Managing Refugee Protection Crises: Policy Lessons from Economics and Political Science

We review and interpret research on the economic and political effects of receiving asylum seekers and refugees in developed countries, with a particular focus on the 2015 European refugee protection crisis and its aftermath. In the first part of the paper, we examine the consequences of receiving asylum seekers and refugees and identify two main findings. First, the reception of refugees is unlikely to generate large direct economic effects. Both labor market and fiscal consequences for host countries are likely to be relatively modest. Second, however, the broader political processes accompanying the reception and integration of refugees may give rise to indirect yet larger economic effects. Specifically, a growing body of work suggests that the arrival of asylum seekers and refugees can fuel the rise of anti-immigrant populist parties, which may lead to the adoption of economically and politically isolationist policies. Yet, these political effects are not inevitable and occur only under certain conditions. In the second part of the paper, we discuss the conditions under which these effects are less likely to occur. We argue that refugees’ effective integration along relevant linguistic, economic, and legal dimensions, an allocation of asylum seekers that is perceived as ‘fair’ by the host society, and meaningful contact between locals and newly arrived refugees have the potential to mitigate the political and indirect economic risks.


Introduction
In 2015, the number of individuals seeking asylum in Europe increased rapidly.For many governments, the situation posed a major challenge, because they quickly had to find ways to host asylum seekers and process their asylum claims.The "European Refugee Protection Crisis" of 2015 also pushed many countries to introduce more restrictive asylum and new integration policies. 1By and large, these decisions were made on an ad-hoc basis, with limited foresight on their likely impacts and a small evidence base.
While 2015 was certainly exceptional in comparison to previous years, relatively sudden increases in asylum applications and refugee reception are quite common, also in Europe.As shown in Figure 1, Europe experienced the arrival of comparatively large numbers of asylum seekers also during the 1990s.Earlier refugee crises not covered in this figure include, among others, the millions of displaced Europeans in the aftermath of World War II.
Given that demand-side drivers, most notably wars and violent conflict (see Hatton 2004), account for much of the variation in asylum application numbers, it is likely that the number of people seeking asylum will remain outside host countries' complete control and keep fluctuating.Accordingly, while governments' reception and integration challenges might have been exceptionally large in the aftermath of the 2015 refugee protection crisis, similar situations are likely to arise also in the future.Thus, it is a good time to take stock on lessons learned from the past crisis and prepare for future challenges.
1 Initially, some policies made seeking asylum easier.Most notably, for a short period, Germany stopped sending Syrian refugees back to the country of first entry and started invoking the sovereignty clause of the Dublin regulation, i.e. to evaluate Syrian asylum seekers' claims in Germany.Subsequent changes in asylum policies had been almost exclusively restrictive and aimed to reduce the number of (staying) asylum seekers -including Germany retracting from the invocation of the sovereignty clause and Sweden no longer offering permanent residency permits upon arrival.Most countries have also introduced new integration policies focusing on asylum seekers with a high probability of staying or already accepted refugees with the main goal to facilitate their labor market integration.This mainly included language trainings and integration courses.For example, Germany started obliging accepted refugees and asylum seekers with a high probability of staying to participate in integration classes.Sweden has increased the availability of vocational introduction jobs and work experience placements for asylum seekers and expanded the range of services in household work that allow for tax deductions (see Swedish Government 2015).This paper aims to help both researchers and policy makers in this task.We present a selective review and an interpretation of the research examining the economic and political effects of receiving refugees in Europe and North America.We focus predominantly on published studies of the political and economic consequences of the arrival of forced migrants (i.e., asylum seekers, refugees, people with subsidiary protection)2 in Europe using credible research designs to identify causal effects.However, two notes are in order.First, where necessary, we also include studies looking at the effects of immigration more broadly, working papers, and expand our scope to other countries (e.g., the U.S.).Second, given that this is a rapidly developing literature, we have no claim to completeness.
We start with a short review of research on labor market and fiscal impacts of refugee immigration.Based on this literature, two main findings emerge.First, the average impact of refugees on native wages or employment is likely to be small.Second, while natives may be affected through public finances, the fiscal effects of receiving refugees are likely to remain quite limited even under pessimistic assumptions.We substantiate this argument by discussing the sources and the extent of uncertainties embedded in any estimate of the long-term fiscal impacts.
We then proceed with a review of research on the political consequences of asylum immigration.Our main argument is that receiving refugees may have more important economic effects through the broader political process.That is, increasing arrivals of asylum seekers often fuel anti-immigrant attitudes and the rise of authoritarian, populist, anti-immigrant parties.Often, these parties not only promote anti-immigrant policies, but tend to support isolationist policies more broadly (e.g., withdrawal from trade agreements or the European Union).In addition, in response to the increasing popularity of such parties, also mainstream parties tend to shift their policy positions towards isolationism.While it is hard to quantify how these political changes will affect policy -or what the economic effects of these policy changes are -we argue that this channel is likely to pose a larger risk for economic effects than any conceivable labor market or fiscal effects of receiving refugees.
The risk of rising right-wing populism leaves non-isolationist policy makers with the dilemma of how to respond.In order to inform policy decisions, we present a review of the relevant economics and political science research on the effects of asylum and integration policy responses.We argue that effective reception and integration policies both promote refugees' integration outcomes and decrease the potential of an isolationist backlash.
We first survey the literature on voter preferences about refugee policy.This research suggests that across European countries, the majority of residents have a skeptical view of current asylum policies and prefer to curb future refugee arrivals.However, they also tend to support the acceptance of refugees deserving of asylum (according to the Refugee Convention) and are willing to accept more asylum seekers as long as the allocation is 'fair' and proportional to the country's capacity.
We next discuss lessons from the existing literature on how to organize the reception of asylum seekers in a way that minimizes the potential for anti-immigrant backlash.Our suggestions include providing timely information to locals; facilitating repeated and meaningful contact between locals and newly arrived refugees; ensuring that the allocation of asylum seekers within countries is perceived as 'fair'; and taking into account that rural and more conservative constituencies tend to exhibit stronger backlash against refugees.
Finally, we review impact evaluations of (broadly defined) integration policies.This body of research suggests four lessons: interventions improving the match quality between active labor market policy measures and the individual characteristics of each refugee have large effects; long waiting periods, such as those arising from lengthy asylum process, reduce later employment prospects; temporary employment bans for newly arrived asylum-seekers, such as those currently enacted by most European countries, have a detrimental impact on the long-term economic integration of refugees; and proficiency in local language is strongly associated with labor market success and there are good reasons to think that at least part of this association reflects a causal effect of language skills on labor market success.
In sum, while aggregate labor market consequences appear small, both potential fiscal and political effects of a sudden increase in the number of asylum seekers are not inevitable.Our overview of the insights accumulated from earlier experiences suggest that a swift integration of refugees, an allocation of asylum seekers perceived to be fair, and policies that encourage meaningful interaction between refugees and host communities have the potential to mitigate the fiscal and political risks associated with large-scale refugee arrival -in addition to their direct effects.

