21
THE REFORM CAPACITY OF COALITION GOVERNMENTS JOHANNES LINDVALL, LUND UNIVERSITY Many scholars of comparative politics believe that countries with propor- tional representation and multi-party systems have ineffective governments since the need to coordinate decision-making among political parties pre- vents coalition governments from responding effectively to changes in the economic and social environment. This paper argues, in contrast, that coali- tion governments have high reform capacity as long as political institutions allow parties to make broad policy agreements that include side payments to parties who would otherwise oppose reform. The paper evaluates this argument empirically through an analysis of national policy responses to a common problem: the persistence of high unemployment in many of the advanced democracies in the 1990s and 2000s. 1. The Case Against Coalition Government I wish to examine the proposition that proportional representation is as- sociated with low reform capacity since coalition governments are less effec- tive than single-party majority governments. By “reform capacity,” I refer to the ability of legislatures and governments to adopt and implement policy changes that are desirable on general grounds, since they can be expected to increase some measure of social welfare (Lindvall 2010b, 362–363). The claim that coalition governments are “ineffective” is based on the assumption that distributional conflicts between coalition partners frequently delay the adoption of welfare-increasing policy changes (or render such policy changes impossible). 1 The belief that proportional representation is associated with ineffective government was widely held in the inter-war years. As William Riker (1984, 25) has noted, the disaffection with proportional representation and multi- party government in the 1930s was a result of the fact that many political commentators believed that the breakdown of democracy in Germany and Italy was, at least partly, a consequence of the ineffectiveness of German and Italian coalition governments (Finer 1935, 1961; Hermens 1941). This was a formative experience for many political scientists. When a new edition 1 Social welfare is not the only aim of government, and one may reasonably use several other criteria to evaluate political institutions in democracies. Representativeness and responsiveness come to mind (Lijphart 1999, chapter 16; Powell 2000, chapters 8–9). Recent work in political economy also suggests that electoral systems have significant distributional effects (Iversen and Soskice 2006), which is a relevant consideration from the point of view of social justice. However, the objection to proportional representation and coalition government that I consider in this paper is based on the idea that these institutions allow some groups to block reforms that are desirable from the point of view of social welfare. I wish to address this objection directly. 1

THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

Embed Size (px)

Citation preview

Page 1: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

THE REFORM CAPACITY

OF COALITION GOVERNMENTS

JOHANNES LINDVALL, LUND UNIVERSITY

Many scholars of comparative politics believe that countries with propor-tional representation and multi-party systems have ineffective governmentssince the need to coordinate decision-making among political parties pre-vents coalition governments from responding effectively to changes in theeconomic and social environment. This paper argues, in contrast, that coali-tion governments have high reform capacity as long as political institutionsallow parties to make broad policy agreements that include side paymentsto parties who would otherwise oppose reform. The paper evaluates thisargument empirically through an analysis of national policy responses toa common problem: the persistence of high unemployment in many of theadvanced democracies in the 1990s and 2000s.

1. The Case Against Coalition Government

I wish to examine the proposition that proportional representation is as-sociated with low reform capacity since coalition governments are less effec-tive than single-party majority governments. By “reform capacity,” I referto the ability of legislatures and governments to adopt and implement policychanges that are desirable on general grounds, since they can be expectedto increase some measure of social welfare (Lindvall 2010b, 362–363). Theclaim that coalition governments are “ineffective” is based on the assumptionthat distributional conflicts between coalition partners frequently delay theadoption of welfare-increasing policy changes (or render such policy changesimpossible).1

The belief that proportional representation is associated with ineffectivegovernment was widely held in the inter-war years. As William Riker (1984,25) has noted, the disaffection with proportional representation and multi-party government in the 1930s was a result of the fact that many politicalcommentators believed that the breakdown of democracy in Germany andItaly was, at least partly, a consequence of the ineffectiveness of German andItalian coalition governments (Finer 1935, 1961; Hermens 1941). This wasa formative experience for many political scientists. When a new edition

1Social welfare is not the only aim of government, and one may reasonably use severalother criteria to evaluate political institutions in democracies. Representativeness andresponsiveness come to mind (Lijphart 1999, chapter 16; Powell 2000, chapters 8–9).Recent work in political economy also suggests that electoral systems have significantdistributional effects (Iversen and Soskice 2006), which is a relevant consideration fromthe point of view of social justice. However, the objection to proportional representationand coalition government that I consider in this paper is based on the idea that theseinstitutions allow some groups to block reforms that are desirable from the point of viewof social welfare. I wish to address this objection directly.

1

Page 2: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

of his anti-proportional representation book Democracy or Anarchy? wasreleased in the 1970s, Ferdinand Hermens still maintained that proportionalrepresentation undermined democracy and national unity, suggesting thatthe apparent success of proportional representation in postwar Europe wasa “fair weather” phenomenon (Hermens 1972, 445).

The contemporary literature on “veto players” and “veto points” providestheoretical support for the old idea that a concentration of political powerprovides for more effective government, the main idea being that a “largenumber of veto players tends to ‘lock in’ economic policy and reduce itsability to respond to shocks” (Persson and Tabellini 2006, 734). One of themain predictions of George Tsebelis’s widely used veto player model is that“policy stability (defined as the impossibility of significant change of thestatus quo) will be the result of large coalition governments, particularly ifthe coalition partners have significant ideological differences among them”(Tsebelis 1999, 591; cf. Tsebelis 2002, Chapter 1). This is a prediction aboutpolicy change in general, not welfare-increasing policy changes in particular,but Tsebelis has stated explicitly elsewhere that policy stability “will makethe change of even an undesirable status quo difficult” (Tsebelis 2000, 443).In other words, even if the theory of veto players does not say whether havingmany veto players is a good or a bad thing in general – “sometimes policystability is desirable; at other times policy change is necessary” (Tsebelis2000, 443) – it clearly implies that countries with many veto players havelow reform capacity, as defined here.2

On the basis of these theoretical ideas, a rich empirical literature on elec-toral systems, government types, and policy outcomes has identified coali-tion governments with a range of undesirable phenomena, such as the mis-allocation of public spending, high budget deficits, and higher governmentdebt (see, for instance, Franzese 2002 and Bawn and Rosenbluth 2006).

The claim that single-party majority governments are associated withhigher reform capacity than coalition governments assumes that parties haveinternal mechanisms that allow them to resolve conflicts between factions.In the absence of such mechanisms, “strong” single-party majority govern-ments would be prone to intra-party bargaining failures that are equivalentto the inter -party bargaining failures that generate political stalemates inmulti-party systems (taking into account that the dominant parties in ma-joritarian systems, such as the United Kingdom or the United States, aremore heterogeneous than the relatively small parties that typically emergein multi-party systems). I will assume, for the sake of argument, that partiesdo have mechanisms of intra-party conflict resolution and control at theirdisposal.

2For another recent statement of the hypothesis that coalition governments are ineffec-tive, consider Giovanni Sartori’s Constitutional Engineering, which argues that coalitiongovernments are only able to govern well if the number of parties in government is smalland coalitions are “coalescent” rather than “conflictual” (1997, 58). If these conditionsare not met, Sartori suggests, coalition governments are often unable to overcome theirinternal divisions.

