especially against the background of the usually rather high public support in Germany for development aid (Schneider and Gleser 2018).

This compromise potential of development aid hinges on a number of conditions, however: First, development aid indeed needs to improve development and reduce migration. Otherwise immigration supporters and critics are unlikely to support this policy in the longer term. Second, people need to trust in development aid to have this effect. Third, they need to be willing to spend money on development aid for this purpose. And fourth, the aforementioned trust and willingness to spend money should be shared by supporters of different political parties.

In this policy brief,² The responsibility for the contents of this publication rests with the authors, not the Institute or the Mercator Dialogue on Asylum and Migration (MEDAM). Any comments should be sent directly to the authors. we analyze these conditions to answer the question of whether the expansion of development aid qualifies as a policy response that can reduce polarization and create consensus among citizens who hold diverse views about immigration. The empirical analysis is based on survey data from the Aid Attitudes Tracker (AAT) project wave 9, which was collected in November and December 2017 (see Clarke et al. 2018).

Development aid does not automatically reduce migration

ODA has been discussed as an important tool to fight the root causes of migration. Poverty, famine, war, persecution, and crime, are seen as the root causes of flight. In response, political decision makers consider and communicate development aid as a remote tool to increase all dimensions of a decent life and lift people out of poverty (for an example, see BMZ 2017). This is assumed to increase incentives for potential migrants to stay in their home countries (see, e.g., Dreher et al. 2018).

In the academic literature, in contrast, empirical evidence on the general effectiveness of development aid is mixed (e.g., Doucouliagos and Paldam 2009; Tarp and Hansen 2003). The findings on the effectiveness of aid in reducing migration (Berthélemy et al. 2009; Lanati and Thiele 2018a, 2018b) and the root causes of refugee flows are similarly mixed (Dreher et al. 2018). Empirical evidence shows that developing countries with higher income have on average higher emigration. The relationship reverses at a middle-income threshold. This ‘migration-hump’ has led some to suggest that development aid too poor countries might increase migration (Clemens 2014). Recent studies suggest that the sectoral allocation of aid, rather than the absolute amount, might be decisive (Gamso and Yuldashev 2018; Lanati and Thiele 2018b). They differentiate between aid that is designed to raise the income of people in developing countries in the short run and aid that is designed to improve public services, such as health systems or other forms of public infrastructure, which leads to higher incomes only in the long run (Lanati and Thiele 2018c). This latter form of aid has been shown to help reduce migration through development, but the amount of aid that would have to be spent to achieve this effect would exceed current levels by far. To increase the effectiveness of aid in addressing migration flows, politicians tie aid to recipient countries’ cooperation in reducing refugee flows, which, however, raises ethical concerns.³ See for instance Zeit Online, “Gabriel droht mit Stopp der Entwicklungshilfe” (18 January 2016). www.zeit.de/politik/deutschland/2016-01/sigmar-gabriel-entwicklungshilfe-fluechtlingskrise-nordafrika Apart from that, development aid can also be used to cover the costs of coping with the refugee crisis within destination countries—indeed, many European countries have declared these costs as ODA to the OECD Development Assistance Committee (DAC, see e.g., Devex 2017; Obrovsky 2015). In these cases, a formal increase in ODA does not contribute to the socioeconomic development of countries of origin, because it does not reach them in the first place (see Lanati and Thiele 2019). While development aid is thus needed for a variety of reasons, the straightforward impact on reducing migration that is often suggested by policy makers is at odds with scientific evidence. The condition that an increase in development aid reduces migration is, hence, not fulfilled, raising doubts about whether an increase of development aid can function as a policy compromise between migration critics and supporters.

People are ambiguous about whether development aid reduces migration

Next, we argued that for development aid to function as a policy compromise, people would also need to believe that development aid reduces migration. Yet, our empirical results based on AAT survey data, reflect the ambiguity that can be found in the academic literature. Figure 1 shows that about 45 percent of survey respondents either agree or agree strongly with the statement that “development aid for humanitarian emergencies in the Middle East and Africa helps to reduce immigration to Germany.” Roughly 20 percent disagree or strongly disagree and approximately 25 percent choose the indifferent scale midpoint neither agree nor disagree. In combination with about 10 percent who don’t know, ae third of the German population is undecided about whether development aid is a proper tool to reduce immigration.