Labor market and fiscal effects
This section discusses two channels through which the reception of refugees could have direct effects on the host country's economy.First, refugees could compete with natives in the labor market and thus drive down wages or employment.Second, refugees pay taxes, receive transfers and use public services, and may thus create fiscal effects.We present a short review on the literature examining these effects and conclude that both effects are likely to be relatively small.

A. Impact on native wages and employment
A simple economics textbook model provides a useful starting point for thinking about the labor market consequences of immigration (see, e.g., Borjas 2015 for discussion).In these models, immigration affects native wages only if the skillmix of immigrants differs from that of natives.If the distribution of skills is identical among immigrants and natives, the arrival of immigrants should increase the size of the economy, but have no impact on long-run native wages or employment.Sudden and large immigration does affect the capital-labor ratio and temporarily lowers wages, but over time capital adjusts and wages revert back to their pre-immigration level.However, if the skill-mix of immigrants and natives differs, some natives will win, and others will lose.More precisely, those who have complementary skills in comparison to immigrants will become more productive and can thus demand higher wages. 3Symmetrically, natives who are substitutes to immigrants in terms of their skills will become less productive.Thus, the labor market effects crucially depend on the extent to which immigrants and natives are substitutes vs. complements in the labor market.
A large empirical literature has examined the impact of immigration on natives' wages and employment by comparing natives working in labor markets differentially exposed to immigration.The key challenge of this research is that immigrants are not randomly allocated into labor markets. 4Thus natives working in labor markets with many immigrants are likely to differ from natives working in labor market with few immigrants also in ways that have nothing to do with immigration.
This identification challenge has led part of the literature to focus on quasi-experimental research designs where some labor markets experience the arrival of immigrants for reasons that are plausibly exogenous to the (unobserved) characteristics of these labor markets.Given that a large number of high-quality reviews on this topic already exists, we do not attempt to provide another one here, but rather refer the reader to Borjas (1999), Hanson (2009), Blau and Mackie (2016) and Dustmann, Schönberg and Stuhler (2016).
Our reading of this literature is that the impact of typical refugee flows on native wages and employment is likely to be modest.This conclusion is partly driven by the relatively small estimates for labor market effects even in situations where the number of immigrants is large (see the reviews cited above for details).On the other hand, in most cases of refugee inflows that create a substantial political reaction, the number of refugees tends to be relatively small compared to the size of the labor market. 5Finally, refugees typically struggle to find employment in the host country's labor market, which further reduces the competitive pressure they exert on natives.Thus, it seems unlikely that labor market effects would be the primary channel for refugees to affect natives' economic well-being.

B. Impact on public finances
The most important direct economic effect of receiving refugees is likely to occur through public finances.Refugees tend to have lower employment rates than other immigrant groups or natives throughout Europe (Brell, Dustmann, andPreston 2020, Fasani, Frattini, andMinale 2021).As a consequence, they tend to receive substantially more social transfers and to pay less taxes than natives or other immigrants.These differences give rise to dramatic differences in some forms of transfers.Ruist (2015) documents the situation in Sweden in 2017, where refugees accounted for 5.1% of the population in 2007 (the largest refugee population share in Europe) and 55% of social assistance spending.On the other hand, public spending on refugees' pensions, health and education was much lower than that of natives.As a consequence, refugees accounted for 5.6% of total public spending, i.e., quite close to their population share.Due to their low employment rate, however, refugees paid less taxes than they received benefits and used public services and thus created a net fiscal cost.In total, one percent of Swedish GDP was redistributed to the accumulated refugee population (including refugees' family members).For comparison, Sweden's foreign aid budget is roughly of the same size (OECD 2013).
While such cross-sectional observations clearly contain information, it is unclear how well they capture the long-term fiscal impacts of immigration.The reason is that the net cost or surplus that an individual creates for the public sector varies dramatically over her lifecycle.Everyone is a net burden during childhood and adolescence.Furthermore, most people create large net costs to the public sector during their last years of life.In order to truly capture the fiscal impacts of immigration, these dynamics would have to be taken into account.That is, the appropriate measure for fiscal impacts is the discounted sum of all future taxes, transfers and costs due to public service consumption (Lee and Miller 2000, Storesletten 2000, Ruist 2020).
The challenge in incorporating lifecycle dynamics into the estimates of the fiscal impact of immigration is that much of the costs and benefits will take place in the future.Thus, researchers have to make strong assumptions about the future labor market performance of immigrants, the cost of providing public services, the structure of the tax and benefit system, overall economic growth and so forth.Clearly, our ability to forecast these factors for the next decades is very limited.Accordingly, all estimates of the long-term fiscal effects of immigration are best understood as scenario exercises.
Figure 2 illustrates this issue using data from Finland.Each point in the figure refers to the net present value of the fiscal impact of an additional immigrant arriving in Finland as a function of age at arrival (x-axis) and future labor market performance (marker style). 6The top series correspond to a scenario where the immigrant immediately starts to follow the average profile of natives.That is, in this scenario, a person arriving at age 30 is assigned the average taxes, transfers and cost of public services of current 30-year-old natives.In the next year, she is assigned the averages of current 31-yearold natives and so forth.Furthermore, her offspring is also assigned the age profiles corresponding to current natives.Fertility and mortality are assumed to follow the age profiles observed in current data and pensions are estimated within the model using the earnings profiles of current natives.The future costs and taxes are then discounted to net present value using 3% discount rate and assuming 1% annual earnings growth.
The top scenario highlights the importance of age at arrival in the (unrealistic) situation where immigrants enter the labor markets immediately at the level of natives.Those arriving after age 50, and those arriving before age six constitute a net burden to the taxpayer.Newborn natives, i.e., those "arriving" at age zero, have a negative net present value in these scenarios.(This observation alone highlights the fact that these scenarios do not aim to provide credible predictions.)In contrast, immigrants arriving as young adults spend a long period working and paying taxes, but the expenses due to their education and health care during childhood is paid somewhere else.Thus, they make a large positive contribution to public finances if they integrate immediately into the labor markets.
Thus far, however, immigrants -and refugees in particular -have experienced difficulties in finding stable employment in Finland (Sarvimäki 2011(Sarvimäki , 2017)).The bottom scenarios illustrate this fact.Immigrants and their offspring are now assigned the observed age profile of immigrants (including non-refugees) in 1995-2012, who arrived in Finland during the 1990s.The net present value on public finances is now negative regardless of the age at arrival and varies between €110,00 and €150,000.
The difference between the top and bottom scenarios highlights the importance of labor market integration in generating the fiscal impact of immigration.Among those arriving to Finland between the ages of 20 and 40, the difference between the two scenarios is almost €300,000 in net present value.One way to interpret these results is that if one could design integration training that would help moving immigrants from the bottom to the top scenario, such program would be cost-efficient even with a cost of €300,000 per participant.In comparison, the average investment in training for immigrants who participate in integration programs in Finland has been around €15,000 (Sarvimäki and Hämäläinen 2016).Of course, designing integration programs that actually move immigrants from the bottom to the top scenarios may not be feasible.Nevertheless, a comparison of these alternative scenarios illustrates that effective integration programs can have substantial fiscal impacts.Sarvimäki et al. (2014).See Section 3.2 for discussion.
The third set of scenarios, presented in the middle of Figure 2, correspond to an integration profile, where immigrants follow the profiles of the 1990s immigrants, but their children follow the profiles of current Finnish natives.The difference between these scenarios and bottom scenarios -roughly €75,000 for immigrants arriving in their mid-20s -illustrates the importance of the integration of the children of immigrants to the host country's labor markets.
We stress that none of the scenarios presented in Figure 2 should be interpreted as "price-tags" of immigration.Clearly, Finland's tax and benefit systems as well as the way public services are provided will change in the future.Other assumptions embedded in these scenarios, such as a steady 1% annual earnings growth, are also unlikely to be accurate predictions of the future.However, scenarios such as those presented in Figure 2 can help to clarify which factors matter the most and thus assist policy makers in focusing on these factors.In particular, Figure 2 shows that increasing employment through efficient integration policies has potential to create large benefits purely from a public finances viewpoint.