2

Page 3: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

2. The Case For Coalition Government

My counter-argument starts from the observation that if a reform is trulywelfare-increasing, political parties in favor of reform are at least in princi-ple able to make side-payments to coalition partners who would otherwiseoppose reform. Welfare-increasing reforms generate a “surplus” that can beredistributed in this manner; in fact, social welfare is sometimes defined interms of the possibility of side-payments, or compensation: according to theso-called Kaldor-Hicks Criterion, a welfare improvement can be defined as“a change which will allow of compensation being paid, and which will yetshow a net advantage” (Hicks 1939, 706).

The possibility of compensation is well-known from the literature on in-ternational trade. It is widely accepted that a policy of free trade increasesnational income (there are gains from trade; Samuelson 1962). But it isalso widely accepted that some groups are hurt by free trade, notably thosefactors of production that are specific to import-competing sectors of theeconomy. As Stolper and Samuelson argued in a seminal article on tradeliberalization, “the harm which free trade inflicts upon one factor of pro-duction is necessarily less than the gain to the other. Hence, it is alwayspossible to bribe the suffering factor by subsidy or other redistributive de-vices so as to leave all factors better off as a result of trade” (Stolper andSamuelson 1941, 73). Authors such as Cameron (1978), Katzenstein (1985),Rodrik (1998), and Hays, Ehrlich, and Peinhardt (2005) have shown thatmany governments in the postwar world – especially coalition governments– used various forms of welfare spending to compensate groups that wereharmed by free trade.

In the empirical part of this paper, I will argue that the “flexicurity”agenda in European politics in the 1990s and 2000s can be seen as a re-form package whereby political parties who wish to liberalize employmentprotection or reform unemployment benefit systems offer to compensateworkers through investments in active labor market programs and unem-ployment compensation for the short-term unemployed (on the potential forsuch trade-offs in labor market policy, see, for example, Pontusson 2005,216–217; for a discussion of “flexicurity,” see Viebrock and Clasen 2009).Similarly, the empirical literature on pension reforms in European democra-cies suggests that governments in multi-actor systems tend to build supportfor reform by designing policy packages that include side-payments to (some)losers (see, for example, Bonoli 2000 and Hausermann 2010; for an extendeddiscussion, see Lindvall 2010b, Section 2).

The role of side payments has not been investigated thoroughly in the ex-isting literature on coalition governments. One of the few empirical studiesthat deal explicitly with the relationship between electoral systems and thequality of public policies is Markus Crepaz’s paper “Constitutional Struc-tures and Regime Performance” (1996), which argues that the inclusivegovernments that are typical of PR systems are more “responsive” to theelectorate and therefore pursue superior policies. Similarly, Arend Lijphartdevoted one chapter of his book Patterns of Democracy (1999, chapter 15) toa defense of the performance of “consensual” democracies, claiming that it

3

Page 4: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

is at least as good as the performance of majoritarian democracies (for a re-analysis, more skeptical about the effects of political institutions, see Roller2005, 228–237). But neither Crepaz’s paper, nor Lijphart’s book, nor thegeneral comparative politics literature that defends consensual democracyagainst its majoritarian critics (see, for example, Powell 2000), addressesthe specific claim that coalition governments in PR system are prone todeadlock since the opposition of some coalition parties may prevent govern-ments from adopting desirable policies. Lijphart’s argument, for instance,is that although single-party majority governments may be more decisive,consensual systems have other advantages, such as a greater capacity fordeliberation and a more “steady” policymaking style (Lijphart 1999, 259–260).

If it is possible to compensate losers, it seems that increasing the numberof coalition partners, or the ideological distances among them, should notreduce reform capacity, at least not in principle.

One potential complication is that even if it is possible to make sidepayments, increasing the number of veto players tends to increase the trans-action costs associated with political bargaining (Dixit 1996). As Cox andMcCubbins note, a large number of veto players makes it more difficult to“structure negotiations,” and to ensure that each party “receives sufficientvalue to accept the deal” (2001, 27).

As Birchfield and Crepaz (1998, 181–182) argue, however, there are im-portant differences between veto players who operate through separate in-stitutions, competing with one another – which is incidentally what Coxand McCubbins are mainly concerned with – and veto players who inter-act face-to-face on a regular basis, working together for common purposes(such as parties in coalition governments). The transaction costs associatedwith political bargaining are significantly lower for “collective” veto players:coalition governments in particular tend to have procedures available thatallow them to “structure negotiations,” and the wide remit of most gov-ernments in advanced democracies arguably provides coalitions with manymeans of sharing the surplus generated by welfare-increasing reforms (en-suring that all parties receive “sufficient value”). This suggests that coali-tion government as such is not a cause of low reform capacity. But it alsosuggests that intra-coalition bargaining may become involved if coalitiongovernments operate in political environments with many institutional vetoplayers. If other agents and institutions – outside the coalition – are able toinfluence the implementation of some policy programs, those programs areeffectively removed from the bargaining table, reducing the government’sroom for maneuver.

Tsebelis criticizes the distinction between competitive and collective vetopoints for being imprecise (Tsebelis 2002, 88). For example, Tsebelis claims,it is not clear why the interaction between two chambers of parliament (com-petitive veto points, according to Birchfield and Crepaz) should be less “face-to-face” than the interaction between government and parliament (collectiveveto points). This is an important point, for one should not underestimatethe possibilities for communication between “institutional” veto players (we

4

Page 5: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

can think of Germany, with its Mediation Committee acting as an interme-diary between the Bundestag and the Bundesrat). However, if one comparesthe interaction between members of a coalition government with the interac-tion between a national government and sub-national governments in federalsystems, for example, it seems clear that the flow of communication betweenparties in a coalition is more dense, increasing their ability to make packagedeals.

The remainder of this paper examines empirical evidence on Europeanlabor market reforms, in order to evaluate the theoretical claims made inthis section. Section 3 and the first half of Section 4 demonstrate that thereis no significant difference between “strong” single-party majority govern-ments and “weak” coalition governments (or minority governments) when itcomes to the adoption of employment-oriented labor market reforms in theWestern European states. The second half of Section 4 shows that wherecoalition governments have not adopted extensive labor market reforms, asin Belgium, this outcome can be attributed to the absence of national con-trol over those programs that have typically been used to compensate thelosers from labor market reform.

3. Labor Market Reforms in Europe

In the 1980s, 1990s, and 2000s – after the economic downturns of themid-1970s and early 1980s – many European countries were confronted witha difficult economic and social problem: high and persistent unemploymentrates. The remainder of this paper compare national policy responses tothe problem of unemployment in order to assess whether countries with het-erogenous coalition governments or minority governments have lower reformcapacity than countries with ideologically cohesive majority governments.

According to an influential literature in labor economics, one reason whyit proved so difficult for many European countries to reduce unemploymentafter the crises of the 1970s and early 1980s is that some of the labor mar-ket policy programs that were established after the Second World War hadbecome unsuitable in the social and economic environment of the late twen-tieth century. Most importantly, labor economists have long argued that anunlimited or very long duration of unemployment benefit entitlements in-creases structural (non-cyclical) unemployment, especially if unemploymentbenefits are not coupled with strong activation requirements (Bassanini andDuval 2006; Fredriksson and Holmlund 2006; Layard, Nickell, and Jackman1991; Nickell and Layard 1999).

