ces open forum 20: eurozone crisis and social models: what we can learn from italy and spain

8/12/2019 CES Open Forum 20: Eurozone Crisis and Social Models: What We Can Learn from Italy and Spain

1/58

ces papers - open forum

2013-2014

Eurozone crisis and social models:

what we can learn from

italy and spain

Author: sofia perez


2/58

CES-

ces papers - open forum# 20

Open Forum CES Paper Series

The Series is designed to present work in progress by current and former Center afliates andpapers presented at Centers seminars and conferences. Any opinions expressed in the papers

are those of the authors, and not of CES.

Editors:

Grzegorz Ekiert and Andrew Martin

Editorial Board:

Philippe AghionPeter Hall

Roberto Foa

Alison FrankTorben IversonMaya Jasanoff

Jytte KlausenMichele Lamont

Mary LewisMichael RosenVivien Schmidt

Kathleen ThelenDaniel Ziblatt

Kathrin Zippel


3/58

ABSTRACT

ces papers - open forum # 20

This paper considers the role of social model features in the economic performance of Italy and Spain during

the run-up to the Eurozone crisis, as well as the consequences of that crisis, in turn, for the two countries social

models. It takes issue with the prevailing view - what I refer to as the competitiveness thesis - which attri-

butes the debtor status of the two countries to a lack of competitive capacity rooted in social model features.

This competitiveness thesis has been key in justifying the liberalization plus austerity measures that European

institutions have demanded in return for nancial support for Italy and Spain at critical points during the crisis.

The paper challenges this prevailing wisdom. First, it reviews the characteristics of the Italian and Spanish

social models and their evolution in the period prior to the crisis, revealing a far more complex, dynamic and

differentiated picture than is given in the political economy literature. Second, the paper considers various ways

in which social model characteristics are said to have contributed to the Eurozone crisis, nding such explana -

tions wanting. Italy and Spains debtor status was primarily the result of much broader dynamics in the Euro -

zone, including capital ows from richer to poorer countries that affected economic demand, with social model

features playing, at most, an ancillary role. More aggressive reforms responding to EU demands in Spain may

have increased the long term social and economic costs of the crisis, whereas the political stalemate that slowed

such reforms in Italy may have paradoxically mitigated these costs. The comparison of the two countries thus

suggests that, in the absence of broader macro-institutional reform of the Eurozone, compliance with EU dic -

tates may have had perverse effects.

Keywords:

Eurozone, Eurozone crisis, social models, Spain, Italy, labor market, welfare state, internal devaluation.

author:

sofia perez

biography:

Soa Perez is a political scientist at Boston University. Her work centers on the politics of regulatory change and the connections

of institutional change across policy areas, including monetary policy, nancial regulation, wage setting, labor market, welfare and

migration policy in Europe. Her work has appeared in Comparative Political Studies, Comparative Politics, International Studies

Quarterly and the Journal of European Public Policy among other journals.


4/58ces papers - open forum# 20

Eurozone crisis and social

models: what we can

learnfrom italy and spaini-

taly and

Explanations of the ongoing crisis of the Eurozonetend to fall into two not necessarily mutually ex-clusive categories: 1) those that emphasize theshortfalls of the currency unions macro-institutionalarchitecture (monetary union without scal or bank-ing union), and 2) those that emphasize structuralcharacteristics of the Eurozones debtor countries

as the source of the intra-Eurozone nancial imbal-ances that became the focus of bond markets during

the sovereign debt crisis. It is the latter view whatFitoussi and Saraceno (2013) have referred to as theBerlin -Washington consensus and has later cometo be known as the Brussels-Frankfurt consensus -that carried the day in the Eurozone and lies behindthe pro-cyclical austerity cum structural reformsformula that has been imposed on debtor countries bytheir European institutions (the Commission and, for

a time, the ECB) and other Eurozone governments(the Eurogroup). Chief among the demands of the