Political effects
We now turn to the literature on refugee arrivals' political effects in the receiving countries.We first present a relatively detailed review of the recent literature on the impact on votes and attitudes.At the end of the section, we also briefly discuss the existing evidence on the tendency of populist parties to support other isolationist policies and their influence on the policy stances of mainstream parties.

A. First-order effects: attitudes and votes
There are two main theories as to how individuals respond to the presence of a 'foreign' group of people.One the one hand, the 'contact hypothesis theory' (Allport 1954) suggests that interaction between natives and immigrants (majority and minority group) can reduce negative attitudes of majority group members toward the minority group and -under some circumstances -tackle xenophobic fears.On the other hand, 'group threat theory' (see, e.g., Blumer 1998;Quillian 1995) suggests that it is actually the presence of immigrants that causes and exacerbates such animus.
Empirically, it is challenging to credibly document the impact of immigrants on their host societies, because immigrants typically make residential choices based on private information about their destinations, such as local labor market conditions or xenophobic attitudes.Since these local conditions or attitudes are likely to also influence support for right-wing parties, refugee self-selection poses a serious challenge to causal inferences about the effect of immigration (see Dustmann and Preston 2001).In other words, it is inherently difficult to distinguish between co-occurrence of refugees and local attitudes towards immigration among natives and the isolated effect of an increase in refugee migration on native citizens' attitudes and political behavior.We believe that the endogeneity issues are potentially severe, which is one reason why we largely focus on studies that employ research designs that can credibly claim to identify causal effects.
Another challenge has to do with the fact that citizens experience immigration through different channels.Immigration, and newly arriving asylum seekers especially, are one of the most salient topics in the media.Almost on a daily basis, newspapers report on asylum-related issues (see, e.g., Spirig 2021).At the same time, experiences with refugees can also take place on the micro-level.Citizens who live in a community that hosts asylum seekers might see and/or meet asylum seekers on a daily basis.As Hopkins' (2010) study on anti-immigrant attitudes in the U.S. shows, these channels are not independent from each other.He presents evidence that "at times when rhetoric related to immigrants is highly salient nationally, those witnessing influxes of immigrants locally will find it easier to draw political conclusions from their experiences" (Hopkins 2010, 44).This finding points to several contextual factors that could moderate the influence of refugee arrival on political outcomes (via its impact on voters' attitudes).We discuss this in more detail below.In addition, while the macro-level channel might be more impactful on individuals' attitudes this is particularly challenging to empirically substantiate -among other reasons, because of the lack of exogenous variation.This is why we emphasize the importance of factors determining citizens' macro-level experience of the arrival of asylum seekers -in particular the salience and framing of the issue in the media and political debates -but focus our review mainly on studies exploring regional differences in right-wing, antiimmigrant party support within states, rather than between states.This focus on regional variation in all likelihood underestimates the total (regional and national) effect of immigration on right-wing, anti-immigrant parties (see, e.g., Steinmayr 2021).
Empirical studies that look at the micro-level and are able to tackle these inferential challenges do not provide a uniform answer as to what the direct effect of immigration on political outcomes is.The results spread from a significant negative effect of accommodating asylum seekers on right-wing party support (Steinmayr 2021), to a strong positive effect of experiencing asylum seeker arrivals on extreme-right party support (Dinas et al. 2019;Steinmayr 2021).The following section summarizes this range of studies.
The recent study by Steinmayr (2021) is one of the first to find an overall negative effect of accommodating asylum seekers on (an increase in) far-right party support: Upper Austrian municipalities that accommodate asylum seekers display an about 3.9 percentage point smaller increase in the 2015 state election vote share for the anti-immigrant FPÖ than municipalities that did not host refugees.According to the study, qualitative interviews imply that "in almost all cases, the level of anxiety declined after the asylum seekers had been there for some time since most of the feared consequences did not materialize" (Steinmayr 2021, 321).In new working papers on the effect of refugee reception centers in Italy and France, Gamalerio et al. (2021) and Vertier, Viskanic, and Gamalerio (2020), respectively, provide additional evidence that when meanigful contact between locals and refugees is promoted and possible (i.e., the number of refugees is small), the presence of refugee centers leads to a reduction in right-wing party support.
Two studies, one focusing on the effects of asylum seeker arrivals on support for the right-wing AfD in Eastern Germany (Schaub, Gereke and Baldassarri 2021) and one focusing on refugee arrival in Danish municipalities (Jensen 2020), document null effects on party vote shares.In addition, Savolainen (2016) examines electoral effects of opening asylum centers in Finnish municipalities in 1990-2011, and finds neither evidence for an impact on anti-immigration nor proimmigration parties' vote share. 7 relatively large set of studies documents positive effects of asylum seeker arrivals on support for anti-immigrant parties.Bratti et al. (2020) show that in Italy, proximity to refugee centers that do not provide integration services increases support for populist parties.Dustmann, Vasiljeva, and Damm (2019) focus on refugees and find that not only votes for the right-wing party, but also the center-right party increase with larger shares of refugees being allocated to a municipality.An increase in the municipality's refugee share by one percentage point leads to a 1.2 percentage point increase in the anti-immigration party's vote share in parliamentary elections, and also the center-right party benefits.Furthermore, they document that anti-immigrant parties are more likely to run in municipal elections when refugee shares are higher and that this relationship does not hold in big urban areas. 