The idea that unemployment insurance systems, and other forms of socialprotection, should be designed to encourage the unemployed to seek newjobs has become broadly accepted among experts on the labor market in thecourse of the 1990s and 2000s (Lindvall 2010a, Chapter 4). As the economistOlivier Blanchard (2006, 45) noted in a recent survey article, there is nowa “consensus” among economic experts on why some labor markets workbetter than others.

While there is a trade-off between efficiency and insurance,the experience of the successful European countries suggestsit need not be very steep. What is important in essence is

5

Page 6: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

to protect workers, not jobs. This means providing unem-ployment insurance, generous in level, but conditional on thewillingness of the unemployed to train for and accept jobs ifavailable.

Similarly, in a recent book, the sociologist Lane Kenworthy (2008, 136)outlines a “policy package” that is intended to “provide generous benefitsto working-age individuals and households who need them without creatingexcessive employment disincentives.”

The package features generous transfers to those unable towork due to involuntary job loss, sickness, disability, or fam-ily responsibilities. However, benefits provided on a tempo-rary basis should be of relatively short duration, and eligibil-ity criteria for those provided on a permanent basis should befairly strict. In exchange for this strictness, extensive sup-port should be provided for those entering or returning tothe work force, in the form of training, job placement, publicemployment, and childcare.

Finally, the political scientist Cathie Jo Martin (2004, 43) describes a new“conception of the welfare state,” which combines “sticks” (stricter benefitrules and shorter benefit duration) and “carrots” (active labor market poli-cies) in order to “reintegrate the long-term unemployed back into the coreeconomy.”

There are several reasons why labor market reform is a good test casefor the purposes of this paper. First, assuming that the literature citedabove has described the relationship between unemployment benefits andstructural unemployment correctly, labor market reforms have clear welfareimplications (there is broad agreement that unemployment has large socialcosts). Second, “demanding” labor market reforms – reforms that put pres-sure on the unemployed to look for work (Eichhorst et al. 2008) – haveclear distributional implications: although they may result in lower struc-tural unemployment, they tend to harm some groups. Third, it is relativelystraightforward to identify the “side payments” that governments may useto compensate the “losers” (principally training programs and other activelabor market policies that improve the employability of the unemployed andhelp to match them with potential employers).

The first column in Table 1 shows the mean number of labor marketreforms per year in fourteen European countries between 1980 and 2006,including only those reforms that strengthened the incentives to work (ac-cording to a new dataset of unemployment benefit reforms that has beendeveloped at the Fondazione Rodolfo Debenedetti in Milan and the Institutzur Zukunft der Arbeit in Bonn). The second column shows the mean num-ber of “structural” reforms per year, where “structural” reforms are reformsthat affect all categories of wage earners. Both columns compare ideolog-ically cohesive majority governments (105 country-year observations) withall other types of governments (273 country-year observations). As Table1 reveals, both the rate of reform in general and the rate of structural re-form in particular were slightly (but not significantly) higher under “weak”coalition governments and minority governments than under ideologically

6

Page 7: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

Table 1. Reforming Unemployment Benefits in the EU-15,1980–2006 (mean number of reforms per year)

All Reforms Structural Reforms

Cohesive majority governments 1.02 0.07All other governments 1.07 0.11

Source: Fondazione Rodolfo Debenedetti (2010). The sample con-sists of Austria, Belgium, Denmark, Finland, France, Germany,Greece, Ireland, Italy, the Netherlands, Portugal, Spain, Sweden,and the United Kingdom (the EU-15 countries, excluding Luxem-bourg). Only reforms that increased work incentives are included.Comprehensive policy changes that address “the broader design ofexisting systems, rather than their minor features,” are coded asstructural. An unemployment benefit reform is “structural” if it af-fects the working-age population as a whole. “Cohesive majoritygovernments” are governments where one party family (left, right,Christian democratic, or centrist) controls more than 50 percent ofall legislative seats, according to the Swank (2006) dataset. Individ-ual observations have been coded as “cohesive majority governments”in election years if an ideologically cohesive majority government heldpower for most of the year.

Table 2. Reforming Benefit Duration in the EU-15, 1980–2006 (mean number of reforms per year)

All Reforms Structural Reforms

Cohesive majority governments 0.05 0.02All other governments 0.05 0.02

Source: Fondazione Rodolfo Debenedetti (2010). Only reforms thatreduced the duration of unemployment benefit entitlements included.For all other definitions, see Table 1.

cohesive majority governments. The data in this table thus provide littlesupport for the idea that reform capacity is higher under “strong,” ideolog-ically cohesive majority governments.

As I noted above, labor economists have identified a system of unlimitedor long-lasting unemployment benefits as highly likely to increase structuralunemployment. Table 2 therefore examines a subset of the reforms in Table1, namely those that involve a reduction of the duration of unemploymentbenefit entitlements. Here, the mean number of reforms is more or lessidentical under the two different types of governments, suggesting once morethat coalition governments and minority governments do not have lowerreform capacity than single-party majority governments.

7

Page 8: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

Table 3. Unemployment in the EU-15, 1980–2006

Unemployment rates

“Strong” governments (percent) 1980 2006 Diff.

Greece 100 2.8 8.8 6.0United Kingdom 100 5.7 5.4 −0.3France 74 5.8 9.0 3.3Spain 41 11.5 8.5 −3.0Portugal 37 7.8 7.7 −0.1Germany 22 3.2 10.4 7.1Austria 11 1.9 4.8 2.9Ireland 4 7.4 4.4 −3.0Belgium 0 8.1 8.3 0.2Denmark 0 6.9 4.1 −2.8Finland 0 4.7 7.7 3.1Italy 0 7.7 6.9 −0.8Netherlands 0 6.2 3.9 −2.3Sweden 0 2.2 7.1 4.9

Data source (for unemployment): Armingeon et al. (2009). “Strong”governments are ideologically cohesive majority governments, as de-fined in Table 1.

Turning from policies to outcomes, Table 3 describes the developmentof unemployment rates in the same fourteen countries. As the figures inthe final column show, unemployment trends have varied a great deal acrosscountries (in some countries, unemployment has decreased; in some countriesit has increased), but there is no evidence that countries with a history ofideologically cohesive majority governments have dealt more effectively withthe problem of unemployment than countries with a history of heterogenouscoalition governments and minority governments. In three of the four coun-tries (in this sample) that have managed to reduce unemployment by morethan one percentage point (Denmark, Ireland, and the Netherlands, two ofwhich are examined in detail below), all governments have been coalitiongovernments or minority governments from 1981 onwards.

The labor market reform in Denmark in 1994, which is one of the few“structural” reforms of unemployment benefit duration in the dataset that Iused for Tables 1 and 2, is a good example of how side-payments can be usedto facilitate reform. By the early 1990s, there was broad political agreementin Denmark that unemployment had become a structural problem in thesense that economic upswings did not translate into lower unemploymentrates (Torfing 1999, 13–14; Larsen and Andersen 2009, 244–249). Therewas also broad agreement that this was a result of misguided labor marketpolicies, notably the introduction of de facto unlimited unemployment ben-efits in the late 1970s. However, the trade unions and the Social Democratswere only willing to accept stricter qualification requirements and shorterbenefit duration if they got something else in return, and what they wantedwere large investments in Swedish-style active labor market policies. This

8

Page 9: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

was the political bargain at the heart of the Danish labor market reforms,which appear to have contributed to the Danish “employment miracle” inthe 1990s (Andersen 2006).