EU - carried out by way of both formal and informalmodalities of conditionality attached to critical -nancial support during the debt crisis - were budget-ary cuts that have fallen heavily on social protectionprograms along with changes in labor market regu-lation intended to facilitate internal devaluation viadownward adjustments in labor costs. Even whenthe perverse effects for growth of this approach of

labor market liberalization coupled with pro-cyclical

scal measures have been recognized by importantactors (the IMF and, less openly, the Commission,[see Bontout and Lokajickova 2013; EuropeanComission 2012b]) warnings against any slow downin scal consolidation and calls for further structuralreforms in debtor countries have remained the order

of the day in intergovernmental negotiations withinthe Eurogroup. Meanwhile, changes to the macro-institutional architecture of the Eurozone that could

limit the damage inicted on debtor states (such as,for instance, the issuance of mutualized debt instru-ments to end the risk of sovereign defaults and re-denomination) have remained blocked by intergov-ernmental impasse or put on timelines that precludereal relief in the foreseeable future (as is the casewith key elements of banking union). All this re-

ects the conviction (central to the Brussels-Frank-furt consensus) that the situation in which the debtorcountries nd themselves is in some way or other toblame on excesses in their social models (Dlvikand Martin, forthcoming), that is to say their systemsof labor market regulation and social protection

The experiences of Italy and Spain are ofparticular relevance to this debate surrounding thesources of the Eurozone crisis and the repercussions

of the austerity centered approach adopted by the

EU. With the third and fourth largest economiesin the currency area, the two countries are widelyseen as determinant for the future of the Eurozone

at large. At the same time, Italy and Spain havetraditionally been identied in the welfare state andVarieties of Capitalism (VOC) literatures as repre-senting a Southern European model of capital-ism characterized by social policy and labor market

features that have pernicious effects: low levels oflabor market participation (in particular by women)dualism between long time labor market insidersand outsiders and collective bargaining institutionsthat undermine competitiveness. From this charac-

terization it is an easy step to the conclusion thatsocial model characteristics must have been behindthe situation the two countries found themselves infrom 2010 onwards when the Eurozone sovereigndebt crisis spread beyond the smaller peripheralstates in the currency area. Indeed, the character-

ization of these countries in much of the VOC lit-erature appears to vindicate the Brussels-Frankfurtconsensus on the roots of the Eurozone crisis

A closer look at the evolution of welfarestate and labor market institutions in the two coun-tries both prior to and following the sovereign debtcrisis, however, sheds a far more complex picture


5/58ces papers - open forum# 20

one that reveals important differences in their socio-economic and political dynamics as well as the ex-tent to which social model change and its econom-ic effects in each country have depended on wider

macro-economic developments in the Eurozone.

This paper takes up the question of whether

Italy and Spains fate in the Eurozone crisis can beattributed to the social model features that have beenthe focus of current reforms. It argues that there are

multiple ways in which the two countries experi-ences challenge the underlying assumptions and ad-equacy of the course that they have been encouraged,when not forced, to follow by their Euro partners viaboth formal and informal conditionality. In the sec-tions that follow, I rst describe the evolution of the

Italian and Spanish social models in the period priorto the crisis, pointing to different reform patterns thatwere strongly affected by the different political andmacro-economic constraints that governments faced

at various points. Prior to the crisis, the domestic pol-itics of institutional reform had moved the Spanishsocial model considerably closer to that of its North-ern EU neighbors than was the case for Italy. This,however, did little to prevent the economic boomand bust that Spain experienced following monetary

union. While Italy experienced less social modelchange and made less headway on such critical mea-sures as labor force activation in the pre-crisis pe-riod, this can be attributed, to a signicant extent,to the radically different macro-economic conditionswith which it joined the Eurozone. Governments of

the Left in Italy failed repeatedly in their attempts toshift the balance of social spending in a more activa-tion friendly direction largely because of these mac-ro-economic conditions. The lack of attention paid tothe relationship between macro-economic constraintand the politics of social model reform is one of thestriking short-comings of the analysis that lies behindthe Brussels-Frankfurt consensus, which has servedto justify the de facto imposition of reforms in areas

well beyond the remit of EU institutions, contributingto economic hardship and raising fundamental ques-tions of democratic accountability in both countries.