8Hangartner et al. (2019) and Dinas et al. (2019) examine the effect of asylum seekers passing through Greek islands close to Turkey.These two papers find that the passing-through of asylum seekers, who do not stay but continue onwards with their journeys, leads to "lasting increases in natives' hostility toward refugees, immigrants, and Muslim minorities; support for restrictive asylum and immigration policies; and political engagement to effect such exclusionary policies" (Hangartner et al. 2019: 442) and, in the short run, an increase of two percentage points (more than 40% at the mean) in the vote share of Golden Dawn, arguably the most extreme rightwing and anti-immigrant party holding office in Europe (Dinas et al. 2019).Steinmayr (2021) relatedly finds that support for the far-right Austrian FPÖ increased about 1.5 percentage points more in municipalities that asylum seekers passed through and did not stay as opposed to comparable municipalities that did not see asylum seekers pass through, and Gessler, Tóth and Wachs (2021) document a similar effect for Hungary.
Taken together, these studies indicate that the (short-term) effect of the presence of newly arrived asylum seekers on electoral outcomes and support for far-right parties is not only theoretically, but also empirically unclear.The most natural interpretation given the different contexts under study is that the effects depend strongly on moderating factors, such as the facilitation of inter-group contact (see, e.g., Steinmayr 2021), the size of the refugee arrivals (see, e.g., Dinas et al. 2019), pre-existing political attitudes (see, e.g., Dustmann, Vasiljeva, and Damm 2019), and macro-level determinants, such as issue salience (see., e.g., Hopkins 2010).Yet, there is a lack of research as to which moderating factors are most crucial when it comes to political consequences of refugee immigration.To gauge which moderating factors are the most relevant, however, we briefly consider a larger set of studies investigating the effect of immigration (in different forms) on political outcomes in the following.
While there are many studies looking at the effect of (refugee) migration on electoral outcomes, there is very limited attention to other outcomes such as native citizens' preferences for redistribution or trust in political institutions.McLaren (2012McLaren ( , 2015) ) is one of the few exceptions, however.Her research focuses on the impact of anti-immigrant attitudes on political trust and argues that some voters feel that immigration threatens a sense of national identity that lies at the heart of the liberal state and/or want to hold the state accountable for 'failing' to control immigration adequately.Accordingly, immigration not only fuels anti-immigrant attitudes, 9 but might also lead to a decrease in political trust: "politicians and institutions are likely to be blamed for failing to control immigration adequately" (2012,171).In a similar vein, a few studies focus on the effect of ethnic diversity on social trust (for an overview, see Dinesen, Schaeffer, and Sønderskov, 2020) and attitudes regarding welfare state spending (Dahlberg, Edmark, and Lundqvist 2012).Both of these papers find a negative effect of ethnic diversity on mentioned outcomes.Dahlberg, Edmark, and Lundqvist (2012) investigate the effect of the arrival of refugees in assigned Swedish municipalities on voters' responses to a survey on welfare state spending and find evidence for the so-called 'in-group bias': Individuals display lower preferences for redistribution if the share of refugees placed in their municipality is larger. 10  The literature dealing with the impact of immigration -studies focusing on asylum sekers are rare -on more extreme forms of political behavior, such as political violence, does not provide a uniform answer, either.Dancygier (2010) investigates the impact of immigration on violent attacks and documents a positive relationship between the two in Greater London.Braun and Koopmans (2010) find similar effects in Germany, as do Krueger and Pischke (1997) in the German Democratic Republic, but not in Western Germany.Accordingly, Krueger and Pischke (1997) note that local political processes -how local conflicts are handled -play a large role: Immigration and violent outbursts are positively correlated, if local political processes facilitate mobilization.A qualitative study by Karapin (2002) also suggests that whether immigration by ethnic minorities led to violent, anti-immigrant riots in 1990s Germany depended on local political processes, such as, among others, facilitation of non-violent political participation.Analyzing more recent data from Germany, Marbach and Ropers (2018) find that in times when immigration is salient on the national level, increases in the number of asylum seekers at the local level are associated with more anti-asylum seeker violence on the local level.Falk, Kuhn and Zweimüller (2011), however, find very little, or no impact of the size of the immigrant community on political violence.
In sum, there is a range of studies employing credible research designs that suggest that immigration leads to political shifts: More immigration appears to lead to increases in votes for anti-immigration, and typically right-wing, parties, more political violence directed at immigrants, and potentially lower levels of political trust and preferences for redistribution.Accordingly, this research sheds light on potential political repercussions of receiving asylum seekers.Rises in right-wing authoritarian attitudes, erosion of trust in political institutions and democratic governance, and electoral success of extreme-right parties (such as, for example, in Greece) have the potential to fundamentally affect the civic fabric of a society and undermine the credibility of its political system.This democratic backsliding could prove to be much more substantial than the short-term negative economic and fiscal consequences of an increase in the number of arriving asylum 9 Like, for example, in the UK, where immigration was not dominating elections in the past because major parties did not found their campaigns on it. 10Nekby and Pettersson-Lidbom (2016) question the validity of Dahlberg, Edmark, and Lundqvist's (2012) identification strategy and the representativeness of their sample (see also Dahlberg, Edmark, and Lundqvist's (2016) reply).
seekers.However, it is crucial to note that taken together, the studies discussed above also imply that the effect of asylum immigration on political outcomes is not deterministic.The consequences appear to be highly dependent on the political context and policy choices.While we gauge the potentially substantial second-order effects below, Section 4 will discuss which policies are most likely to reduce the political effects of refugee protection crises.