4. Four European Experiences

I will now proceed to present an investigation of four European countries,which is organized around two paired comparisons: a comparison of Britainand Ireland and a comparison of Belgium and the Netherlands. The caseselection is based on three considerations. First of all, these four countrieshave all experienced relatively long periods of high unemployment, in bothabsolute and relative terms. The unemployment rates in Belgium, Ireland,the Netherlands, and the United Kingdom were higher than the Europeanaverage for much of the 1980s (the Belgian unemployment rate has remainedrelatively high, whereas the other three countries have managed to reduceunemployment considerably). Second, my argument is concerned with bothcabinet composition and institutional constraints on the central state gov-ernment, and the selection of these four cases enables me to examine theinteraction of those factors. Third, since different types of welfare statesconfront different challenges when they seek to make “employment-friendly”reforms (Hausermann and Palier 2008), I wish to compare countries withsimilar welfare state structures and varieties of capitalism.

Britain and Ireland are two liberal market economies with liberal welfarestates, but their political systems are different: Britain has a tradition ofsingle-party governments and Ireland has a tradition of coalition govern-ments. Belgium and the Netherlands, on the other hand, are coordinatedmarket economies with more conservative welfare state traditions. Bothhave a history of coalition governments (with party systems dominated bythe same ideological families), but the Netherlands is a centralized, unitarystate whereas Belgium is federal and decentralized (Belgium only becameformally federal in 1993, but many labor market policies – such as placementservices and other active labor market programs – were already decentral-ized in the 1980s). The comparison between Britain and Ireland shows thatthe labor market reform trajectories of these two countries have been rela-tively similar, in spite of the differences between their political systems. Thissuggests that Ireland’s multi-party system, and its history of coalition gov-ernment, has not decreased the reform capacity of Irish governments. Thecomparison between Belgium and the Netherlands suggests that the absenceof significant reforms in Belgium – compared to neighboring Netherlands,which has been one of the most reform-intensive countries in Europe – canbe explained by Belgium’s federal structure, or, to be more specific, by theallocation of labor market competencies within the Belgian political system.Apart from the issue of federalism, Belgium and Ireland are constitutionallyquite similar: they are parliamentary monarchies with weak upper chambersand no judicial review (Huber et al. 2004). This makes it easier to identifythe role that federalism has played.

The empirical analysis is mainly concerned with the behavior of governingparties in certain crucial periods (specific to each country), concentrating onthe role of social democratic parties and other left-wing parties since they

9

Page 10: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

are more likely than other parties to oppose reforms that involve a reductionin the overall generosity of unemployment benefits (including the durationof benefit entitlements). When it comes to “side payments,” the discussionis therefore mainly concerned with policy changes that social democraticparties have promoted. I also take the role of the trade unions into account,given that trade unions and social democratic parties are closely affiliatedin many European countries. I suggest, however, that party politics ex-plains the main differences between the four countries that I examine: bothBelgium, Ireland, and the Netherlands have social partnership institutionsthat involve trade unions and employer organizations in national-level poli-cymaking (although in Ireland, this has only been the case since 1987), yetthe outcomes were different. This observation is consistent with the argu-ment that effective corporatist policy concertation requires a governmentthat is able to encourage, implement, and enforce agreements among thesocial partners, and between the social partners and the state (Hemerijckand Vail 2005).

Labor Market Reforms in Britain and Ireland. Britain has a majoritarianfirst-past-the-post system, whereas Ireland’s single-transferable vote systemhas highly proportional effects. As a result of this institutional difference(and the fact that the two countries have different cleavage structures),Ireland has a more fragmented party system, which has had obvious effectson government formation: single-party majority governments are the normin Britain but the exception in Ireland. The last time the dominant Irishparty Fianna Fail had a majority in the Irish parliament, the Dail, was in1981. Since then, all Irish governments have been coalition governmentsor minority governments. In spite of these political differences, British andIrish labor market reforms in the 1980s, 1990s, and 2000s have been quitesimilar – at least when it comes to unemployment compensation and effortsto activate the unemployed. I will concentrate on two major sets of reforms:the elimination of earnings-related unemployment benefits in the 1980s andthe “activation agenda” of the 1990s.

The Irish party system has some peculiar features, which has led scholarsof comparative politics in the past to treat it as a special case. However,Peter Mair (1992, 383) has argued that there is an “essential comparability”and Michael Laver (1992, 380) has shown that the Irish parties can be asso-ciated with the main European party families. Laver finds, for example, thatthe Irish Labour Party “fulfils a role in the Irish party constellation just likesocial democratic parties in almost every European country” (Laver 1992,381), which is noteworthy since most of the labor market reforms that Idescribe below were adopted by coalitions that involved the Labour Party.

In the first half of the 1980s, Britain and Ireland became the first Euro-pean countries that shifted from income-related to flat-rate unemploymentbenefits, homogenizing insurance-based unemployment benefits and means-tested unemployment assistance. The change was more gradual in Ireland,where pay-related benefits were phased out by various coalition governmentsover a ten-year period, but the similarities are more striking than the dif-ferences.

10

Page 11: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

After the introduction of “earnings-related” benefits in Britain (1966) and“pay-related” benefits in Ireland (1973–1974), the generosity of British andIrish unemployment compensation was close to the European average, butafter the elimination of earnings-related benefits in Britain in 1981 and thelarge cuts in pay-related benefits in Ireland in 1983 and 1987 (the com-plete abolition followed in 1994, when pay-related benefits had already be-come largely irrelevant to social protection), unemployment benefits in bothcountries became low by European standards (Scruggs 2004).3 In Britain,earnings-related benefits were eliminated by a Conservative single-partygovernment under Margaret Thatcher, whereas the first major cut in pay-related benefits in Ireland was made by a coalition between the center-rightparty Fine Gael and the Irish Labour Party.

Previous research on the British case shows that the two main motivesfor the Thatcher government’s unemployment benefit reforms were costs anddisincentives. According to Jochen Clasen’s study of British labor marketreforms (2005, 77), these policy changes were intended “to save money and toaddress the ‘why work’ problem, that is, to widen the gap between benefitsand income form work.” Concerning the second motive, Clasen has notedthat there were “frequent claims of a rising degree of workshyness” – anold rhetoric that the Conservatives “revived and emphasized” (Clasen 1992,287).

Irish governments had similar motives. The immediate reason for the re-duction of pay-related benefits was the desire to reduce the budget deficit(gross general government debt increased from less than 50 percent of GDPto almost 100 percent of GDP between 1973 and 1983; see Franzese 1998),but the argument that replacement rates must be reduced in order to en-courage the unemployed to look for work also mattered – the removal of thebenefits was “part of a wide panoply of policies implemented from 1982 on-wards, directed at reducing the replacement ratio facing unemployed work-ers,” according to one Irish social policy expert (McCashin 2004, 200). In1983, when the first cutbacks in pay-related benefits were made, this ar-gument was especially popular within the main coalition party, Fine Gael,but it was also accepted by their partner, the Labour Party. For example,when the Minister for Social Affairs – the Labour politician Barry Desmond– introduced the 1983 Social Welfare Bill, he said that one of the purposesof reducing the pay-related unemployment benefit was to discourage abuseof the unemployment benefit system (Dail Eireann, Volume 340, 9 March1983).