People are split on whether to increase spending to fight the root causes of migration

Moreover, we argued that for development aid to function as a policy compromise, people would also need to be willing to spend more on development aid. Still, survey participants were similarly divided about whether the German government should increase its disbursements for development aid. Terrorism and insecurity are primary causes of flight from major countries of origin such as Afghanistan. By reducing grievances, development aid could help stabilize such countries. We therefore investigated the attitude toward increasing aid to reduce terrorism. As shown in figure 2, there is a relatively even split between supporters, opponents, and those who don`t know. Only 7 percent strongly agree and including those who agree, 30 percent state that Germany should spend more on development aid to this end, while about 35 percent disagree or strongly disagree. The largest group is constituted by those who neither agree nor disagree or say they don’t know. A majority of the population is thus hesitant to support development aid as a means to fight terrorism in these regions.

As a preliminary result, the first three of our conditions to make an increase in development aid a policy compromise between immigration supporters and immigration skeptics do not hold. The next section, in which we turn to the final condition, further shows why this is the case.

Partisan and attitudinal frictions over the effectiveness and increase of ODA remain

Immigration critics and supporters hold very different views about whether development aid helps reduce immigration and on whether they would be willing to spend more on development aid. This divisiveness is more driven by attitudes toward immigration than by partisan preferences. As figure 3 shows, individuals with skeptical attitudes toward immigration are much more likely to be skeptical of development aid, too. More specifically, the more skeptical the attitude toward refugees is the lower the probability of agreeing that ODA helps to reduce refugee inflows and the lower the probability of supporting increasing ODA to tackle terrorism. This effect is hardly noticeable with regard to partisan preferences. True, the more a respondent moves to the right pole of the political ideology scale, the lower the probability of agreeing with either policy statement, as figure 3, panels c and d reveals. Nonetheless, the effect of ideology is substantially smaller than the previously mentioned effect of attitudes toward refugees. This pattern also shows in figure 4, in which voters for parties that are usually considered more right-leaning (CDU/CSU, FDP, and AfD) do not follow a clear pattern in their attitudes toward development aid, unlike parties traditionally classified as left-leaning (SPD, Die Linke, Bündnis 90/Die Grünen).⁴ AfD refers to Alternative for Germany, CDU refers to the Christian Democratic Union, CSU to the Christian Social Union, FDP to the Free Democratic Party, and SPD to the Social Democratic Party. Die Linke is the Left Party, and Bündnis 90/Die Grünen are the Greens. Figure 4, panel a shows that agreement is highest among supporters of the Greens, which is then, however, followed by those of the conservatives (CDU/CSU), and only then by the SPD and the Left Party (Die Linke). For all of these, percentages above 50 indicate that an absolute majority agrees with the statement that ODA reduces refugee inflows. By contrast, a considerably lower share of voters for the Liberal Party (FDP) and the right-wing populist party AfD agrees. Finally, fewer than one out of three non-voters agrees that ODA reduces refugee inflows.

Moving to the support for increasing ODA disbursements to combat terrorism, figure 4, panel b shows that while the overall level of support is lower, the aforementioned gap between the political left and right is more pronounced than in panel a. Among voters for the Greens and left-wing parties (SPD and the Left Party) support is around 40 to 45 percent. Amid CDU/CSU, FDP, and AfD voters—all belonging to the right-wing party bloc—support is considerably lower. Among the latter only slightly more than 10 percent are in favor of increasing ODA for this purpose. Finally, 20 percent of non-voters support increasing ODA to combat terrorism.⁵ Additionally, part of the sociodemographic background explains attitudes to some extent (see table 1 in the appendix). We find evidence of an education effect for both indicators that goes beyond the positive correlation between education and attitudes toward refugees. People with a high level of formal education have a higher probability of supporting the indicators compared with people who have no or only a low level of education and qualifications. Furthermore, the support for indicators grows convexly with age (i.e., an extra year of age has a higher positive effect among the old compared with the young). And while women are more critical of increasing ODA to combat terrorism, people belonging to the highest household income category support it more often.