B. Second-order effects
In contrast to asylum immigration's first-order effects on (far-)right party support, there is much less research on its second order-effects: How does the immigration-fueled rise of anti-immigrant parties affect isolationist policies and the political platform of mainstream parties?The existing literature suggests a twofold answer.First, anti-immigrant parties (e.g., the French Rassemblement National or the U.K.'s Brexit Party) or candidates (e.g., President Trump) tend to favor isolationist policies such as exiting the European Union, reinstating border controls, and -more recently -taking antiglobalization stances more broadly (see, e.g., Walter 2021), such as curbing international trade.Second, the rise of populist parties may affect policy even if these parties do not enter the government.This view is supported by a related literature in political science examining how successes of far-right parties exert electoral pressures on mainstream parties, and are thereby able to shift government parties' policy positions closer to their ideal point (see, e.g., Abou-Chadi and Krause (2020), Spoon and Klüver (2020), as well as the review article by Golder (2016) and the references therein).
A prominent example of such second-order effects of asylum salience is the Brexit referendum.According to Moore and Ramsey's (2017) analysis of all articles published by leading U.K. news outlets during the 2016 EU referendum campaign, immigration, and in particular asylum migration, was the most prominent campaign issue (based on the number of times it led newspaper print front pages), with almost 80% of them appearing in Leave-supporting newspapers.In contrast, economic issues and, in particular, the vexing question of the impact of leaving the Single Market, received significantly less frontpage attention.While assessing the impact of the UK leaving the EU is extremely challenging, it seems likely that the significance of refugees for the U.K. economy is miniscule, and the connection to the question of EU membership tenuous, especially in comparison to the importance of accessing the European Single Market.This focus on the issue of asylum migration threatened to crowd out attention on the many other, and arguably more consequential, legal and economic issues tied to Brexit.

How to mitigate the fiscal and political risks?
The research reviewed above illustrates the gravity of challenges and constraints policy makers face when their countries receive refugees and asylum seekers.Thus far, we have provided little guidance on how to respond to these challenges.In this section, we review research that we hope will help decision makers to design appropriate policy responses.We first provide an overview of voter preferences and proceed with a relatively thorough review of the available evidence on the impacts of alternative reception and allocation of asylum seekers, integration policies and language training as well as eventually giving refugees permanent residence permits and citizenship.

A. Voter preferences
We first survey the literature on voter preferences about refugee policy.This research suggests that across all Western countries, the majority of residents have a skeptical view of current asylum policies and prefer to curb the number of future asylum seekers.However, they also tend to support the acceptance of refugees deserving of asylum and are willing to accept more asylum-seekers as long as the allocation is 'fair' and proportional to the country's capacity.
The increase in asylum applications and the differences across European countries have not gone unnoted.Asylum law experts and policy-makers alike have repeatedly expressed concern about the unequal distribution of asylum seekers across Europe, the Dublin regulation, and how the EU is handling the increasing pressure at its borders (see, e.g., Thie-lemann 2010; Angenendt, Engler, and Schneider 2013;Malmström 2014).Furthermore, citizens across Europe share the impression that the Dublin system is unfair.Employing an online survey experiment involving 18,000 voters across fifteen European countries, Bansak, Hainmueller and Hangartner (2017) find that only 18 percent of respondents support the current Dublin regulations, which state that asylum seekers usually have to submit their claim in the European country of first entry.Interestingly, the support is very low even in countries that benefit from the current status quo in the sense that they receive relatively few asylum claims.In stark contrast, 70 percent of respondents prefer proportional allocation of asylum seekers based on the country's capacity (a function of population size, GDP, unemployment rate, and number of past applications).When voters are randomly prompted with the actual numbers of asylum seekers their country would receive under each allocation, they are somewhat more likely to support the allocation that yields the lowest number of asylum seekers for their own country.However, even under this treatment condition, in all but three countries (Czech Republic, Poland, and the UK) a majority of voters prefers proportional allocation over the status quo.(Note that ten out of the fifteen countries would have to host more asylum seekers under proportional allocation.)These findings indicate that a majority of citizens is willing to provide refuge to additional asylum seekers as long as they know that the overall allocation across 'Dublin countries' is proportional to a country's capacity.
In a companion paper based on the same survey, Bansak, Hainmueller and Hangartner (2016) employ a conjoint analysis asking the 18,000 respondents to evaluate fictitious profiles of asylum seekers that randomly varied along personal attributes.They find that asylum seekers who are highly skilled, contribute to the host country's economy, have more consistent asylum testimonies and severe vulnerabilities, and are Christian rather than Muslim receive the greatest public support.Bansak, Hainmueller and Hangartner (2016) argue that these results point to tough challenges for policy makers who are struggling to meet their legal responsibilities to protect refugees in line with the 1951 Refugee Convention.The public's strong anti-Muslim bias and preference for highly skilled asylum seekers who can speak the language of the host country hinder the acceptance and integration of asylum seekers given that most currently originate from Muslim-majority countries and may lack the desired professional and language skills.At the same time, Bansak, Hainmueller and Hangartner (2016) argue that the findings also point to opportunities for policy makers: the fact that citizens across Europe share common humanitarian concerns for refugees with consistent asylum claims suggests that large segments of the public have at least partially internalized the central pillars of international refugee law.

B. Reception and allocation of asylum seekers
We next discuss lessons from the existing literature on how to organize the reception of asylum seekers in a way that minimizes the potential for anti-immigrant backlash.Our suggestions include providing timely information to locals; facilitating repeated and meaningful contact between locals and newly arrived refugees; ensuring that the allocation of asylum seekers across and within countries is perceived as 'fair'; and taking into account that rural and more conservative constituencies tend to exhibit stronger backlash against refugees.
First, a distribution of asylum seekers across receiving communities that is perceived as fair by voters increases support for reception.As discussed in detail above, Bansak, Hainmueller and Hangartner's (2017) cross-country survey that examines Europeans' attitudes towards different asylum seeker allocation mechanisms shows that most citizens in Europe prefer a proportional allocation of asylum seekers over the status quo of country of first entry.
Second, large, liberal municipalities (big urban areas, cities) appear to exhibit smaller (attitudinal and) electoral responses to increases in asylum immigration.Right-wing, anti-immigrant parties usually do not receive their largest support from cities (see Dustmann, Vasiljeva, and Damm 2019).In addition, both Barone et al. (2016) andDustmann, Vasiljeva, andDamm (2019) show that in large urban areas, immigration does not increase right-wing party support in Italy and Denmark, respectively.Accordingly, it seems worth to take this strong treatment effect heterogeneity into consideration when deciding where to host asylum seekers.
Third, macro-level salience matters.Therefore, information campaigns and narratives prior to the arrival of asylum seekers could help mitigate initial backlashes.Hopkins' (2010) study on attitudinal changes of Americans in response to im-migration highlights the importance of the macro-level narrative on immigration and how that might influence and structure citizens' perception of face-to-face encounters with refugees in their community.