[One paragraph removed since it contained unconfirmed interview quotes.]Nevertheless, the unemployment insurance reform in 1983 was politically

difficult for the Labour Party, especially with respect to its affiliation withsome of the leading Irish trade unions. So why did the Labour Party agree

3In 1983, the rate of the pay-related unemployment benefit in Ireland declined from40 to 25 percent of previous earnings, and a cap of 80 percent of previous income wasintroduced for total unemployment benefit payments, declining to 75 percent the nextyear. There were further cuts in 1987, under the new Fianna Fail government, whichreduced the replacement rate to 12 percent. By the early 1990s, the benefit was thus “ashadow of the original benefit” (Tansey 1991, 59–62; see also Department of Social Welfare1985, 15 and Department of Social Welfare 1989, 31).

11

Page 12: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

to these cuts? The answer appears to be that the Labour Party was moreconcerned with long-term unemployment benefits and other benefits for thepoorest, and the reduction of short-term benefits was a part of a packagedeal that involved the maintenance of basic social protection.

[One paragraph removed since it contained unconfirmed interview quotes.]The consequence was that the difference between long-term and short-

term benefits became smaller, as short-term benefits were cut while thebasic rate was increased.4 The strategy of eliminating pay-related benefitsin order to free resources for the expansion of more basic social welfare ar-rangements was also advocated by the Commission of Social Welfare, whichwas established by Barry Desmond in 1983, delivered its final report in 1986,and had significant influence over social policy in the following years.

Today, the homogenization of unemployment benefits and unemploymentassistance has become common in Western Europe, but Britain and Irelandwere the first countries that moved in this direction. It is not clear that thecuts in replacement rates in the 1980s had any positive effects on employment(unemployment remained relatively high in Britain and Ireland until themid- to late 1990s). However, the fact that governments in Britain andIreland made such similar reforms, with similar motives and similar effectson the nature of the unemployment protection system as a whole, suggeststhat Ireland’s history of coalition governments did not stand in the way ofradical policy change.

In the mid-1990s, governments in Britain and Ireland developed new acti-vation policies, in order to deal with the problem of “unemployment traps”– the fact that many of the unemployed could only hope for low-paying jobsthat were less attractive, all things considered, than the income from un-employment benefits or unemployment assistance payments combined withselective programs available to the long-term unemployed. Eichhorst et al.(2008, 4–7) distinguish between “demanding” and “enabling” activationpolicies, where reforms that make policies more “demanding” are designedto increase the job search activity of the unemployed and the likelihood thatthey will accept job offers, whereas reforms that make policies more “en-abling” are designed to increase the employability of the unemployed andfacilitate the transition to paid employment. British and Irish governmentshave used a combination of “demanding” and “enabling” policies. AlthoughBritain introduced some of these programs earlier, the similarities are againmore striking than the differences: in his study of activation reforms in Eu-rope, Weishaupt (2008, 256–257) concludes that both Britain and Irelandhave “endorsed the ‘rights and duty’ rhetoric,” seeking to “increase activitylevels and employment rates.”

In the course of the late 1980s and early to mid-1990s, the Conservativegovernments in Britain emphasized “demanding” reforms, making benefitprovision stricter by “tightening the conditions of active job-search for thereceipt of unemployment benefits, relaxing the definition of appropriate em-ployment, and toughening sanctions for noncompliance” (Clasen and Clegg2006, 540). After the Labour Party’s election victory in 1997, Tony Blair’s

4By 1989, this process had gone so far that the maximum level of unemploymentassistance exceeded the maximum unemployment benefit (Tansey 1991, 56, 59).

12

Page 13: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

new government launched its “New Deal” agenda, which made job-searchrequirements and work requirements even stricter but combined those mea-sures with more “enabling” efforts to assist and counsel the unemployed(Clasen 2005, 81).

The Irish approach in the mid-1990s was based on “enabling” activation.Between 1994 and 1997, Ireland was governed by a “rainbow” coalition be-tween three ideologically diverse parties: Fine Gael, Labour, and the Demo-cratic Left. Perhaps the most important labor market policy measure of thisgovernment was the introduction of the Back to Work Allowance in 1994,which allowed long-term unemployed returning to work to keep some of theirbenefits over a three-year period, including “secondary” benefits (FitzGerald2005, 135). As Weishaupt (2008, 291) notes, the Back to Work Allowanceaddressed with the problem of unemployment traps since it “made virtuallyany low-paid employment more attractive than remaining on unemploymentbenefits.” Other measures introduced around this time included lower taxesand social insurance contributions for low-wage earners (McCashin 2004,210–211).

[One paragraph removed since it contained unconfirmed interview quotes.]The Back to Work Allowance is an “in-work benefit,” intended to facilitate

the transition from unemployment to paid work. British governments haveused similar instruments – as Martin Rhodes (2000, 47) noted in his overviewof the British welfare state, the first in-work benefits were introduced inBritain already in the 1970s. However, as Rhodes also points out, suchinstruments became more widespread and generous in the mid-1990s, whenthe so-called Family Credit was extended to childless couples and singleadults. In that sense, the timing was similar in Britain and Ireland.

The right-wing Fianna Fail–Progressive Democrats coalition that wasformed in Ireland under Bertie Ahern in 1997 introduced a new program of“demanding” activation, called the “Preventative Strategy.” This programincreased the job search requirements for the young unemployed, and intro-duced, Weishaupt (2008, 294) argues, “Ireland’s first formal and systematicapproach to interacting with all job-seekers,” establishing a new approachto activation based on “early intervention, individual case management, andcontractualized action plans.” These measures were in many ways similarto what the new Labour government did in Britain at this time.

To sum up, both Britain and Ireland made a series of labor market policychanges in the 1990s. The mix of “demanding” and “enabling” measureswas slightly different in the two countries, but the essential ideas were thesame. Political institutions appear to have had some effect on the policystyles of governments, but not so much on policy substance.

The British government is the paradigmatic example of a government witha lot of room for maneuver: no institutional or partisan veto players has theauthority to prevent a party with a majority in the House of Commonsfrom implementing its policy agenda. Yet, when we compare labor marketreforms in Britain to labor market reforms in Ireland, it is not clear thatthe reform capacity of British governments has been significantly higher,even if most of the Irish reforms that I have considered were adopted bygovernments that included the Labour Party – a social democratic party

13

Page 14: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

that might have been expected to resist the reduction in social protectionthat some of the reforms have resulted in.

One important explanation is arguably that Ireland is a unitary statewith few institutional veto players. Intra-coalition bargaining is facilitatedby a concentration of power in central government, so Irish governmentshave been able to use package deals to establish broad support for poten-tially controversial reforms. Both in the 1980s and 1990s, the Labour Partyagreed to cutbacks in the unemployment benefit system since their coali-tion partners backed other policies that the Labour Party wanted: in the1980s, the Labour Party prioritized long-term benefits for the poor, and inthe 1990s, the Labour Party gave priority to “enabling” activation measuressuch as in-work benefits and active labor market programs.