In sum, the refugee issue seems to close the gap between the left and right wing at least partially with voters for the conservative CDU/CSU being close to those in the left-wing party bloc (Green Party, Left Party, and SPD). At the same time, our data suggest that the dividing line with regard to the perception of development aid as a tool to curb refugee inflows runs between people who are critical of refugees and those who are more welcoming. Promoting and increasing ODA in order to fight the root causes of refugee inflows is supported less by immigration critics.⁶ The dependent variable for the binary logistic regression models is a dummy contrasting ‘agree’ and ‘strongly agree’ with all other available answer categories, including ‘don’t know.’ We analyzed the effect of an individual’s attitude toward refugees (additive index summing up 6 items and ranging from 1 to 5; Cronbach’s Alpha 0.83; 1 = positive attitude, 5 = negative attitude) controlling for political ideology (as a substitute for party preference; measured on the classical 10-point left-right scale; 0 = left, 10 = right) and basic sociodemographic characteristics (see table 1 in the appendix). Furthermore, it must be noted that the left-right scale is positively correlated to the attitudes toward refugees (Pearson’s R = 0.34). The more to the right a respondent is located, the more critical the stance toward refugees. Running the model and including either only political ideology or attitude toward refugees shows that this correlation poses no severe problem. The substantial interpretations remain unaltered. Development aid is thus unlikely to be a bridge-builder in the polarized debate about immigration.

Development aid is not an effective tool to comfort immigration critics

This policy brief has shown that there is no consensus across partisan groups in Germany when it comes to ODA as a tool to tackle refugee inflows. Rather, attitudes toward refugees predict whether respondents consider development aid an effective tool for addressing migration issues or not. We concluded that relying on development aid per se as a tool to address immigration issues is likely to prove ineffective in bridging in the divides in the German population that have opened during the European ‘refugee crisis.’

Based on these results, we recommend that policy makers who want to fight the root causes of migration with development aid target this aid based on evidence on whether and what kind of ODA is effective in reducing migratory pressures in countries of origin (see Lanati and Thiele 2018c). Although almost all relevant political parties support fighting the root causes of refugee flows, what really matters is the eventual effectiveness of this policy. Development aid should be adjusted accordingly, but there are strong arguments in favor of not just focusing on those countries that are currently the most important origin countries of refugees and irregular migrants. Often, the reasons for upping sticks and leaving one’s home country will not easily be addressed by development aid (see Schneiderheinze and Stöhr 2017). In addition, poor institutions and insecurity might render aid ineffective (see, e.g., Zürcher 2017). Thus, regional and sectoral foci should be set with care and trade-offs between (short-term) humanitarian aid to support host countries of refugees in the region and ODA, which more likely takes effect in the medium or long run, must be considered. Inevitably, policy makers should also communicate these aspects transparently and honestly to the electorate.

¹ In this policy brief, we use both terms interchangeably.

² The responsibility for the contents of this publication rests with the authors, not the Institute or the Mercator Dialogue on Asylum and Migration (MEDAM). Any comments should be sent directly to the authors

³ See for instance Zeit Online, “Gabriel droht mit Stopp der Entwicklungshilfe” (18 January 2016). www.zeit.de/politik/deutschland/2016-01/sigmar-gabriel-entwicklungshilfe-fluechtlingskrise-nordafrika

⁴ AfD refers to Alternative for Germany, CDU refers to the Christian Democratic Union, CSU to the Christian Social Union, FDP to the Free Democratic Party, and SPD to the Social Democratic Party. Die Linke is the Left Party, and Bündnis 90/Die Grünen are the Greens.

⁵ Additionally, part of the sociodemographic background explains attitudes to some extent (see table 1 in the appendix). We find evidence of an education effect for both indicators that goes beyond the positive correlation between education and attitudes toward refugees. People with a high level of formal education have a higher probability of supporting the indicators compared with people who have no or only a low level of education and qualifications. Furthermore, the support for indicators grows convexly with age (i.e., an extra year of age has a higher positive effect among the old compared with the young). And while women are more critical of increasing ODA to combat terrorism, people belonging to the highest household income category support it more often.

