Immigration, Globalization, and Unemployment Beneﬁts in

Immigration, Globalization, and UnemploymentBenefits in Developed EU States

Christine S. Lipsmeyer Texas A&M UniversityLing Zhu Texas A&M University

At a time of mounting concern about how traditional welfare states will react to globalization, there has been increasinginterest in specifying how global economic forces affect welfare policies in industrialized states. Building on theoriesfrom the political economy and comparative institutional literatures, we analyze the influence of an important aspect ofglobalization—the flow of immigration. Focusing on states in the European Union, we present a theoretical model thatillustrates the interactive relationships between immigration, EU labor market integration, and domestic institutions. Ourfindings highlight how immigration in conjunction with domestic political institutions affects unemployment provisions,while labor market integrative forces remain in the background. The story of immigration and unemployment compensationin the EU is less about the opening of borders and the market forces of integration and more about the domestic politicalpressures.

Immigration has become a crucial issue for Euro-pean welfare states. The economic integration pro-cess has loosened border controls among most Eu-

ropean Union member states, generating large flows ofimmigrants looking for employment (Nannestad 2007).Not surprisingly, a substantial body of literature on laborimmigration and Western welfare states argues that im-migration tends to be a disadvantage for recipient stateswhile an advantage for immigrants themselves (Banting2000; Hunger 2000; Nannestad 2007). Immigrants canfind new jobs, but existing labor markets may strain underthe burden of additional workers, adding to the pressureon welfare budgets.

In both the comparative welfare policy and global-ization literatures, researchers have overlooked how thismovement of workers may affect governments’ unem-ployment policies. Rather than viewing globalization asa capital-led enterprise, we theorize about it from thestandpoint of labor, emphasizing the intertwining of la-bor markets across EU states. By starting from a neoliberalscenario of a perfect labor market and distorting the mar-ket with integrative and political institutions, we balance

Christine S. Lipsmeyer is Assistant Professor ([email protected]). Ling Zhu is PhD Candidate ([email protected]). Bothcan be reached at Department of Political Science, TAMUS 4348, Texas A&M University, College Station, TX 77843.

We presented previous versions of this article at the 2009 annual meeting of the Midwest Political Science Association and the 2010Maastricht Conference of the Association for Public Policy Analysis and Management. We particularly thank the editor and the anonymousreviewers for helpful comments; any remaining errors are ours alone. Replication data for our model can be found at http://www-polisci.tamu.edu/faculty/lipsmeyer/.

the literatures’ typical foci on capital-driven theories byfocusing on the integration of labor markets.

We are interested in uncovering whether im-migration, combined with integration, influencesunemployment provisions. Our theory is that the influ-ence of immigration depends on the economic pressurescaused by the globalizing effects of integrating the EUlabor market, so the effect of immigration is constrainedby institutional arrangements that can affect the supply-demand relationship in labor markets. We argue that bothcross-state pressures (i.e., EU labor market integration)and domestic political institutions are important factorsfor understanding how economic pressures are trans-ferred into policy changes. In both cases, they alter theexpected outcome of the labor market, causing changesin unemployment benefits.

Using pooled time-series data for 15 developed EUstates, we systematically assess how the effects of immi-gration on unemployment provisions are filtered by EUlabor market integration and domestic institutional ar-rangements. In countries less integrated into the EU la-bor market, increased immigration rates lead to more

American Journal of Political Science, Vol. 55, No. 3, July 2011, Pp. 647–664

C©2011, Midwest Political Science Association DOI: 10.1111/j.1540-5907.2011.00509.x

647

648 CHRISTINE S. LIPSMEYER AND LING ZHU

unemployment provisions, while the same relationshipin more integrated countries is less strong. The influenceof immigration comes largely through the domestic in-stitutions: higher immigration rates are associated withlarger benefit levels when left parties hold a larger pro-portion of parliamentary seats and when labor unions arestronger.

Immigrants, Open Labor Markets,and Welfare States

Explaining the relationship between globalization, EUeconomic integration, and welfare states in advanced in-dustrialized countries has been a central aim for bothpolitical economists and policy makers. For countries inEurope, the economic integration process mirrors theglobal economic trend in that it increases capital mo-bility, trade liberalization, and labor mobility (Rhodes1995). One direct effect of economic integration is thatthe establishment of an economic and monetary unionhas led to rapid growth of capital mobility and the expan-sion of the financial market (Cameron 2001). Anothermacroeconomic impact of EU integration is trade liber-alization, removing trade barriers and facilitating the freemovement of commodities across states (Hall 2001). Theextensive literature on globalization has focused on thesetwo aspects rather than on the mobility of labor. Someargue that integration and globalization are destabilizingwelfare states by affecting wage formation through labormarket competition and increased job mobility (Palierand Sykes 2001), while others view this as a way to en-force labor market flexibility (Haldrup, Anderson, andSorensen 2000). Still, others say it signals an erosion ofsocial standards and welfare support (Halfmann 2000;Lundborg and Segerstrom 2002).

How integration affects countries’ welfare policiesremains a debatable issue. There have been fears thatas the EU develops its labor market integration agendaeastwards, a more integrated labor market could reducethe ability of states to implement or maintain generoussocial insurance and redistributive policies (Huber andStephens 2001; Lundborg and Segerstrom 2002; Peridy2007; Razin and Sadka 2000). Scholars who argue forthis idea of welfare reduction believe that labor immigra-tion can lead to a negative self-selection process; betterjob opportunities and more generous social welfare moti-vate workers to migrate (Borjas, Freeman, and Katz 1996;Brooks 2002, 2007; Favell and Hansen 2002; Geddes 2003;Kahler 1992; Walz 1997). Opening the labor market toless developed EU states allows cheap labor to flow toward

states with more generous welfare provisions.1 In order tomaintain economic competitiveness, domestic employersmay pressure governments to reduce welfare compensa-tion. Furthermore, countries with generous welfare poli-cies will face pressures when they compete with countriesthat do not have high levels of social provisions in terms ofattracting capital investment (Clarke 2004; Rhodes 2001).

Alternatively, the effect of labor integration may de-pend on a country’s economic and political environ-ments. First, whether immigrants are assets or burdensto a country may depend on the demand in a country’slabor market. If a country desires foreign workers andthey do not directly threaten job opportunities for nativeworkers, then labor immigration could be mutually ben-eficial to both immigrants and natives and would not addto the fiscal burden of maintaining generous social welfare(Zimmermann 1995). Second, the extensive comparativepolitical economy literature points to a long tradition ofegalitarian policymaking in highly open and internation-ally competitive economies (Scheve and Slaughter 2006).The core argument is that economic globalization maydisplace native workers, causing an increase in politicaldemands on social insurance and redistributive policies(Cusack 1999; Iversen and Cusack 2000; Pierson 1994).

The existing literature on immigration, economic in-tegration, and welfare policies reaches no conclusions onhow these global pressures affect domestic policy. Manyscholars conceptualize immigration and integration asglobal economic shocks, but provide opposing predic-tions in terms of how globalizing economic forces leadto welfare policy changes.2 In addition, the relationshipbetween immigration, integration, and welfare changeis viewed as unidirectional. We contend that this think-ing might oversimplify the complex causal mechanismin terms of how global economic shocks are translated

1 This “push and pull” literature on international immigrationasserts that immigrants react to economic incentives created byincome gaps (Borjas 1995). Immigrants move from low-incomecountries to high-income countries with social welfare policies act-ing as a possible “pulling” mechanism (Marques 2010). Empiricalevidence for this classic “pull” argument has been somewhat lack-ing, with Castles and Miller (2003) and Feld (2005) finding thatimmigrants respond more directly to labor market demands, ratherthan welfare provisions. In fact, they do not necessarily flow fromthe poor areas into the rich areas.