C. Length of the asylum process
Once the asylum seekers have been received and allocated to asylum centers, policy makers need to decide how much resources to use for their applications to be processed.The available resources, in turn, largely determine how fast the applications can be processed.Hainmueller, Hangartner and Lawrence (2016) show that the processing time affects subsequent integration of refugees into the host society.More specifically, they provide evidence as to how the length of time that refugees 'wait in limbo' for a decision on their asylum claim impacts on their subsequent economic integration.Exploiting exogenous variation in wait times and using registry panel data covering asylum seekers who had applied for asylum in Switzerland between 1994-2004, they find that one additional year of waiting reduces the subsequent employment rate by 4 to 5 percentage points, a 16% to 23% drop compared to the average rate.This deleterious effect is remarkably stable across different subgroups of refugees stratified by gender, origin, and age at arrival.The findings of Hainmueller, Hangartner and Lawrence (2016) are consistent with previous cross-sectional and qualitative evidence (Stepick and Portes 1986;Waxman 2001;Bakker, Dagevos and Engbersen 2014) suggesting that waiting in limbo dampens refugee employment through psychological discouragement, rather than a skill atrophy mechanism.In other words, whereas recent reductions in refugees' labor market access waiting times point to the importance of early labor market access for economic integration, Hainmueller, Hangartner and Lawrence ( 2016) highlight an additional factor that affects asylum seekers' ability to integrate: the degree of, and time period, in uncertainty about the future.Their partial equilibrium cost benefit analysis suggests that even policy reforms marginally reducing the waiting period for asylum seekers would help refugees to navigate the difficult transition from a life in legal limbo to a successful integration into the host community better.Moreover, from a host country perspective, such reforms would reduce public expenditures for welfare benefits significantly due to the increase in employment and the resulting increase in tax contributions of employed refugees.

D. Integration programs and active labor market policies
Many countries provide integration programs for refugees and asylum seekers if their applications are approved.These integration programs are often arranged as part of active labor market policies and are offered also to other unemployed immigrants.We next review research based on plausible research designs that aim to evaluate the impacts of these programs.Overall, these studies suggest that integration programs can be remarkably efficient in increasing employment and earnings of refugees (as well as other immigrants struggling to find their way into the host country's labor market).
Andersson Joona and Nekby ( 2012) study a Swedish program, where newly arrived immigrants were provided extensive counseling and coaching on employment prospects.A trial of this intervention was conducted through Public Employment Service (PES) in 2006-2008.The caseworkers participating in intensive coaching were trained to work exclusively with newly arrived immigrants and would handle less than 20% of the caseload in comparison to regular caseworkers.The intervention aimed to facilitate direct contacts with employers and to improve the match quality between the immigrants and ALPM measures.The evaluation was conducted as a randomized controlled trial (RCT), where PES officers were asked to randomly assign newly arrived immigrants into treatment (intensive coaching) and control (regular introduction programs) groups.
The results suggest that intensive coaching increased the share of immigrants in regular employment by 6 percentage points two years after the start of the intervention.Given that the employment rate of the control group was only 14 percent, this corresponds to 43% increase in employment rate.The effect was sufficient to cover the cost of the program in 2-3 years.Furthermore, the overall effect appeared to be due to men being responsive to the treatment, while no impact was found for women.
An unfortunate feature of this trial is that the treatment and control groups were not fully balanced on pre-assignment characteristics.A possible reason for the unbalance is that the PES officers conducted the randomization and may not have followed the randomization protocol.In particular, they may have attempted to select better participants into the program.If this were the case, the estimated impact of the intervention would overstate the true impact.This hypothesis is supported by the fact that controlling for observed characteristics reduces the estimated employment impact from seven to six percentage points.While this reduction is not particularly dramatic, it suggests that the estimates should be interpreted as upper bounds for the true impact.Furthermore, the results highlight the importance of avoiding situations where parties who might have a stake in the results are responsible for the randomization process.
Åslund and Johanson ( 2011) examine an earlier Swedish intervention called Special Introduction Programs (SIN), piloted in 2003.It was based on methods originally used for helping workers with disabilities to find employment and focused on immigrants and refugees who were considered to be at risk of becoming long-term unemployed.PES offices executed the intervention and the participating offices were given additional funding to hire caseworkers.These caseworkers had only 10% of the caseload of regular caseworkers, which allowed them to work intensively with each of their clients.
The intervention consisted of caseworkers finding suitable jobs for their clients and running an introductory session in these jobs together with the employer, colleagues, and union representatives.This was followed by an internship period lasting up to six months after which the caseworker organized a follow-up session with the aim of turning the internship into a regular job.
The program was piloted in 20 Swedish municipalities in 2003.Åslund and Johanson (2011) evaluate its impacts using a difference-in-differences strategy, where they compare changes in treatment municipalities with changes in non-participating locations in the same local labor market.They find that the intervention increased transitions from unemployment to work experience schemes and improved future employment probabilities for those who entered these schemes.
In a recent contribution, Dahlberg et al. (2020) present evidence from a randomized control trial implemented in Gothenburg, Sweden in 2016-2020.The program targeted newly arrived, low-educated refugees and included language training, supervised work practice, job search assistance, and extended cooperation between the local public sector and firms.The baseline services provided to the control group include all of these elements, but the experimental intervention was much more intensive.For example, the baseline language training provided 15 hours of teacher per week, while those participating in the experimental program received 40 hours per week.The results suggest that these additional investments doubled employment rates during the first year following the program's end (from 15% among to control group to roughly 30% among the treatment group).
In comparison to these rather intensive Swedish programs, the "integration plans" introduced in Finland in 1999 were very light and cheap.These integration plans are prepared in a joint meeting of a caseworker, the immigrant, and an interpreter with the aim to find a sequence of training and other measures that would be the most suitable for each immigrant.In principle, similar meetings took place with all unemployed immigrants already before the reform.However, the integration plans aimed to improve the communication between caseworkers and immigrants.For example, the new guidelines stated that the caseworker had to make sure that the immigrant fully understood the content of her integration plan and knew how to follow it.In addition, the reform aimed to increase the caseworker's capacity to better take into account the specific skills and circumstances of each immigrant.
Sarvimäki and Hämäläinen (2016) evaluate the impact of the integration plans using a research design based on a phasein rule dictating that participation was mandatory only for unemployed immigrants who had entered the population register after May 1st, 1997.This rule creates a discontinuity, where there is a 35 percentage point difference in the likelihood of receiving an integration plan between those who had arrived on May 1st, 1997 and those who had arrived slightly earlier.Comparing the two otherwise similar groups shows that those who had arrived on May 1st, 1997 and were thus much more likely to receive an integration plan earned cumulatively roughly €7,000 more in the ten-year follow up period than those arriving just before the specified date.Scaling this effect with the change in the likelihood of receiving an integration plan suggests a local average treatment effect of 47% increase in cumulative earnings.Using the same approach, they also find a 13% decrease in the reception of cumulative social benefits.These effects seem to be due to in-creased language training and other training courses specifically designed for immigrants, replacing more "traditional" active labor market training such as job-seeking courses.That is, there is no detectable impact on the overall amount of training.Furthermore, Pesola and Sarvimäki (2021) find that the integration plans also have large intergenerational effects on grades and educational attainment of the children of the affected immigrants.Arendt et al. (2021) examine the impacts of another reform on integration policies using a similar empirical approach as Sarvimäki and Hämäläinen (2016).Refugees arriving to Denmark after January 1st, 1999, were required to take substantially more language training than those arriving before the cutoff date.The reform also reduced welfare benefits available for some refugees, altered the way refugees were allocated across municipalities and shifted the responsibility to provide integration training from the central government to municipalities.These changes increased refugees' employment and earnings and facilitated skill-upgrading.In addition, male children of refugees whose both parents arrived just after the threshold date were more likely to complete lower secondary school and committed fewer crimes.
The research discussed in this section shares two central themes.First, all studies examine interventions aimed at improving the match quality between immigrants and training programs.Second, all papers find much larger effects than what is typically documented in the literature on the impacts of active labor market policies on natives' labor market integration (see e.g.Card, Kluve and Weber 2010, 2018 for reviews).These observations are consistent with the hypothesis that refugees (as well as some other immigrant groups) may lack the type of skills that can be improved through training provided by the employment offices, and that they first of all need support navigating the system.Thus, even the very small interventions such as Finland's integration plans can have large effects.These findings also suggest that further policy experimentation on how to improve training and counseling could yield high returns on the public investment.