Labor Market Reforms in Belgium and the Netherlands. In the early 1980s,open unemployment peaked in Belgium and the Netherlands, reaching ap-proximately 11 percent of the labor force in both countries. At this time,Belgium and the Netherlands had high unemployment in both absolute andrelative terms. Since then, the two countries have diverged. By the early2000s, the Netherlands was widely regarded as one the best examples of suc-cessful labor market reform in Europe. For example, in a contribution to theWelfare and Work project (Scharpf and Schmidt 2000), Hemerijck, Unger,and Visser (2000, 229) wrote that a series of reforms in the Netherlands inthe course of the 1990s had contributed to the large fall in unemployment:

The reform of the employment service, the strengthening ofactivation requirements in social security, additional job pro-grams in the public sector, wage subsidies to encourage em-ployers hiring low-skilled workers, tax reduction for those inwork, negotiated flexibility in working hours, statutes andpensions, and policies to enhance the growth of part-timework – are all components of the Dutch job miracle.

In the 2000s, the system “flexicurity” that the Netherlands exemplified be-came one of the main blueprints for labor market reform in Europe (Viebrockand Clasen 2009).

Meanwhile, unemployment remained high in Belgium in the 1990s, andalthough there have been a few important policy changes in the 2000s (asI explain below), a long series of OECD reports have been critical of theabsence of reform, especially when it comes to the unemployment bene-fit system. For example, the OECD’s 2007 Economic Survey of Belgiumcalls for more “efficient activation,” pointing out that the unlimited dura-tion of unemployment benefits is now an almost unique exception amongOECD countries, and suggesting that the generosity of unemployment ben-efits should be reduced for long-term unemployed in order to encourage jobsearch activity (OECD 2007, 66–71). At least until the activation reformthat was implemented by the liberal–socialist coalition government in 2004,Belgium was a clear example of a policy combination that labor economistsare particularly critical of – largely unconditional benefits with unlimitedduration (Layard, Nickell, and Jackman 1994, 91–93).

14

Page 15: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

The fact that Belgium and the Netherlands have diverged in this manneris rather puzzling, since there are important political similarities betweenthe two countries. Belgium and the Netherlands are both dominated bythe same three ideological traditions: Christian democracy, social democ-racy, and secular liberalism. Belgium has an additional, linguistic dimen-sion – there are two social democratic parties (one French, one Flemish), twoChristian democratic parties, and two liberal parties – but when it comesto the ideological composition of governments, the similarities are striking.Both countries had right-wing coalitions under Christian democratic primeministers for most of the 1980s, but in the late 1980s, Wilfried Martens andRuud Lubbers reshuffled their cabinets, bringing in the social democrats(this happened in 1988 in Belgium and 1989 in the Netherlands). The newcoalitions lasted until 1994 (the Netherlands) and 1999 (Belgium), whenthey were replaced by “purple” coalitions between the social democrats andthe liberals (with the social democratic leader Wim Kok as prime ministerin the Netherlands and the Flemish liberal leader Guy Verhofstadt as primeminister in Belgium. The Dutch purple coalition lasted between 1994 and2002 whereas the Belgian lasted from 1999 to 2007. (Here, the similaritiesend, for Kok’s government in the Netherlands was replaced by a new right-wing coalition whereas the Belgian governments after 2007 have been broadcoalitions of left- and right-wing parties.)

The fact that the Netherlands was governed by ideologically heteroge-neous coalitions from the late 1980s to the early 2000s did not stop the Dutchgovernment from adopting significant reforms in this period. The reason wasarguably that governments were able to use side payments, making pack-age deals that linked “demanding” activation measures (such as strongerjob search requirements for the unemployed) to “enabling” measures (suchas social protection for part-time workers and an expansion of active labormarket programs). Gerrit Zalm, a liberal politician who served as financeminister in Wim Kok’s “purple” governments in 1994–2002, says that thiscombination of policies was a “trade-off” between the Social Democrats andhis own Liberal Party (Lindvall 2010a, 159). Zalm did not have much faithin active labor market programs: “I am from the Liberal Party; I was morein favor of tax incentives for work,” he said in an interview with the author.He also says, however, that being able to offer subsidized employment al-lowed the public authorities to put more pressure on the unemployed: “Theway you execute the benefit schemes can become more harsh. And of course,from the point of view of a liberal, that is not a bad thing.”

From the point of view of the theoretical arguments presented in section2, the fact that the purple governments in the 1990s were able to engagein quid-pro-quo bargaining is not surprising: the Netherlands is a unitarystate where the central government controls policy implementation, and theDutch constitution does not provide for judicial review of acts of parliament(the Netherlands does not have a constitutional court), so no institutionalveto players can challenge the national government; the transaction costsassociated with intra-coalition bargaining are consequently low.

The main question concerning Belgian labor market policy is why therehas been no overhaul of the unemployment benefit system. Throughout the

15

Page 16: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

1980s and 1990s – and arguably still in the 2000s – Belgian governmentshave contented themselves with minor adjustments, as opposed to any moresignificant changes (Van der Linden 1997, 290). Since it is difficult to studythe causes of a non-event, I will concentrate on the most important changethat did occur in the Belgian unemployment benefit system in the 1990sand 2000s: the 2004 activation reform, which was initiated by Frank Van-denbroucke, a Flemish social democratic politician and federal minister oflabor in the liberal–social democratic coalition in the mid-2000s. This re-form did not change the basic parameters of the Belgian unemploymentinsurance system (benefits still have unlimited duration and a constant re-placement rate), but it involved closer monitoring of the job search behaviorof the unemployed and new sanctions for unemployed individuals that donot actively seek work (Faniel 2005, 133–134). Although this policy changewas quite modest compared to unemployment benefit reforms elsewhere inEurope in the 1980s, 1990s, and 2000s, it came late, proved difficult forthe government to adopt, and resulted in widespread social unrest acrossBelgium in the spring of 2004.

[One paragraph removed since it contained unconfirmed interview quotes.]The distribution of power between the national government and Belgium’s

three regions (Flanders, Wallonia, and Brussels) is complicated. In the areaof labor market policy, the social security system, including unemploymentbenefits, is a federal matter, as is employment protection (although this isalso negotiated with the social partners), but the regions are responsible foractive labor market programs, which means that there is a “fragmentationof competences” in the domain of labor market and social policy (Cantillonet al. 2006, 1035–1036; for a comprehensive overview of the distribution ofcompetencies in social and labor market policy in Belgium, see also de Deken2007).

In order to reform the unemployment benefit system in 2004 it was nec-essary for the federal government to make a so-called “cooperation agree-ment” with the regional governments. This process began in December2003. One important reason why it proved difficult to reach agreementwithin the coalition – and to persuade the main trade union confederations(who are affiliated with Belgium’s Christian democratic and social demo-cratic parties) – is that this cumbersome political process complicated thedevelopment of policy packages that combined “demanding” and “enabling”activation measures. Although, as we have already seen, the responsibleminister, Frank Vandenbroucke, presented the reform as a move towardsa “Scandinavian” model, combining relatively strict conditionality require-ments with programs that would help the unemployed to train for and findnew jobs, active labor market policy is largely the domain of the regions, sothe government was unable to include specific commitments to “positive”activation measures in its proposal: Mr. Vandenbroucke could only promisethat he would initiate negotiations with regional governments concerningthe accompagnement of the unemployed (Faniel 2005, 140).