⁶ The dependent variable for the binary logistic regression models is a dummy contrasting ‘agree’ and ‘strongly agree’ with all other available answer categories, including ‘don’t know.’ We analyzed the effect of an individual’s attitude toward refugees (additive index summing up 6 items and ranging from 1 to 5; Cronbach’s Alpha 0.83; 1 = positive attitude, 5 = negative attitude) controlling for political ideology (as a substitute for party preference; measured on the classical 10-point left-right scale; 0 = left, 10 = right) and basic sociodemographic characteristics (see table 1 in the appendix). Furthermore, it must be noted that the left-right scale is positively correlated to the attitudes toward refugees (Pearson’s R = 0.34). The more to the right a respondent is located, the more critical the stance toward refugees. Running the model and including either only political ideology or attitude toward refugees shows that this correlation poses no severe problem. The substantial interpretations remain unaltered.

Acknowledgments

We thank Martin Bruder, Matthias Lücke, and Rainer Thiele for very helpful comments and suggestions. The usual disclaimer applies.

References

Ademmer, E. and T. Stöhr. 2018. “Europeans are more accepting of immigrants today than 15 years ago: Evidence from eight waves of the European Social Survey” MEDAM Policy Brief 2018/1.

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

BMZ. 2017.“Helping refugees build a future. Tackling the root causes of displacement, stabilizing host regions, supporting refugees.” Bonn/Berlin: Federal Ministry for Economic Cooperation and Development.

Clemens, M. A. 2014 “Does development reduce migration?”, Chapter 6 in International Handbook on Migration and Economic Development, edited by R. E. B. Lucas, Edward Elgar.

Devex. 2017. “Debating the rules: What in-house refugee costs count as aid?”, Devex blog, 5 July 2017.

Doucouliagos, H., and M. Paldam. 2009. The aid effectiveness literature: The sad results of 40 years of research. Journal of Economic Surveys 23(3): 433–461.

Dreher, A., A. Fuchs, and S. Langlotz. 2018. “The Effects of Foreign Aid on Refugee Flows.” European Economic Review 112: 127-147.

Gamso, J., and F. Yuldashev. 2018. “Targeted Foreign Aid and International Migration: Is Development-Promotion an Effective Immigration Policy?” International Studies Quarterly, online first.

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

Lanati, M., and R. Thiele. 2018b. “Foreign assistance and migration choices: Disentangling the channels”, Economics Letters 172: 148–151

Lanati, M., and R. Thiele. 2018c. “Foreign aid can dampen migration if it improves public services” MEDAM Policy Brief 2018/2.

Lanati, M., and R. Thiele. 2019. “Foreign assistance and emigration : accounting for the role of non-transferred aid.”  EUI Working Paper.

Mushaben, J. M. 2017. “Wir schaffen das! Angela Merkel and the European Refugee Crisis.” German Politics 26(4): 516–533.

Obrovsky, M. 2015. Ausgaben für Flüchtlingshilfe—Aufputz der ODA-Statistik oder reale Erhöhung der EZA? Policy Note 17/2015. Österreichische Forschungsstiftung für Internationale Entwicklung.

UNHCR. 2017. Global Report 2017. Geneva: United Nations High Commissioner for the Refugees (UNHCR).

Schneider, S. H., and S. H. Gleser. 2018. Meinungsmonitor Entwicklungspolitik 2018: Einstellungen zu Entwicklungszusammenarbeit und nachhaltiger Entwicklung. Deutsches Evaluierungsinstitut der Entwicklungszusammenarbeit (DEval), Bonn.

Schneiderheinze, S., and T. Stöhr. 2017. “4.1. Determinants of migration: Who leaves, who arrives, and what drives their destination choices?” In 2017 MEDAM Assessment Report on Asylum and Migration Policies in Europe. Sharing Responsibility for Refugees and Expanding Legal Immigration. Kiel: Kiel Institute for the World Economy (IfW).

Tarp, F., and H. Hansen. 2003. “Aid effectiveness disputed.” In Foreign Aid and Development, edited by F. Tarp. London: Routledge.

Wiesendahl, E. 2016. “Der Kulturkonflikt um die Flüchtlingskrise und die politischen Folgen.” Zeitschrift für Staats- und Europawissenschaften 14(1): 53–79.

Zürcher, C. 2017. “What do we (not) know about development aid and violence? A systematic review”, World Development 98: 506–522.