The second section of the paper focuses on

the contrasting economic performances of Italy andSpain in the run-up to the Eurozone crisis. It con-

siders various arguments that have been offered for

how social model features may have contributed tothe two countries situation at the time of the cri-

sis. While social model features and developmentscan explain some of the contrast in the labor mar-ket performance of the two countries during this

period, there is little to suggest that they played adeterminant role in producing the much touted lossof competitiveness the two countries suffered during

the rst eight years of monetary union or their vul-nerability to contagion as the sovereign debt crisisdeveloped. Developments within the Eurozone at

large involving a rise in household savings in someof the richer countries (Germany in particular) andthe intensication of nancial ows within the cur-rency zone from from richer to poorer countries

over the pre-crisis period (Lane and Pels 2012; Pet-tis 2012) appear to have played a more fundamentalrole, along with other country characteristics that donot form part of what we can call their social models

The third and last section of the paper reviews

the responses of the Italian and Spanish governmentsto the Eurozone crisis and the impact this has had on

the two countries social models. While only Spainwas forced to sign a memorandum of understanding

with the troika in order to obtain a credit line for therescue of its banking sector in 2011, reforms in bothcountries were heavily determined by what Sacchiet al. (2010) have referred to as informal but strongconditionality imposed by the Eurogroup in coor-dination with the ECB starting in 2010. With an al-ready more market-oriented social protection systemand the advent of a center-right government with an

absolute majority in 2011, Spain moved considerablyfaster and further in the direction of labor market de-regulation advanced by the troika than Italy, whereelectoral results forced a slower pace of reform and adifferent trade-off between productivity growth and

employment. Indeed, Italian governments and social


6/58ces papers - open forum#

actors up to 2012 were able to limit the degree ofjob destruction by building on existing non-marketmechanisms of employment-sharing without incur-ring any substantially higher bond market penaltythan Spain, even though Spains internal devaluationand competitiveness measures improved consider-

ably and Italys continued to deteriorate. These

facts - due largely to the negative feedback effect ofemployment and household income insecurity on do-mestic demand and lending in Spain present a realchallenge for the assumptions underpinning the Brus-sels-Frankfurt consensus. They raise the possibilitythat Spains faster and more decisive embrace of the

troikas prescriptions in the midst of an economic

downturn and in the absence of substantial macro-institutional reform of the Eurozone may in the endbear little advantage. Indeed, they illustrate the dan-ger that liberalizing labor market reform imposed inthe context of pro-cyclical spending cuts and withoutdecisive change in the Eurozones institutions may

only lead countries to fall into a bad economicequilibrium (de Grauwe 2012; Grauwe and Ji 2012;Wyplosz 2012; 2013) and impair growth in the longterm by disregarding the relationship between pres-ent and future output (Fitoussi and Saraceno 2013).

I.The Spanish and Italian Social Models: evolu-tion in the pre-crisis period

The social models of the countries of South-ern Europe have often deed easy categorization.Nonetheless, following the work of Leibfried (1992)and Ferrera (1996), who emphasized the limited de-velopment of social policy, the male breadwinnerbias in social spending and labor market regulationand the central role of the family as a safety net, Italy

and Spain have commonly been described in the Va-rieties of Capitalism (VOC) literature as sharing adistinct South European, or familial model ofwelfare capitalism (see Rhodes 1996; Bonoli 1997:Triletti 1999; Crouch 1999; Naldini 2003; Amable2004). This view of the two countries, however, wasbased fundamentally on snap-shots taken in the late1980s and early 1990s when their welfare state and

centrated in old age pensions and little if any incomesupport going to people who did not have extensivecontribution records (younger population cohortsworking age families, and women in general). Neither country had an adequate system of minimumincome support for those of working age and this