2 We find predictions for both welfare reduction and expansionin the empirical literature. Scholars of neoliberal reform of socialwelfare policies—e.g., cutting benefits, tightening eligibility, etc.—have focused their concerns on whether economic globalization hasplayed a significant role in fostering these reforms (Bommes andGeddes 2000; Favell and Hansen 2002; Kivst 2004; Rhodes 1995).Scholars who focus on domestic institutions find evidence, instead,of welfare expansion (Basinger and Hallerberg 2004; Cusack 1999;Ha 2008; Iversen and Cusack 2000; Pierson 1994; Swank 2005).

IMMIGRATION, GLOBALIZATION, AND UNEMPLOYMENT BENEFITS 649

into policy changes. Instead, we argue that how EU wel-fare states incorporate immigrants into their systems maybe determined by the institutional context of the recipi-ent states. By omitting the link between global economicforces and a variety of institutional settings, one cannotgain a comprehensive understanding of the relationshipbetween immigration and welfare states.

An Interactive Modelof Welfare Change

In our theoretical approach, we focus on the interactiverelationship between immigration, integration, and do-mestic political institutions. Our argument focuses onthe effect of immigration and how it will depend on thecontext of the country—how integrated it is in the EUand the state’s domestic political situation. We transcendprevious research on economic globalization by theoriz-ing about the labor aspect of the globalizing process.3 Weargue that labor immigration is an important aspect ofglobal economic forces. Without considering the impactof immigration, we cannot have a comprehensive under-standing of the relationship between globalization anddomestic welfare policies.

We contend that immigration is a force that con-strains welfare changes. More specifically, we rely on aneoliberal theoretical understanding of economic forcesto inform our argument concerning international laborpressures. In a competitive labor market, workers’ in-comes are determined by the supply-demand relationshipof the labor market. Unemployment provisions, as a ma-jor source of wage compensation, are also determined bymarket mechanisms. In a pure market-driven scenario,an increase in immigrants will inflate the labor supply indomestic markets and thus generate pressures for reduc-ing wage rates and wage compensations (Borjas, Freeman,and Katz 1996). Under this theory, the relationship wouldbe: Immigration → Welfare.

When we consider institutional factors that can affectlabor supply and demand in domestic markets, however,the relationship between immigration and unemploy-

3 Previous scholars have focused primarily on the capital side ofglobalization. For example, some scholars used indicators for tradevolume (Hicks and Swank 1992; Huber, Ragin, and Stephens 1993;Rodrik 1998), while others focused on capital flows. Garrett andMitchell (1997) use import penetration from low-wage countriesand financial market integration to construct their measures forglobalization. In more recent studies, scholars focused on interna-tional trade exposure (Castles 2001), policy restrictions on capitalmobility (Ha 2008; Quinn 1997), and tax rate interdependency(Sanz and Velazquez 2001).

ment provisions becomes less obvious. Various institu-tional arrangements may influence this supply-demandrelationship, as well as market competition. Therefore, inour theoretical framework, we consider three institutionsthat can affect labor market competition and thus shapethe relationship between labor immigration and welfareentitlements. Thus, under our theory, the relationshipwould be: Immigration × Institutions → Welfare.

First, we argue that the influence of immigration de-pends on how integrated a country is in the EU’s labormarket structure. When a country moves toward moreEU integration, there are fewer domestic restrictions onimmigrants and less protection for the residential work-ers. Relaxing domestic protections could bring challengesfor a country to maintain a high level of redistribution,because domestic workers in such an open economy havefewer comparative advantages when faced with immi-grant labor than those workers in a more closed eco-nomic situation (Bhagwati 1982; Kemnitz 2005). Withless of a comparative advantage, workers faced with thiscompetition will find their wages reduced, and coun-tries will see their taxes from net income decrease. Coun-tries with more open labor markets may not be able tomaintain high redistribution levels. Therefore, the im-pact of immigration on welfare policy may be partic-ularly strong when there are very few protections fordomestic workers (Baur and Zimmermann 1997; Boeriand Terrell 2002; Ireland 2004; Kivst 2004). When acountry moves away from EU integration, it can im-plement more restrictive labor market policies that canbuffer the shocks brought by immigration, culminating infewer competitive pressures on welfare reduction. There-fore, we contend that net immigration exhibits a morenegative impact on welfare policies in countries thatare moving toward integration than countries movingaway from integration (Brucker, Frick, and Wagner 2006;Rhodes 1995).

EU labor market integration, however, is not the onlyinstitutional factor that can affect how immigration al-ters a perfect labor market scenario. We argue that do-mestic political pressures and institutions also interplaywith the market mechanism and affect the resilience ofwelfare policies. Overall, depending on the political influ-ence of a welfare policy’s supporters, a policy may find areprieve from the threat of retrenchment (Pierson 1994).For example, when groups of misplaced native work-ers are large, democratic governments may respond totheir demands by maintaining high levels of unemploy-ment compensation. Prior studies demonstrate that twospecific institutions often play essential roles in affectingwage-bargaining and protecting domestic workers: left-wing political parties and labor unions.


Traditionally, workers are the core constituents ofleft-wing political parties. When increased labor mar-ket competition leads to more market risks for do-mestic workers, left-wing parties may respond to theirconstituents’ needs by providing more generous welfareprovisions. Research finds that countries governed by leftand social democratic parties tend to have higher levelsof welfare provisions (Huber and Stephens 2001; Swank2000). When facing pressures for welfare policy changes,governments must resolve the trade-offs between mak-ing the domestic market more attractive to capital in-vestment and sacrificing domestic workers’ interests bycutting welfare benefits (Basinger and Hallerberg 2004).Reducing unemployment entitlements may be costly ifit hurts the left parties’ constituents’ interests; therefore,left-wing parties may have few political incentives to re-duce entitlement levels.

Labor unions can have a significant influence duringthe wage-setting process. The collective labor-capital bar-gaining systems may play an intervening role by bufferingpolicies from economic shocks brought by liberal mar-ket forces (Basinger and Hallerberg 2004; Streeck 1993;Swank 2002). In countries where union density is high,labor markets become less competitive for domestic cit-izens. The existence of organized labor unions can helpto sustain higher levels of wage compensation and so-cial protection for domestic workers, because bargainingmechanisms block the economic forces that might re-duce the level of social protections in a more competitivemarket.

Therefore, the challenges brought by economic open-ness combined with a strong left-labor power context (i.e.,left parties and labor unions) can produce a compensa-tion strategy that retains a generous welfare state (Cusack1999; Hall and Soskice 2004; Iversen and Cusack 2000).According to Iversen and Cusack (2000), economic open-ness and the integration process can become forces forwelfare expansion instead of retrenchment.4 This high-lights how the relationship between economic pressures(i.e., inflow of immigrant workers) and welfare changesmay be contingent upon the configuration of the politicalsystem. Depending on a country’s mix of political institu-tions, a government may choose to compensate workersand citizens for market pressures or failures.

To this end, our theoretical argument is built on theinterplay between economic shocks (brought by immi-

4 Rudra and Haggard (2005) extend the debate by providing an em-pirical analysis of developing countries and show that when facingeconomic pressures brought by globalization, welfare states wouldhave different reactions. They find that authoritarian regimes tendto react more by cutting their welfare benefits than democratizedstates.

gration) and two types of institutional arrangements:EU labor market integration and domestic institutions.When combined with increased immigration, EU in-tegration should expand the competition in the labormarket, ultimately causing a decline in unemploymentprovisions. Immigration should not have a retrenchingeffect, however, in countries with a “compensating” do-mestic institution—left-wing parliaments or stronger la-bor unions. We argue that in order to buffer the domesticlabor market and shore up their support base, these polit-ical actors will either stabilize or increase welfare benefits,rather than retrench their assistance.