It seems reasonable to expect that a similar reform would have comesooner if it had been easier for Belgian coalitions to make comprehensivepackage deals. Indeed, it is not unlikely that more radical reforms would

16

Page 17: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

have been possible if that were the case, since there are proposals for radicalreform at the highest levels: the coalition government that was formed in2008 said in its coalition agreement that it will increase the replacement ratefor the first period of unemployment but lower it for long unemploymentspells, in order to encourage job search, but as yet nothing has come of this.

For Hemerijck and Schludi (2000) and Hemerijck, Unger, and Visser(2000), Belgium’s low reform intensity is associated with the conflict betweenthe French- and Flemish-speaking communities (see also Beland and Lecours2005, 277–281). The ongoing federalization process has undermined the “ca-pacity for coordination and issue linkage at the national level,” Hemerijckand Schludi argue, which means that reforms will only be possible if new“political and societal coalitions” can “reduce the plethora of veto powersin the Belgian political economy” (Hemerijck and Schludi 2000, 166). Myconclusion is similar, although I have a slightly different interpretation ofwhy “veto powers” matter: in my view, the institutional fragmentation ofpolitical power makes it difficult for political parties at the central level toconstruct broad policy packages.

In previous sections, I have shown how left-wing parties and trade unionsin Ireland and the Netherlands have agreed to political trade-offs betweenunemployment benefit reforms, active labor market policies, and other ef-forts to support the unemployed – “enabling” activation measures were aform of side payment that social democratic parties got in return for agree-ing to the “demanding” measures that other parties favored. The fact thatthis type of bargain does not appear to have been possible in Belgium is atleast partly a consequence, I argue, of the allocation of policy competencieswithin Belgium’s federal structure.

5. Conclusions

This paper argues that the reform capacity of coalition governments isjust as high as the reform capacity of single-party majority governments, aslong as political institutions allow coalition partners to develop broad policypackages that provide side payments to parties who would otherwise opposereform. In other words, when coalition governments fail to take desirablepolicy action in response to pressing economic and social problems, this isnot a result of inherent flaws in systems with proportional representationand multiparty competition; it is a problem associated with certain un-fortunate combinations of institutions. The notion that strong, single-partymajority governments are more effective than coalition governments, so thatthe choice of electoral system involves a trade-off between effectiveness andrepresentativeness, has always been one of the strongest arguments in favorof majoritarian politics in general and plurality voting in particular. Thispaper throws some doubt over this common belief.

17

Page 18: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

References

Andersen, Torben M. 2006. “From Excess to Shortage – Recent Devel-opments in the Danish Labor Market.” In Structural Unemployment inWestern Europe, edited by Martin Werding. Cambridge: MIT Press, 75–101.

Armingeon, Klaus, Marlene Gerber, Philipp Leimgruber, and MichelleBeyeler. 2009. “Comparative Political Data Set 1960–2007.” Instituteof Political Science, University of Berne.

Bassanini, Andrea and Romain Duval. 2006. “The Determinants of Unem-ployment Across OECD Countries.” OECD Economic Studies 42.

Bawn, Kathleen and Frances Rosenbluth. 2006. “Short versus Long Coali-tions: Electoral Accountability and the Size of the Public Sector.” Amer-ican Journal of Political Science 50:251–265.

Birchfield, Vicki and Markus M. L. Crepaz. 1998. “The Impact of Constitu-tional Structures and Collective and Competitive Veto Points on IncomeInequality in Industrialized Democracies.” European Journal of PoliticalResearch 34:175–200.

Blanchard, Olivier. 2006. “European Unemployment: The Evolution of Factsand Ideas.” Economic Policy 21 (45):7–59.

Beland, Daniel and Andre Lecours. 2005. “Nationalism, Public Policy, andInstitutional Development: Social Security in Belgium.” Journal of PublicPolicy 25 (2):265–285.

Bonoli, Giuliano. 2000. The Politics of Pension Reform. Cambridge: Cam-bridge University Press.

Cameron, David. 1978. “The Expansion of the Public Economy: A Com-parative Analysis.” American Political Science Review 72 (4):1243–1261.

Cantillon, Bea, Veerle de Maesschalck, Stijn Rottiers, and Gerlinde Verbist.2006. “Social Redistribution in Federalised Belgium.” West EuropeanPolitics 29 (5):1034–1056.

Clasen, Jochen. 1992. “Unemployment Insurance in Two Countries: a Com-parative Analysis of Great Britain and West Germany in the 1980s.” Jour-nal of European Social Policy 2 (4):279–300.

———. 2005. Reforming European Welfare States. Oxford: Oxford Univer-sity Press.

Clasen, Jochen and Daniel Clegg. 2006. “Beyond Activation.” EuropeanSocieties 8 (4):527–553.

Commission on Social Welfare. 1986. Report of the Commission on SocialWelfare. Dublin: Stationery Office.

Cox, Gary W. and Mathew D. McCubbins. 2001. “The Institutional Deter-minants of Economic Policy Outcomes.” In Presidents, Parliaments, andPolicy. Cambridge: Cambridge University Press, 21–63.

Crepaz, Markus M. L. 1996. “Constitutional Structures and Regime Perfor-mance in 18 Industrialized Democracies.” European Journal of PoliticalResearch 29:87–104.

de Deken, Johan. 2007. “Distribution of Responsibility for Social Securityand Labour Market Policy. Country Report: Belgium.” AIAS workingpaper 2007-53, Amsterdam: University of Amsterdam.

18

Page 19: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

Department of Social Welfare. 1985. Report 1983 and 1984. Dublin: Sta-tionery Office.

———. 1989. Report 1987 and 1988. Dublin: Stationery Office.Dixit, Avinash K. 1996. The Making of Economic Policy. Cambridge: MIT

Press.Eichhorst, Werner, Otto Kaufmann, Regina Konle-Seidl, and Hans-Joachim

Reinhard. 2008. “Bringing the Jobless into Work?” In Bringing theJobless into Work?, edited by Werner Eichhorst, Otto Kaufmann, andRegina Konle-Seidl. Berlin: Springer, 1–16.

Faniel, Jean. 2005. “Reactions syndicales et associatives face au ‘controlede la disponibilite des cholmeurs’.” Annee sociale 2004 :133–148.

Finer, Herman. 1935. The Case Against Proportional Representation. Lon-don: Fabian Society, second ed.

———. 1961. The Theory and Practice of Modern Government. London:Methuen & Co.

FitzGerald, Eithne. 2005. “Employment Policy.” In Contemporary IrishSocial Policy, edited by Suzanne Quin, Patricia Kennedy, Anne Matthews,and Gabriel Kiely. Dublin: University College Dublin Press, second ed.,115–137.

Fondazione Rodolfo Debenedetti. 2010. “The fRDB-IZA Social ReformsDatabase.” Milan: Fondazione Rodolfo Debenedetti (http://www.frdb.org/scheda.php?id=1&doc_pk=9027).

Franzese, Jr., Robert J. 1998. “The Political Economy of Public Debt: AnEmpirical Examination of the OECD Postwar Experience.” Paper for the1998 Wallis Conference on Political Economy, Northwestern University, 7November.

———. 2002. Macroeconomic Policies of Developed Democracies. Cam-bridge: Cambridge University Press.