was reected in high levels of dependence upon the

family as a safety net, low female labor force participation rates and fertility, and higher than averagerates of poverty risk. Most importantly, the logic othis familial model of welfare provision was seento be self-reinforcing as high levels of employmenprotection and old age pensions served to compensate the mostly male breadwinners of earlier decades ensconced as union insiders for the lackof other forms of income support and security fo

themselves and their families. According to someauthors, this combination of welfare state featureswas also believed to undergird a tendency towardlow-wage production strategies and/or regional divisions and a rm structure skewed towards overlysmall rms, thereby locking these countries into aself-perpetuating Southern syndrome of low employment rates and productivity growth (Amable2005; Rhodes 1996; Hancke and Herrmann 2007)

The perception that the two countries represented a distinct variety of capitalism with itparticular logic and set of deleterious consequenceswas also reinforced by work on their industrial relations systems, characterized by the coexistenceof competing labor confederations, a complicatedcombination of sometimes overlapping levels owage bargaining (in particular in Italy) (Regin1979, 1997) or a national employer confederationsthat, although highly encompassing, did not have

adequate leadership capacity vis vis its afliate(the case of Spain) (Perez 2000; Regini 1979; 1997Ebbinghaus and Visser 2000). An intermediatelycentralized bargaining structure combined with highlevels of employment protection and divided labororganizations were believed to undermine efforts todeliver coordinated wage moderation (Regini 1984Crouch 1985). And although both countries shared

20



a substantial history of nationally negotiated incomespolicy - or concertation pacts designed to deliverwage restraint, by the late 1980s these efforts appearedto have broken down indenitely in Italy and be onthe verge of collapse in Spain (Negrelli ; Perez 2000).

This in a nutshell, was the scenario at the

start of the 1990s when efforts to characterize thetwo economies in relation to the VOC literaturesjuxtaposition of liberal and coordinated marketeconomies took place. The subsequent two de-cades saw signicant changes in the two countries.Nonetheless, the view that the Southern econo-mies were subject to a distinct historical logic so-cially rooted in their welfare and labor market in-stitutions has proven inuential. And even afterpointing out a much wider set of institutional fac-tors on which the two countries often differ, Am-

able (2005, p. 242) would go as far as to write thatthe high-skill-level job-security route seems mostlyout of reach for Mediterranean capitalism,102 anda competitive model based on exibility is muchmore in order. Implementing this model potentiallymeans meeting some considerable opposition frominsiders. Leaders of these countries would there-fore welcome any political support at the supra-

national level. It would be easier to overcome do-mestic opposition if the labour exibility measuresappeared to come from a European initiative

As is evident, such VOC interpretationsof Italy and Spains social models encourage pre-cisely the type of interventions and prescriptionsthat lie at the heart of the Berlin-Washington con-sensus and have guided the informal but strong(Sacchi 2013) conditionality that both countrieshave faced in the course of the Eurozone crisis. A

review of developments in the two decades prior tothe crisis, however, suggests that the conclusionsthat might be drawn from these rather static efforts

at classication fail to take into account the dynam-ics of social model change in the two countries.

By the mid 2000s some of the postulatedsimilarities between the two cases had waned andwhat was once thought to be a common set of self-perpetuating and path-dependent socio-economic

dynamics appeared more a matter of temporary co-

incidence. First, while from the standpoint of theearly 1990s, the social protection system of the two

countries might not have appeared different, thishad changed considerably. On one hand, between1996 and 2007 Italys total social expenditures rose(from 20 percent to 26.7 percent of GDP) right upto the EU15 average, while Spains (at only 21 per-cent of GDP in 2007) continued to place that coun-try at the very low end of the Eurozone (and closerto the highest spenders among the new EU mem-ber states from the CEE). But on the other handthe distribution of spending in Italy remained veryskewed towards old age pensions (see Figure 1 be-low), consistently around 51-52 per cent of all socialspending (compared with an EU15 average of 37-39per cent), while in Spain the overall weight of oldage pensions in social spending had dropped from41 percent in 1996 to 38 percent in 2007. This wasthe result of a far more successful process of pen-sion bargains with unions (which had lowered pen-sion spending to 9% in 2007, well below the EU15

average) and a fairly consistent effort by Spanishgovernments to shift spending to benets support-ing those of working age workers through family,labour market, health care and sickness policies