Research Design

According to a neoliberal theoretical approach, the eco-nomic shocks brought by immigrants will produce pres-sures for governments to reduce their welfare provisions.But the relationship between immigrant workers and do-mestic welfare changes may be more complex than this.Our more comprehensive theoretical model alters thesegeneral expectations, and we argue that these market-based expectations are distorted by EU labor market in-tegration forces and domestic political institutions.

Hypotheses

Therefore, our theoretical expectations are that (1) inte-grative forces at the EU level will condition the effects ofincreased immigration as an economic shock on unem-ployment provisions; (2) domestic political institutionswill mediate the effects of economic pressures, (2a) whenimmigration increases in states with parliaments havinga larger percentage of left political party seats, we wouldnot expect to observe benefit retrenchment, (2b) whenincreasing immigration into states with larger unionizedworkforces, we also would not expect to find decreas-ing unemployment benefits.5 From these expectations,we derive the following set of hypotheses:

H1: Increased immigration into states moving towardEU labor market integration will have a negativeeffect on unemployment provisions.

5 The literature on both of these domestic institutions offers ad-ditive hypotheses that expect them to have positive effects on un-employment provisions. Although we agree with these theoreticalexpectations and they inform our own expectations, in this article,we are concerned with their interactive effects when combined withimmigration.


H2a: Increased immigration into states with a largerpercentage of left party seats in parliament willhave a nonnegative effect on unemployment pro-visions.

H2b: Increased immigration into states with a largerpercentage of unionized workers will have a non-negative effect on unemployment provisions.

Data, Measures, and Method

To test our hypotheses, we pool data on labor immigra-tion, EU labor market integration, domestic institutions,and macroeconomic conditions for 15 European statesfrom 1971 to 2007. This period includes the major wavesof EU enlargement in 1973, 1981, 1986, and 1995.6

Unemployment Entitlements. For our dependentvariable, we use unemployment benefits measured as re-placement rates. This OECD summary measure of aver-age unemployment entitlement is defined as the averageof the gross unemployment benefit replacement rate. TheOECD calculates this measure based on cash replacementrates for two earnings levels, three family situations, andthree durations of unemployment (Martin 1996; OECD1994, 2007).7

Immigration. To measure labor immigration, we usedata from the OECD’s International Migration Statisticswebsite. In our empirical models, we measure immigra-tion by the net migration rate for each country in a givenyear.8 We compute it using the following equation:

6 We include 14 EU member states (Austria, Belgium, Denmark,Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Por-tugal, Spain, Sweden, United Kingdom) and Norway in our study.Ideally, we should include all 27 EU member states and analyzeall (six) waves of EU enlargement. Due to data availability, we in-clude 15 countries in this article. Although Norway is not an EUmember state, it signed the Agreement on the European EconomicArea (EEA) and participates in the Schengen/Dublin Agreement.Therefore, we include Norway as one of our country cases.

7 Data are accessed through the following webpage: http://www.oecd.org/dataoecd/52/9/42625593.xls. Biannual data from 1971 to2007 are available for the entitlement measure. To obtain annualdata for our panel analysis, we interpolate the data by using theaverage values for the year before and after the missing data point(e.g., Blanchard 1998; Blanchard and Wolfers 2000).

8 A more direct measure of labor immigration would be an indicatorconstructed by counting the stock of foreign workers flowing fromless developed countries to EU member states. Unfortunately, datafor employed foreign workers by country of origin are only availablefrom 1990 to 2007 in the OECD migration database. A comparisonof the two measures for the period when both are available yields apairwise correlation of .9774.

Net Migration Rate

= Inflow of Immigrants-Outflow of Immigrants

1,000 Inhabitants (1)

Globalization. We include three indicators of global-ization to reflect the international flow of labor, capital,and commodities: EU labor market integration, foreigndirect investment (FDI), and foreign trade.

EU Labor Market Integration. Labor market integra-tion reflects the international flow of labor. The existingempirical literature provides various ways to measure in-tegration of the capital market in the EU; yet it provideslittle information on how to measure labor market inte-gration. Scholars who study the labor market and integra-tion in the EU generally rely on enlargement treaties anddomestic labor market restrictions as indicators of labormarket openness. Commonly, they quantify integrationtreaties or domestic policies by creating a set of dummyvariables that distinguish which countries participatedin a particular EU treaty or adopted a particular labormarket restriction policy (Ha 2008; Hansson and Olofs-dotter 2008). While both EU treaties and domestic labormarket restrictions reflect important policy dimensionsof labor market integration, these variables are relativelystatic measures that reflect little variation in the level oflabor market integration across countries and time. Toconstruct our integration measure of EU labor markets,we rely on the rule of “one price” for an integrated mar-ket. In the presence of a competitive market structure andin the absence of transport costs and other barriers totrade, prices of homogenous products sold in differentmarkets would converge to “one price” because of marketforces (Baele et al. 2004; Funke and Koske 2008; Gold-berg and Verboven 2005; Rosenbloom 1990; Sarno andTaylor 2002). The rule of one price implies that if assetsin spatially separated markets have sufficiently compara-ble characteristics and are evaluated by the same set ofmarket rules, “cross-market prices or yield differentialsover time constitute a measure . . . of time-varying inte-gration” (Baele et al. 2004, 510). More specifically, in afully integrated labor market, the price of labor in dif-ferent countries would converge toward the same marketequilibrium over the long run. If a particular labor marketsegment (i.e., the labor market in one country) is less inte-grated into the common market, then the price for laborin that market segment would be more dispersed fromthe market equilibrium. Based on this concept of priceconvergence, we construct a labor-price-based measurefor EU labor market integration using the OECD annual


data for the cost of an average unit of labor in each coun-try (OECD 2009). The integration measure is calculatedbased on the following equation:

EU Integrationit = 1 − Xit

Mt(2)

where:

Xit = |EU Average Unit Labor Costt −Unit Labor Costi t |(3)

Mt = maxi

(Xit) (4)

In equation (2), the numerator is the EU average unitlabor cost at time t minus the unit labor cost in country i attime t (equation 3), and the denominator is the maximumdiffernce between these two costs for every country i andyear t (equation 4).9

Figure 1 is a plot of the integration measure for eachcountry. The measure ranges from 0 to 1 with a “1” rep-resenting the perfect integration scenario, whereby thedomestic labor price is the same as the EU average laborprice.10 Overall, the ratio measure shows variation acrosscountries and years with more dispersion from 1 in theearlier years. The Nice Treaty witnessed a movement inthe ratio measure toward 1, with countries’ labor marketsbecoming more integrated into the EU market. Compar-ing across countries, the founding states’ domestic labormarkets are more integrated into the EU than memberstates that entered after the SEA.

FDI and Foreign Trade. We use foreign direct invest-ment as an indicator of capital mobility and measure it as ashare of GDP. We measure foreign trade as total importsand exports as a share of GDP to reflect internationalflows of goods and commodities. Both of these measuresof economic openness are drawn from the Penn WorldTables (Heston, Summers, and Aten 2009).11

Domestic Political Institutions. As discussed above,we include indicators for two domestic political institu-tions: (1) left-wing parliamentary seats and (2) the ef-fective power of labor unions. We rely on a measure of

9 The unit labor cost indexes (exchange rate adjusted) are calculatedas the quotient of total labor costs and real output in a total economy(including manufacturing, industry, construction, etc.).