Fredriksson, Peter and Bertil Holmlund. 2006. “Improving Incentives inUnemployment Insurance.” Journal of Economic Surveys 20 (3):357–386.

Hays, Jude C., Sean D. Ehrlich, and Clint Peinhardt. 2005. “GovernmentSpending and Public Support for Trade in the OECD.” InternationalOrganization 59 (2):473–494.

Hemerijck, Anton C. and Martin Schludi. 2000. “Sequences of Policy Fail-ures and Effective Policy Responses.” In Welfare and Work in the OpenEconomy, Vol. I, edited by Fritz W. Scharpf and Vivien A. Schmidt. Ox-ford: Oxford University Press, 125–228.

Hemerijck, Anton C., Birgitte Unger, and Jelle Visser. 2000. “How SmallCountries Negotiate Change.” In Welfare and Work in the Open Economy,Vol. II, edited by Fritz W. Scharpf and Vivien A. Schmidt. Oxford: OxfordUniversity Press, 175–263.

Hemerijck, Anton C. and Mark I. Vail. 2005. “The Forgotten Center. StateActivism and Corporatist Adjustment.” In The State After Statism,edited by Jonah D. Levy. Cambridge: Harvard University Press.

Hermens, Ferdinand A. 1941. Democracy or Anarchy? A Study of Propor-tional Representation. Notre Dame: University of Notre Dame.

———. 1972. Democracy or Anarchy? A Study of Proportional Represen-tation. New York: Johnson Reprint, second ed.

19

Page 20: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

Hicks, John. 1939. “The Foundations of Welfare Economics.” EconomicJournal 49 (196):696–712.

Huber, Evelyne, Charles Ragin, John D. Stephens, David Brady, and JasonBeckfield. 2004. “Comparative Welfare States Data Set.” NorthwesternUniversity, University of North Carolina, Duke University and IndianaUniversity.

Hausermann, Silja. 2010. The Politics of Welfare State Reform in Conti-nental Europe. Cambridge: Cambridge University Press.

Hausermann, Silja and Bruno Palier. 2008. “The Politics of Employment-Friendly Welfare Reforms in Post-Industrial Economies.” Socio-EconomicReview 6 (3):559–586.

Iversen, Torben and David Soskice. 2006. “Electoral Institutions and thePolitics of Coalitions.” American Political Science Review 100 (2):165–181.

Katzenstein, Peter J. 1985. Small States in World Markets. Ithaca: CornellUniversity Press.

Kenworthy, Lane. 2008. Jobs With Equality. Oxford: Oxford UniversityPress.

Larsen, Christian Albrekt and Jørgen Goul Andersen. 2009. “How New Eco-nomic Ideas Changed the Danish Welfare State.” Governance 22 (2):239–261.

Laver, Michael. 1992. “Are Irish Parties Peculiar?” In The Development ofIndustrial Society in Ireland, edited by John Goldthorpe and ChristopherWhelan. Oxford: Oxford University Press, 359–381.

Layard, Richard, Stephen Nickell, and Richard Jackman. 1991. Unemploy-ment. Oxford: Oxford University Press.

———. 1994. The Unemployment Crisis. Oxford: Oxford University Press.Lijphart, Arend. 1999. Patterns of Democracy. New Haven: Yale University

Press.Lindvall, Johannes. 2010a. Mass Unemployment and the State. Oxford:

Oxford University Press.———. 2010b. “Power Sharing and Reform Capacity.” Journal of Theoret-

ical Politics 22 (3):1–18.Mair, Peter. 1992. “Explaining the Absence of Class Politics in Ireland.”

In The Development of Industrial Society in Ireland, edited by JohnGoldthorpe and Christopher Whelan. Oxford: Oxford University Press,383–410.

Martin, Cathie Jo. 2004. “Reinventing Welfare Regimes.” World Politics57 (1):39–69.

McCashin, Anthony. 2004. Social Security in Ireland. Dublin: Gill andMacmillan.

Nickell, Stephen and Richard Layard. 1999. “Labor Market Institutions andEconomic Performance.” In Handbook of Labor Economics. New York:Elsevier, 3030–3084.

OECD. 2007. Economic Survey: Belgium. Paris: OECD.Persson, Torsten and Guido Tabellini. 2006. “Electoral Systems and Eco-

nomic Policy.” In The Oxford Handbook of Political Economy, edited byBarry R. Weingast and Donald A. Wittman. Oxford: Oxford University

20

Page 21: THE REFORM CAPACITY OF COALITION GOVERNMENTS · PDF fileItalian coalition governments ... Recent work in political economy also suggests that electoral systems have ... consider Giovanni

Press.Pontusson, Jonas. 2005. Inequality and Prosperity. Ithaca: Cornell Univer-

sity Press.Powell, G. Bingham. 2000. Elections as Instruments of Democracy. New

Haven: Yale University Press.Rhodes, Martin. 2000. “Restructuring the British Welfare State.” In Welfare

and Work in the Open Economy, Vol. II, edited by Fritz W. Scharpf andVivien A. Schmidt. Oxford: Oxford University Press, 19–68.

Riker, William H. 1984. “Duverger’s Law Revisited.” In Electoral Lawsand Their Political Consequences, edited by Bernard Grofman and ArendLijphart. New York: Agathon Press, 19–42.

Rodrik, Dani. 1998. “Why Do More Open Economies Have Bigger Govern-ments?” Journal of Political Economy 106 (5):997–1032.

Roller, Edeltraud. 2005. The Performance of Democracies. Oxford: OxfordUniversity Press.

Samuelson, Paul A. 1962. “The Gains from International Trade OnceAgain.” Economic Journal 72:820–829.

Sartori, Giovanni. 1997. Comparative Constitutional Engineering. Hound-mills: Palgrave, second ed.

Scharpf, Fritz W. and Vivien A. Schmidt, editors. 2000. Welfare and Workin the Open Economy. Oxford: Oxford University Press. 2 Vols.

Scruggs, Lyle. 2004. “Welfare State Entitlements Data Set: A ComparativeInstitutional Analysis of Eighteen Welfare States.” Version 1.1.

Stolper, Wolfgang F. and Paul A. Samuelson. 1941. “Protection and RealWages.” The Review of Economic Studies 9 (1):58–73.

Swank, Duane. 2006. “Politics Data Set.” Marquette University.Tansey, Paul. 1991. Making the Irish Labour Market Work. Dublin: Gill

and Macmillan.Torfing, Jacob. 1999. “Workfare With Welfare.” Journal of European Social

Policy 9 (1):5–28.Tsebelis, George. 1999. “Veto Players and Law Production in Parliamentary

Democracies.” American Political Science Review 93:591–608.———. 2000. “Veto Players and Institutional Analysis.” Governance

13 (4):441–474.———. 2002. Veto Players. Princeton: Princeton University Press.Van der Linden, Bruno, editor. 1997. Chomage. Reduire la fracture. Paris:

De Boeck Universite.Viebrock, Elke and Jochen Clasen. 2009. “Flexicurity and Welfare Reform:

A Review.” Socio-Economic Review 7 (2):305–331.Weishaupt, Timo. 2008. The Emergence of a New Labor Market Policy

Paradigm. Ph.D. thesis, University of Wisconsin-Madison.

21