E. Proficiency of local language
There is little doubt that proficiency in the host country language is crucial for a successful integration.Yet, estimating the causal effect of language on immigrants' integration is challenging because the correlation of language proficiency and labor market outcomes raises a well-founded fear of endogeneity.11In addition, the general focus on language learning (almost) throughout this literature entails critical challenges for the estimation of the importance of language on labor market outcomes.Most importantly: The fact that language learning ability is correlated with many other, often unobservable, characteristics that could also influence immigrants' job search and earnings is often noted, but difficult to overcome.Yet, even though many of the studies discussed below might not be able to deliver causal estimates, they provide, at the very least, an upper bound on the benefits of language courses, as the following summary illustrates.
Labor market participation seems to increase with local language skills in various countries.Dustmann and Fabbri (2003) find that in the UK, English language acquisition (ELA) increases the chances for a male job-seeker to find gainful employment by 26 percent.For women, the estimates are not statistically significant.In addition, they also find a significant positive effect of English proficiency on earnings. 12Grondin (2007) shows that the same positive relationship between English speaking ability and probability of employment also exists in Canada.Aldashev, Gernandt and Thomsen (2009) find that in Germany, language proficiency does not only affect immigrants' labor market participation, chances of employment, and earnings, but also their occupational choice.
A positive effect of local language skills has also been documented for earnings.In an early analysis, Tainer (1988) finds a statistically significant positive effect of English proficiency for foreign-born men in the U.S. The extent of the effect, however, varies across ethnicities: There is a larger effect for Hispanics and Asians than for European-born men.Chiswick and Miller (1995) analyze the impact of English language fluency on earnings in Australia, Canada, the U.S. and Israel.In all countries, they find a significant positive effect, varying between 5.3 percent higher earnings in Australia and 16.9 percent higher earnings in the U.S..The results are also confirmed in later studies for Israel (Chiswick and Repetto 2000), the U.S. (Chiswick and Miller 2002) and Canada (Chiswick and Miller 2003).Bleakley and Chin (2004) use an instrumental variables strategy based on age of child's arrival in the U.S. and her source country's language to determine the effect of English language skills in the U.S..They find a significant positive effect of language ability on earnings, arguably mainly driven by years of schooling.Dustmann (1994) confirms this positive effect also for German language ability in former-West Germany. 13Simliarly, Ispohrding, Otten and Sinning (2014) find a strong positive effect of language ability on wages, arguably mainly mediated through occupational choice.Similar results also exist for ELA in the UK.Shields and Wheatley Price (2002) and Miranda and Zhu (2013), both using an IV strategy, estimate a large positive effect of ELA on wages. 14Finally, Budría and Swedberg (2015) find that in Spain as well, there is a general positive effect of language abilities on earnings, but it is more pronounced for high-skilled workers.They earn about 50 percent more if they speak Spanish.
Focusing on France, Lochmann, Rapoport and Speciale (2019) leverage a discontinuous assignment role to governmentoffered language training to document a significant effect of assignment to training on labor force participation of immigrants.This effect is increasing in the immigrants' education levels.Using administrative data from Switzerland, Hangartner and Schmid (2021) are also able to address above-mentioned concerns about endogeneity with a differencein-differences design.They exploit the quasi-random placement of refugees to Swiss states (cantons) and the existence of a sharp language border dividing German and French-speaking areas and examine the size of the economic gains from proficiency of the host country's language.Compared to otherwise similar English-speaking African asylum seekers, French-speaking asylum seekers have an 80 percent higher probability of finding a job in the first year after arrival due their proficiency in French.This effect is persistent for at least the first five years upon arrival.
Despite some shortcomings, these studies leave little doubt about the importance of proficiency in the host country's language of immigrant integration.They suggest that for arriving asylum seekers' economic integration and the receiving country's public expenditures, providing extensive language training to asylum seekers (and future residents) could prove highly beneficial.