10 The top panel in Figure 1 presents the five founding members ofthe EU (Luxembourg is not included in our sample). The middlepanel shows the six countries that joined the EU after the SingleEuropean Act (SEA), and the bottom panel displays the countriesthat became EU members after the Maastricht Treaty.

11 PWT (version 6.3) accessed: http://pwt.econ.upenn.edu/phpsite/pwt index.php.

left parliamentary seats from Swank (2002) that is cal-culated based on left party legislative seats as a share ofall legislative seats.12 Following the work of Wallerstein,Golden, and Lange (1997), our indicator of union densityis net union membership as a share of wage and salaryearners in employment. This indicator reflects the stateof the domestic wage bargaining system and industrialrelations.13

Economic Controls. To isolate the relationship be-tween immigration, integration, and unemployment en-titlements, we include two economic control variablesthat are commonly used in empirical models investigatingthe political economy of welfare policy. First, we includeper capita GDP as a control for the domestic economy,which we expect to be positively associated with unem-ployment entitlements.14 Second, we include the unem-ployment rate as a control for the domestic labor market,which we expect to be positively associated with unem-ployment provisions.15

Method and Model Specification

We construct the dataset for our empirical analyses bypooling data for 15 EU states from 1971 to 2007. Acommon practice for analyzing panel data for compar-ative political economy models is to use a lagged de-pendent variable with country fixed effects and panel-corrected standard errors based on Beck and Katz’s (1995)

12 Swank’s “Comparative Parties Datasets: Political Strength of Po-litical Parties by Ideological Group in Capitalist Democracies”(http://www.marquette.edu/polisci/faculty swank.shtml).

13 Data for the union density measure are drawn from two sources:(1) Golden, Lange, and Wallerstein’s (2009) dataset on industri-alized democracies, and (2) ICTWSS, a database of the institu-tional characteristics of trade unions, wage setting, state inter-vention, and social pacts, maintained by the Amsterdam Insti-tute for Advanced Labor Studies (AIAS). Both datasets use thesame equation to calculate the union density measures. Data inthe Golden, Lange, and Wallerstein dataset do not cover yearsafter 2000, and the ICTWSS database includes data from 1960to 2007. Since the correlation between these union density mea-sures is .9847, we maximize our year observations by merging thetwo datasets. Golden, Lange, and Wallerstein’s measure of uniondensity is from http://dvn.iq.harvard.edu/dvn/dv/golden, “UnionCentralization among Advanced Industrial Societies: An EmpiricalStudy.”

14 From the PWT (version 6.3), we measure GDP using cgdp, “realgross domestic product per capita, current price.” As a robustnesscheck, we ran the models using other per capita measures from thedataset (i.e., rgdpl, rgdpl2, and rgdpch) and found substantivelysimilar results.

15 Data are from the International Labor Organization (ILO) YearlyLabor Statistics.


FIGURE 1 Changes in EU Labor Market Integration:1971–2007 (Data Source: OECD Statistics, theUnit Labor Costs–Annual Indicators)

recommendation. But if a nonstationary dependent vari-able is used and if random effects are present acrosscountry-year cases, then this model will produce biasedresults (Baltagi 2008). When testing our dependent vari-

able for a panel unit root, we find our variable stationaryusing a Phillips-Perron test (� 2 = 51.39, p = .01).

The second methodological consideration concernsthe error structure in our panel data. Beck and Katz


(1995) argue that if data have panel-level heteroskedas-ticity and only spatial autocorrelation is present, using alagged dependent variable with panel-corrected standarderrors is more efficient. In addition, because our interestis in a public policy, we assume a path-dependent pro-cess whereby the lagged dependent variable controls forthe level of entitlements in the previous year. However, wealso need to consider the problem of autocorrelation whenusing a lagged dependent variable (Baltagi 2008), espe-cially since we find both panel-wise heteroskedasticity andfirst-order autocorrelation.16 In addition to the lagged de-pendent variable in our model, we implement an AR(1)correction for the autocorrelated error structure. Basedon our theory and statistical diagnostics, we use ordinaryleast squares (OLS) with a lagged dependent variable andcountry fixed effects, as well as panel-corrected standarderrors (PCSE) and an AR(1) correction.

In order to test our hypotheses, we specify an interac-tive model to decipher how immigration, along with inte-gration and domestic institutions, shapes unemploymentbenefits. The model includes interaction terms betweenimmigration and integration, as well as immigration andeach of our domestic institutional variables—left partyseats and union density. Because we model immigrationand integration as economic shocks, we treat them in thesame way as the other economic and international vari-ables and use their differences in our model. We contendthat the changes in immigration and integration, not theirabsolute values, influence unemployment entitlements.17

Following in the footsteps of welfare spending models,we lag our domestic political institutions by one year. Al-though policy benefits and policy spending levels denotedifferent parts of the policy process, political actions cantake time to affect policy outcomes in the same way asspending levels. Taking these theoretical arguments intoconsideration, we specify the following model, noting thati and t index countries and years studied:

Entitlement it = �i + �Entitlementit−1

+ �I � Immigrationit +�N� Integrationit

+ �L Left Partiesi t−1 + �U Unionsit−1

+ �IN� Immigrationit × �Integrationit

+ �IL� Immigrationit × Left Partiesi t−1

+ �IU� Immigration × Unionsit−1

+ �E � E conomic Contr ol s + �it(5)

16 Using the White Test, we detect group-wise heteroskedasticity,and the Arellano and Bond Test (Arellano and Bond 1991) showsfirst-order autocorrelation.

17 Our index of integration is not stationary, but the first differencesare. The chi-square statistic based on the Fisher Test is 32.9182,p = 0.3261 (Maddala and Wu 1999).

TABLE 1 Effects of Immigration, Integration,and Domestic Institutions on the Levelof Unemployment Entitlement inDeveloped EU Welfare States,1971–2007

Variable Coefficient (PCSEs)

�Immigration −0.250 (0.169)�Integration 1.921 (1.984)Union Densityt−1 −2.612 (1.653)Left Seatst−1 0.010 (0.012)�Immigration × �Integration −0.417 (0.941)�Immigration × Union Densityt−1 0.385 (0.260)�Immigration × Left Seatst−1 0.004∗ (0.002)�GDP −0.0001 (0.0002)�Unemployment −0.125 (0.083)�Trade 0.004 (0.014)�FDI −0.071∗ (0.043)Entitlementt−1 0.890∗∗ (0.019)Intercept 3.988∗∗ (1.100)N 496� 0.414R2 0.961

∗p < .10, two-tailed t-test.∗∗p < .05, two-tailed t-test.Notes: Dependent variable is the level of unemployment entitle-ments. Coefficients for country dummies not reported.

Results of Pooled Time-SeriesRegression Analysis

Our argument focuses on how the movement of immi-grant labor shapes unemployment benefits;18 however, itdoes not hinge on the independent effects of immigra-tion, integration, or domestic institutions affecting enti-tlements. We contend that immigration plays an interven-ing role in influencing unemployment entitlements, withthe domestic or integrative context of a country affect-ing how immigration shapes unemployment provisions.Table 1 shows the results from our model.19

Because we interact our variables of interest—immigration, integration, and domestic politicalinstitutions—with each other, it is best to use statistical

18 To test for a possible reversed causality issue in our analysis, we re-gressed lagged unemployment benefits on immigration but foundno statistically significant relationship (details in the SupplementalInformation).

19 Note that we performed robustness checks on our results in theform of jackknifing them for each country and year and found nosubstantive differences between the results.


simulations to gauge both the substantive and statisticalsignificance of marginal changes in our variables (Bram-bor, Clark, and Golder 2006; Kam and Franzese 2007).Our figures rely on the creation of scenarios where we setone variable of interest to two values (i.e., the values at the10th and 90th percentiles) while allowing another vari-able of interest to vary across its observed range of values.For clarity, we create one graph to correspond with eachof the two values, and for comparison purposes, showthem side-by-side.20 Note that we hold all other variablesin the model constant at their mean levels. For the simu-lated values of the two interacted variables, we calculatethe 95% confidence intervals for unemployment entitle-ment and show those along with the predicted levels ofunemployment entitlement.21 Note that for each pair ofgraphs we are interested in two relationships: (1) if thereare significant differences in the dependent variable’s levelbetween the two sets of values—so between the graphs,and (2) if there are significant differences in the dependentvariable’s level across the values of the second variable—so within each graph. From these, we can make inferencesabout the impact of each interactive relationship on thelevel of unemployment entitlement.

Figure 2 presents the predicted level of unemploy-ment entitlement across the range of observed values ofchange in immigration for low and high levels of thechange in integration (the 10th and 90th percentiles inour data). First, when comparing across the right andleft graphs, we can see that the two integration scenariosoverlap, indicating that the predicted level of entitlementis not statistically different when comparing low and highlevels of change in integration across the same changesin the immigration rate. This result runs contrary to Hy-pothesis 1. Second, to assess the impact of changes inimmigration, we need to compare across the confidenceintervals for each figure. In the figure on the left, for thelow level of change in integration, the confidence intervalon the left side of the figure does not overlap that on theright side, illustrating that the influence of immigrationon unemployment benefits depends on the level of changein integration. In instances where change in integrationis low, the level of entitlements can range from 23% to35% depending on the change in immigration. But thissame relationship does not hold for a high level of changein integration (in the right figure), where the confidenceintervals barely overlap at the lower and higher levels ofimmigration change. Therefore, we can see that changes

20 One figure that includes the two overlapping figures is availabilefor all scenarios in the Supplemental Information.

21 We use STATA 11 for all statistical work and the Clarify programfor the figures (Tomz, Wittenberg, and King 2003).

in immigration do have a statistically significant effect onunemployment entitlement when change in integrationis at a lower level.

By allowing change in integration to vary across itsobserved values and setting the change in immigrationat low and high levels, in Figure 3 we see another ver-sion of this relationship that reinforces the findings fromFigure 2. First, by analyzing the two graphs in Figure 3together, we find overlap of the confidence intervals andtherefore, no statistically significant differences betweenthe low and high levels of change in immigration on theimpact of change in integration on unemployment bene-fits. Regardless of the level of the change in immigration,changes in integration have no influence on the level ofunemployment entitlements. Second, when analyzing theimmigration scenario in each graph, the confidence in-tervals overlap from less to more change in integration inboth cases, indicating that there is not a statistically signif-icant relationship between change in integration and un-employment benefits regardless of immigration change.Therefore, we find no support for Hypothesis 1.

When changes in immigration interact with differ-ent domestic political contexts, does this lead to vary-ing levels of unemployment entitlement? In other words,can the domestic arrangments influence whether or notimmigration has an effect on benefits? Beginning withgovernment institutions, Figure 4 shows the results fromscenarios illustrating the differences in benefit levels be-tween low and high levels of left parliamentary seats (10thand 90th percentiles—28% and 56%) when allowing thechange in immigration rates to vary across its observedrange of values. We can see how these political scenar-ios lead to different benefit outcomes. First, when com-paring the right and left graphs, there is no discerniblestatistical difference between the two scenarios, becausethe confidence intervals for the two left party scenariosoverlap as change in immigration moves from lower tohigher levels. Secondly, when looking within each fig-ure, there are statistically significant differences in thelevel of benefits as the change in immigration varies butonly when left parties control a majority of parliamentaryseats (the right figure). The confidence intervals for thescenario in the right figure do not overlap when com-paring them at the lower and higher levels of immigra-tion change. This same relationship does not hold forthe figure on the left, when left party seats are at a lowlevel. Therefore, larger changes in immigration combinedwith strong left party support in parliament lead to moregenerous unemployment provisions, offering support forHypothesis 2a. But we find no corresponding relation-ship when left parties control only 28% of parliamentaryseats.


FIGURE 2 Levels of Change in Integration Across the Range of Change in Immigration• All other variables are held at their mean values.

FIGURE 3 Levels of Change in Immigration Across the Range of Change in Integration• All other variables are held at their mean values.


FIGURE 4 Levels of Left Party Seats Across the Range of Change in Immigration• All other variables are held at their mean values.

FIGURE 5 Levels of Change in Immigration Across the Range of Left Party Seats• All other variables are held at their mean values.


FIGURE 6 Levels of Union Density Across the Range of Change in Immigration• All other variables are held at their mean values.

To flesh out this relationship further, in Figure 5 wecan see a more nuanced picture of how change in immi-gration and government ideology interact to alter unem-ployment entitlements. The two graphs show scenariosfor low and high levels of immigration when left partyseats vary, and a comparison of the two illustrates that amajority of left party seats is not necessary for statisticallysignificant changes in benefit levels. When immigrationincreases substantially from the previous year and theleft holds at least 40% of the parliamentary seats, thenunemployment entitlements are significantly more thanwhen the change in immigration is low. These findingsprovide further support for Hypothesis 2a. When ana-lyzing each figure separately, we find that the confidenceintervals overlap for the each scenario in both figures,indicating no statistical significant relationship betweenleft party seats and unemployment provisions regardlessof the change in immigration. Since we have a directionalhypothesis, we take another look at this relationship using90% confidence intervals and find that for a high level ofimmigration, there is a significant statistical relationshipbetween left party seats and unemployment provisions,while this relationship does not hold for a low level ofimmigration.22

22 This figure is available in the Supplemental Information.

Next, we turn to how changes in the immigrationrate and the power of labor unions interact to affect un-employment entitlements. The two graphs in Figure 6illustrate the relationship between low and high levels ofunion density across change in immigration. First, a com-parison of the two graphs shows no significant differencesbetween the levels of entitlement for low and high levelsof union density; the confidence intervals overlap acrossthe range of observed values for immigration change. Sec-ond, when comparing within each union density figure,we find overlap of the confidence intervals within bothscenarios, indicating that there is no statistically signif-icant relationship between changes in immigration andunemployment provisions regardless of the level of uniondensity. When analyzing the relationship using 90% con-fidence intervals for our directional hypothesis, we find astatistically significant relationship, indicating that whenunion density is high, immigration has a positive effecton unemployment provisions.23 The results from this sce-nario offer support for Hypothesis 2b, which posited thatincreased immigration into states with substantial laborunion power would result in higher unemployment enti-tlements. In Figure 7, with two graphs for low and highlevels of immigration change across the range of union

23 This figure is available in the Supplemental Information.


FIGURE 7 Levels of Change in Immigration Across the Range of Union Density• All other variables are held at their mean values.

density, we find no statistically significant differences inunemployment entitlements across the two immigrationscenarios. When looking within each figure, we find over-lap of the confidence intervals for low to high union den-sity for each one, indicating that the different rates ofimmigration change do not significantly vary across thepercentage of unionization.