F. Permanent residency permits and citizenship
Another policy that has the potential to facilitate the integration of immigrants is to allow for faster access to permanent residency and citizenship.Faster access to a more permanent form of residence eliminates fears of deportation, and at the same time incentivizes immigrants to invest in a long-term future in the host country.However, that does not necessarily imply that residency permits and citizenship should be offered at the earliest stage.In theory, lowering the threshold for residency permits and citizenship could also have the opposite effect: Rather than incentivizing integration, issuing residency permits and citizenship too early might destroy immigrants' strive to integrate and learn the local language (Hainmueller, Hangartner and Pietrantuono 2017).While causal evidence on the impact of permanent residency and citizenship is fairly scant, there are a few studies that generally show that giving immigrants permanent residency and citizenship has i) a positive effect on political and social and, to a lesser degree, economic integration and ii) that these 'integration returns' are larger if immigrants receive these statuses earlier in the residency period.
14 Shields and Wheatly Price (2002) estimate a positive effect of about 16.5% on immigrants' mean hourly occupational wages and Miranda and Zhu (2013) estimate that English deficiency leads to 23% lower wages in the UK.
With regard to residency permits, Hainmueller, Hangartner and Lawrence (2016) provide panel data evidence based on Swiss registry data (see above) that asylum seekers probability of finding a job increases by 10 percentage points (a 50% percent increase over the average) if they receive subsidiary protection in the first year after arrival.The boost associated with subsidiary protection, which arguably captures both the increase in refugees' motivation of finding work and decrease in employer's uncertainty about the refugee being deported, fairly linearly decreases the longer the refugee has to wait for receiving protection status and is essentially zero after five years upon arrival.
In the domain of citizenship rights, there are several panel data studies that show a positive association between naturalization and labor market outcomes (see Bevelander andVeenman 2008 andOECD 2011 and the references therein).One common problem with these studies is that even when employing panel data, the coefficient for naturalization might not have a causal interpretation if an unobserved factor, such as the decision to stay in the host country for good, causes immigrants to simultaneously apply for citizenship and finding a (better) job. 15However, Gathmann and Keller (2014) can exploit discontinuities in eligibility rules for immigration reforms in Germany that changed the residency requirements for naturalization.Based on an intention-to-treat analysis, they find only few economic returns for men, but significant, albeit substantively small, returns for women.
To circumnavigate the confounding issue associated with panel data, Hainmueller, Hangartner and Pietrantuono (2015;2017;2019) exploit the quasi-random assignment of citizenship in Swiss municipalities that used referendums to decide on naturalization applications of immigrants.Comparing otherwise similar immigrants who narrowly won or narrowly lost their naturalization referendums, they find that receiving Swiss citizenship strongly improved long-term economic, political and social integration.More specifically, they present evidence that barely naturalized -as compared to barely non-naturalized -immigrants have higher earnings, higher levels of political efficacy and knowledge, are more likely to read also Swiss and not exclusively foreign newspapers, are less likely to plan to return to their (or their parents') country of origin, and are less likely to feel discriminated against.Using an index of these outcomes, their studies show that naturalization increases both political and social integration by one standard deviation.They also find that the integration returns to naturalization are much larger for more marginalized immigrant groups16 and somewhat larger when naturalization occurs earlier, rather than later, in the residency period.
Taken together, these studies support the policy paradigm arguing that naturalization is a catalyst for improving the economic, political and social integration of immigrants -rather than merely the crown on the completed integration process.

G. Fostering meaningful interaction between locals and refugees
In combination, recent studies suggest that it may be important that interactions between asylum seekers and locals are meaningful and sustained (as opposed to mere exposure).Several of the papers discussed in this review speak to the importance of how contact happens.Steinmayr (2021) shows that the clear trend towards more support for right-wing, anti-immigrant parties overall was less extreme in municipalities that were assigned to host asylum seekers.In these municipalities, the arrival and integration of asylum seekers was accompanied, encouraged and facilitated by professionals and volunteers (see also, Gamalerio et al. (2021) on integration-promoting refugee reception centers in Italy).Large positive effects of arriving asylum seekers on anti-immigration party support, however, were documented where refugees predominantly just passed through on their way to other European countries (Dinas et al. 2019, Steinmayr 2021).This implies that the problem-centered media coverage of asylum issues (see, e.g., Eberl et al. 2018) might be less effective at structuring locals' perception of asylum seekers if direct contact is meaningful and repeated (see also Allport 1954, Hopkins 2010).
Recent papers focusing on individual interactions between members of different groups complement the observational evidence mentioned above.Mousa (2020) shows that repeated interactions on the soccer pitch are able to reduce exclusionary attitudes between Christians and Muslims even in a challenging post-conflict context.In addition, not only repeated interaction between asylum seekers and the local community, but also between the (local) government and the local community could prove important.Both Krueger and Pischke (1997) and Karapin (2002) indicate that local political participation seems an important moderator for political violence against immigrants.Despite a huge literature on the 'contact hypothesis' in social psychology and related fields, there is still limited experimental and actionable evidence for policymakers about how best to foster interaction (Paluck and Green 2009).This remains one of the most important avenues for further research.

Conclusions
This paper was inspired by our conversations with policy makers, journalists and fellow researchers in and outside of academia.Our role has been to present summaries of the research on labor market and fiscal impacts of (asylum) immigration, political effects of asylum seeker arrival and presence, and the effectiveness of various integration policies.Sooner or later, these conversations inevitably gravitated towards the question: "OK, but what should we do?" In this paper, we offer a twofold answer.First, we argue that an essential component of the response to increases in asylum seeker arrivals is to remain calm.We acknowledge that this may be a formidable task as media coverage of refugees tends to capture the public imagination and worries about arriving refugees may thus receive a disproportionate weight in public debate.We do not have to look far for anecdotes supporting this view.For example, fears about refugees loomed large in the debate preceding the vote for Britain's exit from the EU despite the UK receiving relatively few asylum seekers in 2015.While it is hard to predict the economic impact of the UK leaving the EU, it seems safe to assume that these effects are likely substantially larger than the direct labor market or fiscal effect of refugees living in the UK.More generally, heated public debate increases the risk that important decisions will be made without a sufficient analysis of their first-order effects.
We stress that we do not make a statement about policy objectives, but on the quality of decision-making.Regardless of the objectives, cool heads are needed to evaluate whether the proposed policies are likely to lead to the desired outcomes.We also recognize that just telling people to calm down is unlikely to be helpful.A substantial share of voters holds deeply skeptical views of current refugee policies, and a growing literature shows that higher numbers of arriving refugees fuel the rise of populist, anti-immigrant parties.We do not advice policy makers to neglect these facts.Rather, we argue that these findings highlight the importance of seeking ways to mitigate the impact of receiving refugees on the broader political process.
Accomplishing this goal likely requires a multifaceted approach, but we view efficient integration policies as a central part of a response.Existing work suggests that such polices can have surprisingly large effects.However, we still lack a sufficient body of research to determine which policies are the most efficient (and for whom).
The second part of our recommendation is thus to increase policy experimentation and evaluation.We would particularly like to see more work on integration policy and programs, and on designing the asylum process (length of the process, labor market access, welfare support) with an eye towards rapid integration of asylum seekers who have a high likelihood of obtaining some form of protection status.Piloting new policies in a way that they are amenable to evaluation is, in principle, relatively straightforward.Many interventions can be tested with RCTs (for example by randomizing the timing when a new policy is implemented) or, alternatively, by creating research designs through the staggered rolled out of policies or the use of discontinuities in eligibility criteria.Given the prominence of refugees in the policy debate, researcher may have a good chance to persuade governments to engage in such experimentation.For example, the Finnish government is already running a large RCT to test a new approach for improving employment of refugees.We believe that similar opportunities are available also elsewhere.□

Figure 1 :
Figure 1: New asylum applications since 1980 in the OECD and the EU

Figure 2 :
Figure2: Scenarios of the long-term fiscal impact of immigration fromSarvimäki et al. (2014).See Section 3.2 for discussion.