Thus far we have concentrated on the estimated im-pact of changes in immigration with integrative and do-mestic institutions on unemployment entitlement in asingle year. One of the attractive aspects of models with alagged dependent variable is that they allow us to estimatethese short-run effects together with long-run effects (de-Boef and Keele 2008). In these models, the short-run ef-fects of an independent variable X are captured by theparameter estimate for that variable (�̂), while the long-

run effects are captured by �̂

1−�̂where �̂ is the parameter

estimate for the lagged dependent variable.In our model, the variables in which we are most

interested (integration, left party seats, and union den-sity) are each interacted with change in immigration.This specification means that a change in immigrationwill work through all three of these variables to influenceunemployment entitlement in both the short run and thelong run. In order to obtain estimates of the short-runand long-run impacts of change in immigration contin-

gent on these conditions, we conducted a series of simu-lations, which are presented in Table 2.24 In each of thesescenarios, our institutional measures are held constant ateither their 10% or 90% level and immigration change ismoved from 0 to 2.76 (one standard deviation) to sim-ulate a positive shock of immigrant workers. At the topof Table 2, we can see that under these circumstancesa nation with a strong left and strong unions that wasmoving away from integration would react by increas-ing unemployment entitlements by almost .78 in the firstyear. The long-term effect of this labor shock would bean increase in entitlements of 7.09. Both of these effectsare statistically significant at conventionally accepted lev-els (p < .05). As we move further down Table 2, we cansee that both the short-run and long-run impacts of thisimmigration shock, though slightly reduced in magni-tude, are positive and significant in a nation with a strongleft and strong unions that is moving toward integration.In the case where there is a weak left, strong unions, andmovement away from integration, the impact of a positiveimmigration shock is increases in entitlements over boththe short run (.46) and long run (4.18) that are borderlinein terms of statistical significance (p < .10). In none of

24 We conducted these simulations using the “changex” commandin Clarify (Tomz, Wittenberg, and King 2003).


TABLE 2 Short- and Long-Run Effects of Institutional Scenarios

Short-Run Effects Long-Run Effects

Strong Left Party, Strong Union Density, Negatively Integrating 0.78∗∗ 7.09∗∗

Strong Left Party, Strong Union Density, Positively Integrating 0.71∗∗ 6.45∗∗

Weak Left Party, Strong Union Density, Negatively Integrating 0.46∗ 4.18∗

Weak Left Party, Strong Union Density, Positively Integrating 0.39 3.55Strong Left Party, Weak Union Density, Negatively Integrating 0.16 1.45Strong Left Party, Weak Union Density, Positively Integrating 0.08 0.73Weak Left Party, Weak Union Density, Negatively Integrating −0.16 −1.45Weak Left Party, Weak Union Density, Positively Integrating −0.24 −2.18

∗p < .10, two-tailed test.∗∗p < .05, two-tailed test.

the remaining scenarios is the simulated impact of thispositive immigration shock statistically significant. Witha strong left but weak unions, the point estimate for ashort-run effect is very close to zero and the point esti-mates for the long-run effects are quite small. When boththe left and unions are weak, we get negative point esti-mates for both the short-run and long-run effects of thisimmigration shock. This series of simulations illustratesthat the short-run impact of a change in immigrationvaries substantially depending on the strength of left par-ties and union density in the nation at the time of thechange. Over time, these variations are exacerbated bytheir long-run effects. The relative impact of integrationis, as we saw in earlier simulations, fairly modest.

Discussion

How immigration affects unemployment protection isnot a straightforward relationship. From a neoliberalviewpoint, increasing immigrant labor should producepressures on a domestic labor market, resulting in fallingwages and decreasing unemployment compensation. Butif we consider pressures that may alter a pure market sce-nario, then depending on a country’s domestic politicalcontext, increased immigration may lead to higher levelsof welfare provisions. Our empirical findings in this articleprovide some support for our argument that, dependingon domestic political institutions, increased immigrationcan instead cause an increase in unemployment entitle-ments. The political mechanisms that support more gen-erous unemployment provisions hinge on the interactiverelationships among economic forces and various insti-tutional arrangements. Interestingly, however, our resultspoint to globalization playing less of a role than previousresearch would have us expect.

Turning to how the movement of immigrant labor in-teracts with different institutional mechanisms, our em-pirical results suggest that EU integrative forces demon-strate less of an impact on unemployment entitlementsthan domestic political forces. The effect of changes inimmigration on unemployment benefits is the same re-gardless of a nation’s integration into the common EUlabor market. Allowing for varying immigration levels,countries becoming more open to the market are neithermore likely nor decrease nor increase their unemploy-ment benefits.

We argue that one reason why the effect of immigra-tion depends on domestic politics is because left partiesand labor unions directly represent domestic workers’policy preferences. To make decisions on unemploymententitlements, governments need to balance the economicinterests of capital and labor (Basinger and Hallerberg2004). As long-term policy provisions for compensatingjob market risks, unemployment entitlements are deter-mined by both domestic workers’ demands and whetherthere are policymaking institutions that can translatethese demands into policies (Burgoon 2001; Iversen andCusack 2000). As the long-term effects from our modelindicate, with rising immigration rates, substantial leftparty strength or labor union strength will have moresalient impacts in sustaining high levels of unemploy-ment entitlements. Corresponding with others’ conclu-sions (Huber and Stephens 2001; Swank 2002), we findthat when immigration rates increase and left partieshave a sizeable percentage in parliament, these partiesare able to translate their constituents’ preferences intorising unemployment benefits. Similar results appear forthe strength of unions, where a high level of union den-sity combined with greater changes in immigration re-sults in significantly higher levels of unemployment ben-efits. We find in both cases that as the share of left partyseats in parliament and the extent of unionization reach


substantial levels, the generosity of unemployment enti-tlements is greater than under other domestic contexts.Therefore, we find evidence in line with the literaturein terms of the impact of domestic institutions in con-ditioning the effects of immigration on unemploymententitlements.

Under certain conditions, labor market openness andwelfare policy may be able to reinforce each other. Ourfindings suggest the importance of considering variousaspects of globalization in order to further disentanglethe connections between economic integration, open-ness, and welfare policy (Burgoon 2001). Domestic policychanges, meanwhile, could vary across time due to differ-ent immigration scenarios. The varying policy responsesto globalizing forces might generate very different com-pensation politics and benefit different labor groups inEU member states. Further disaggregating both welfarepolicy efforts and immigration patterns would contributeto the understanding of the complex dynamics of welfarecompensation.

Conclusion

In this article, we argue that immigration, integration,and domestic politics combine to determine unemploy-ment policies. Our theoretical focus on labor integrationextends the literature on globalization beyond its focuson capital and links it with the multiplying forces of im-migration. Modeling immigration as an economic shockto unemployment benefits, we show that focusing on thecapital side of globalization may limit our understandingof how countries adjust their welfare policies under globaleconomic pressures. By including the influences of bothcapital and labor in our model of welfare policy, we showthat open labor markets and free labor movement cancreate pressures for welfare states to maintain generouslevels of unemployment benefits.

The impact of immigrants on unemployment com-pensation is filtered through domestic political insti-tutions. By viewing these political pressures as com-pensating forces, we transcend previous institutionalarguments by emphasizing the complex and interactivenature of the relationships between immigration and in-stitutions in shaping welfare provisions. Our empiricalfindings demonstrate that the EU labor market integra-tion mechanism has less of an impact on welfare changesthan domestic political institutions. As global economicforces have generated a more open and interdependentworld, changes in labor demographics and relevant wel-fare policies will be an important arena to further the

investigation of the relationship between growth, devel-opment, and inequality. The story of immigration andunemployment compensation in the EU is less about theopening of borders and the movement of labor than it isabout the domestic political makeup; in a world of nar-rowing borders, domestic politics still matters.

References

Arellano, Manuel, and Stephen Bond. 1991. “Some Tests ofSpecification for Panel Data: Monte Carlo Evidence and anApplication to Employment Equations.” Review of EconomicStudies 58(2): 277–97.

Baele, Leiven, et al. 2004. “Measuring European Financial In-tegration.” Oxford Review of Economic Policy 20(4): 509–30.

Baltagi, Badi H. 2008. Econometric Analysis of Panel Data. 4thed. West Sussex, UK: John Wiley and Sons.

Banting, Keith G. 2000. “Looking in Three Dimensions: Migra-tion and the European Welfare State in Comparative Per-spective.” In Immigration and Welfare: Challenging the Bor-ders of the Welfare State, ed. Michael Bommes and AndrewGeddes. New York: Routledge, 13–34.

Basinger, Scott J., and Mark Hallerberg. 2004. “Remodelingthe Competition for Capital: How Domestic Politics Erasethe Race to the Bottom.” American Political Science Review98(2): 261–76.

Baur, Thomas, and Klaus F. Zimmermann. 1997. “Integratingthe East: The Labor Market Effects of Immigration.” In Eu-rope’s Economy Looks East: Implications for Germany and theEuropean Union, ed. Stanley W. Black. New York: CambridgeUniversity Press, 269–310.

Beck, Nathaniel, and Jonathan N. Katz. 1995. “What to Do (andNot to Do) with Time-Series Cross-Section Data.” AmericanPolitical Science Review 89(3): 634–47.

Bhagwati, Hagdish N. 1982. “Shifting Comparative Advan-tage, Protectionist Demands, and Policy Response.” In Im-port Competition and Response, ed. Hagdish N. Bhagwati.Chicago: University of Chicago Press, 153–96.

Blanchard, Olivier. 1998. The Economics of Post-CommunistTransition. New York: Oxford University Press.

Blanchard, Olivier, and Justin Wolfers. 2000. “The Role ofShocks and Institutions in the Rise of European Unemploy-ment: The Aggregate Evidence.” Economic Journal 110(462):1–33.

Boeri, Tito, and Katherine Terrell. 2002. “Institutional Deter-minants of Labor Reallocation in Transition.” Journal of Eco-nomic Perspectives 16(1): 51–76.

Bommes, Michael, and Andrew Geddes. 2000. “Immigrationand Welfare States.” In Immigration and Welfare: Challengingthe Borders of the Welfare State, ed. Michael Bommes andAndrew Geddes. New York: Routledge, 1–13.

Borjas, George J. 1995. “The Economic Benefits from Immigra-tion.” Journal of Economic Perspectives 9(2): 3–22.

Borjas, George J., Richard B. Freeman, and Lawrence F. Katz.1996. “Searching for the Effect of Immigration on the LaborMarket.” American Economic Review 86(2): 246–51.


Brambor, Thomas, William Clark, and Matt Golder. 2006. “Un-derstanding Interaction Models: Improving Empirical Anal-yses.” Political Analysis 14(1): 63–82.

Brooks, Sarah M. 2002. “Social Protection and Economic Inte-gration: The Politics of Pension Reform in an Era of CapitalMobility.” Comparative Political Studies 35(5): 491–532.

Brooks, Sarah M. 2007. “When Does Diffusion Matter? Ex-plaining the Spread of Structural Pension Reforms AcrossNations.” Journal of Politics 69(3): 701–15.

Brucker, Herbert, Joachim R. Frick, and Gert G. Wagner. 2006.“Economic Consequences of Immigration.” In Immigrationand the Transformation of Europe, ed. Craig A. Parsons andThimothy M. Smeeding. Cambridge: Cambridge UniversityPress, 111–46.

Burgoon, Brian. 2001. “Globalization and Welfare Compensa-tion: Distangling the Ties and Binds.” International Organi-zation 55(3): 509–51.

Cameron, David. 2001. “Unemployment, Job Creation, andEconomic and Monetary Union.” In Unemployment in theNew Europe, ed. Nancy Bermo. New York: Cambridge Uni-versity Press, 87–118.

Castles, Francis G. 2001. “Decentralization and the Post-WarPolitical Economy.” Journal of European Social Policy 11(3):195–211.

Castles, Stephen, and Mark J. Miller. 2003. The Age of Migration:International Population Movements in the Modern World.London: Palgrave Macmillan.

Clarke, John. 2004. “Dissolving the Public Realm? The Logicsand Limits of Neo-liberalism.” Journal of Social Policy 33(1):27–48.

Cusack, Thomas R. 1999. “Partisan Politics and Fiscal Policy.”Comparative Political Studies 32(4): 464–86.

deBoef, Suzanna, and Luke Keele. 2008. “Taking Time Seriously:Dynamic Regression.” American Journal of Political Science52(52): 184–200.

Favell, Adrian, and Randal L. Hansen. 2002. “Markets AgainstPolitics: Migration, EU Enlargement and the Idea of Eu-rope.” Journal of Ethnic and Migration Studies 28(4):581–601.

Feld, Serge. 2005. “Labor Force Trends and Immigration inEurope.” International Migration Review 39(3): 637–62.

Funke, Katja, and Isabell Koske. 2008. “Does the Law of OnePrice Hold within the EU? A Panel Analysis.” InternationalAdvances in Economic Research 14(1): 11–24.

Garrett, Geoffrey, and Deborah Mitchell. 2001. “Globalization,Government Spending and Taxation in the OECD.” Euro-pean Journal of Political Research 39(2): 145–77.

Geddes, Andrew. 2003. “Migration and the Welfare State inEurope.” The Political Quarterly 74(1): 150–62.

Goldberg, Pinelopi K., and Frank Verboven. 2005. “Market Inte-gration and Convergence to the Law of One Price: Evidencefrom the European Car Market.” Journal of InternationalEconomics 65(1): 49–73.

Golden, Miriam, Peter Lange, and Michael Wallerstein.2009. “Union Centralization among Advanced Indus-trial Societies: An Empirical Study.” http://hdl.handle.net/1902.1/10193.

Ha, Eunyoung. 2008. “Globalization, Veto Players, and WelfareSpending.” Comparative Political Studies 41(6): 783–813.

Haldrup, Neils, Torben M. Anderson, and Jan Ros Sorensen.2000. “Labor Market Implications of EU Product MarketIntegration.” Economic Policy 15(30): 38–68.

Halfmann, Jost. 2000. “Welfare State and Territory.” In Immi-gration and Welfare: Challenging the Borders of the WelfareState, ed. Michael Bommes and Andrew Geddes. New York:Routledge, 34–51.

Hall, Peter A. 2001. “Organized Market Economics and Un-employment in Europe: Is It Finally Time to Accept Lib-eral Orthodoxy?” In Unemployment in the New Europe, ed.Nancy Bermo. New York: Cambridge University Press, 52–86.

Hall, Peter A., and David Soskice. 2004. Varieties of Capital-ism: The Institutional Foundations of Comparative Advantage.New York: Oxford University Press.

Hansson, Asa, and Karin Olofsdotter. 2008. “Integration and theStructure of Public Spending.” Comparative Political Studies41(7): 1001–7.

Heston, Alan, Robert Summers, and Bettina Aten. 2009. PennWorld Table Version 6.3. Center for International Compar-isons of Production, Income and Prices at the University ofPennsylvania.

Hicks, Alexander, and Duane Swank. 1992. “Politics, Institu-tions, and Welfare Spending in Industrialized Democra-cies, 1960–1982.” American Political Science Review 86(3):658–74.

Huber, Evelyne, Charles Ragin, and John D. Stephens. 1993.“Social Democracy, Christian Democracy, ConstitutionalStructure, and the Welfare State.” American Journal of Soci-ology 99(3): 711–49.

Huber, Evelyne, and John D. Stephens. 2001. Development andCrisis of the Welfare State: Parties and Policies in Global Mar-kets. Chicago: University of Chicago Press.

Hunger, Uwe. 2000. “Temporary Transnational Labor Migra-tion in an Integrating Europe and the Challenge to the Ger-man Welfare State.” In Immigration and Welfare: Challengingthe Borders of the Welfare State, ed. Michael Bommes and An-drew Geddes. New York: Routledge, 186–204.

Ireland, Patrick R. 2004. Becoming Europe: Immigration, Inte-gration, and the Welfare State. Pittsburg, PA: University ofPittsburg Press.

Iversen, Torben, and Thomas R. Casack. 2000. “The Causes ofWelfare State Expansion: Deindustrialization or Globaliza-tion.” World Politics 52(3): 313–49.

Kahler, Mills. 1992. “External Influence, Conditionality and thePolitics of Adjustment.” In The Politics of Economic Adjust-ment: International Constraints, Distributive Politics, and theState, ed. Stephan Haggard and Robert Kaufman. Princeton,NJ: Princeton University Press, 89–136.

Kam, Cindy D., and Robert J. Franzese Jr. 2007. Modeling and In-terpreting Interactive Hypotheses in Regression Analysis. AnnArbor: University of Michigan Press.

Kemnitz, Alexander. 2005. “Discrimination and Resistance toLow-Skilled Immigration.” Labor 19(2): 177–90.

Kivst, Jon. 2004. “Does EU Enlargement Start a Race to theBottom? Strategic Interaction among EU Member States


in Social Policy.” Journal of European Social Policy 14(3):301–18.

Lundborg, Per, and Paul S. Segerstrom. 2002. “The Growth andWelfare Effects of International Mass Immigration.” Journalof International Economics 56(1): 177–204.

Maddala, Gangadharrao S., and Shaowen Wu. 1999. “A Com-parative Study of Unit Root Tests with Panel Data and aNew Simple Test.” Oxford Bulletin of Economics and Statis-tics 61(S1): 631–52.

Marques, Helena. 2010. “Migration Creation and Diversion inthe European Union: Is Central and Eastern Europe a ‘Nat-ural’ Member of the Single Market for Labour?” Journal ofCommon Market Studies 48(2): 265–91.

Martin, John P. 1996. “Measures of Replacement Rates for thePurpose of International Comparisons: A Note.” OECD Eco-nomic Studies 26: 99–115.

Nannestad, Peter. 2007. “Immigration and Welfare States: ASurvey of 15 Years of Research.” European Journal of PoliticalEconomy 23(2): 512–32.

OECD. 1994. The OECD Jobs Study: Facts, Analysis, Strate-gies. Organization for Economic Co-operation and Devel-opment.

OECD. 2007. Benefits and Wages 2007: OECD Indicators. Paris,France: OECD.

OECD. 2009. Unit Labor Costs: Annual Indicators. Oecd.stated. Organization for Economic Co-operation and Develop-ment.

Palier, Bruno, and Robert Sykes. 2001. “Challenges andChanges: Issues and Perspectives in the Analysis of Glob-alization and the European Welfare States.” In Globaliza-tion and European Welfare States: Challenges and Change, ed.Robert Sykes, Bruno Palier, and Pauline M. Prior. New York:Palgrave Macmillan, 1–16.

Peridy, Nicholas. 2007. “Welfare Magnets, Border Effects or Pol-icy Regulations: What Determinants Drive Migration Flowsinto the EU?” Global Economic Journal 6(4): 14–39.

Pierson, Paul. 1994. Dismantling the Welfare State? Reagan,Thatcher, and the Politics of Retrenchment . New York: Cam-bridge University Press.

Quinn, Dennis. 1997. “The Correlates of Change in Interna-tional Financial Regulation.” American Political Science Re-view 91(3): 531–51.

Razin, Assaf, and Efraim Sadka. 2000. “Unskilled Migration:A Burden or a Boon for the Welfare State.” ScandinavianJournal of Economics 102(3): 463–79.

Rhodes, Martin. 1995. “‘Subverse Liberalism’: Market Integra-tion, Globalization and the European Welfare State.” Journalof European Public Policy 2(3): 384–406.

Rhodes, Martin. 2001. “Globalization, Welfare States and Em-ployment: Is There a European ‘Third Way’?” In Unem-ployment in the New Europe, ed. Nancy Bermo. New York:Cambridge University Press, 87–118.

Rodrik, Dani. 1998. “Why Do More Open Economies HaveBigger Governments?” Journal of Political Economy 106(5):997–1032.

Rosenbloom, Joshua L. 1990. “One Market or Many? LaborMarket Integration in the Late Nineteenth-Century UnitedStates.” Journal of Economic History 50(1): 85–107.

Rudra, Nita, and Stephan Haggard. 2005. “Globalization,Democracy, and Effective Welfare Spending in the Devel-oping World.” Comparative Political Studies 38(9): 1015–49.

Sanz, Ismael, and Francisco J. Velzquez. 2001. “The Evolutionand Convergence of the Government Expenditure Compo-sition in the OECD Countries.” Public Choice 119(1): 61–72.

Sarno, Lucio, and Mark P. Talyos. 2002. The Economics of Ex-change Rates. New York: Cambridge University Press.

Scheve, Kenneth F., and Matthew J. Slaughter. 2006. “PublicOpinion, International Economic Integration, and the Wel-fare State.” In Globalization and Egalitarian Redistribution,ed. Pranab Bardhan, Samuel Bowles, and Michael Waller-stein. Princeton, NJ: Princeton University Press.

Streeck, Wolfgang. 1993. “The Rise and Decline of Neocorpo-ratism.” In Labor and an Integrated Europe, ed. Lloyd Ulman,William T. Dickens, and Barry Eichengreen. Washington,DC: Brookings Institute, 80–101.

Swank, Duane. 2002. Global Capital, Political Institutions, andPolicy Change in Developed Welfare States. New York: Cam-bridge University Press.

Tomz, Michael, Jason Wittenberg, and Gary King. 2003. “Clar-ify: Software for Interpreting and Presenting Statistical Re-sults.” Journal of Statistics Software 8(1): 1–30.

Wallerstein, Michael, Miriam Golden, and Peter Lange. 1997.“Unions, Employer Associations, and Wage-Setting Institu-tions in Northern and Central Europe, 1950–1992.” Indus-trial and Labor Relations Review 50(2): 379–401.

Walz, Uwe. 1997. “Dynamic Effects of Economic Integration: ASurvey.” Open Economics Review 8(3): 309–26.

Zimmermann, Klaus F. 1995. “Tackling the European MigrationProblem.” Journal of Economic Perspectives 9(2): 45–62.

Supporting Information

Additional Supporting Information may be found in theonline version of this article:

Table 1: Effects of Lagged Unemployment Benefits onImmigrationTable 2: Including the Years of EU Treaty Enactments onthe Level of Unemployment Entitlement in DevelopedEU Welfare State: 1971–2007Table 3: Variance Inflation Factor (VIF) for VariablesTable 4: Correlations Between Explanatory VariablesFigure 1: Levels of Change in Immigration Across theRange of Left Party SeatsFigure 2: Levels of Union Density Across the Range ofChange in ImmigrationFigure 3: Levels of Integration Across the Range of Im-migrationFigure 4: Levels of Immigration Across the Range of In-tegrationFigure 5: Levels of Left Party Seats Across the Range ofImmigration


Figure 6: Levels of Immigration Across the Range of LeftParty SeatsFigure 7: Levels of Union Density Across the Range ofImmigrationFigure 8: Levels of Immigration Across the Range ofUnion Density

Please note: Wiley-Blackwell are not responsible for thecontent or functionality of any supporting materials sup-plied by the authors. Any queries (other than missingmaterial) should be directed to the corresponding authorfor the article.

Documents

Immigration, Globalization, and Unemployment Beneﬁts in