Figure 1. Social expenditures as a percentage of GDP

Source: OECD.StatExtract iLibrary (downloaded May 4, 2012)

0

5

10

15

20

25

30

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Italy total

Italy old ag

Spain tota

Spain old a

20



minimum non-contributory pensions for those with-

out signicant work histories, income maintenanceof last resort for those of working age (above the ex-tremely rudimentary levels of local social assistancethat existed at the end of the 1980s), universal freechildcare for children from the age of 3 introduced in1990 (and a plan to extend this to children from age

0 introduced in 2007 that later had to be cancelledas a result of austerity measures), the extension ofunemployment insurance to workers on temporarycontracts (a feature that clearly distinguished Spainfrom Italy) and to the self-employed in 2007, the in-troduction of an emancipation subsidy to facilitatethe ability of young people between 22-30 years ofage to afford rental housing, and the so-called de-pendency law of 2006 that requires regions to allo-cate part of their social spending budgets to providenon-means tested, in-home care for the elderly andthe disabled. This offers a striking contrast to Italysrepeated failure to establish even the most rudimen-tary piece of a more developed welfare state - a na-tional income of last resort scheme - and, until theFornero reform of 2012, to expand the availabilityof unemployment insurance beyond those on inde-nite work contracts with long contribution periods.In spite of an increasingly apparent public awareness

of the biases of the Italian welfare state and its con-sequences, little progress had been made by the endof the 2000s in these areas, or in others such as fam-ily policy and active labor force activation policies.

Alongside the Spanish shift in the structureof social spending, the other key point of departurebetween the two countries involved the regulation ofemployment. Spain initiated a far earlier and radicalliberalization of employment through the legaliza-

tion of xed-term contracts as early as 1984 . WhileItaly would follow course in stepwise fashion in thelate 1990s and early 2000s (see below), the Spanishdualization of its labor market went well beyond thatof any other EU15 state, with temporary employmentcontracts accounting for over 30 percent of total fromthe early 1990s on. At the same time, however, Spainalso gradually expanded its unemployment protection

Thus, while overall social spending in Spain remainedparticularly low, the structure of that spending hadmoved much closer into line with that of the rest of theEU 15 (see also Moreno ; Guillen 2011 for discussion).

From 1996 to 2007, Spanish governmentsincreased their spending on sickness and healthcare

from 29 to 31 per cent, social exclusion from 0.8to 1.3 per cent (compared to Italys minimal movefrom 0.1-0.2 per cent), and benets for families andchildren from 2.3 per in 1996 (then just a quarterof the EU15 average) to 6 per cent in 2007, whileItaly increased its spending in this area from just3.5 to 4.7 per cent. Between 2004 and 2008 Spainalso spent twice as much as Italy on labour marketpolicy, increasing that expenditure as a proportionof GDP from 2.1 per cent in 2004 to 2.5 percent in2008, while Italys fell from 1.3 to 1.2 per cent ofGDP over the same period. By 2007-2008 Spain wasspending as much of its GDP on labour market poli-cies as the EUs leaders in this area (the Netherlandsand Denmark). Nonetheless, the greater part of thatspending went to income maintenance benets rath-er than active labour market policy. Indeed, most ofthe difference between Italy and Spains labor mar-ket expenditure was due to Spains more developed

non-contributory minimum pension and unemploy-ment benet system which protected all categories ofworkers, including workers on xed-term contracts.Although overall coverage in 2008 was still on thelow side (47 per cent, compared with 80-90 per centin Denmark and Austria), this compared well withItalys 17 per cent. Moreover, the gross earnings re-placement rate during the rst year of unemployment(though not thereafter) and the overall duration of ben-ets in Spain were relatively high by EU standards.

The difference in the evolution of the Italianand Spanish social model was not just one of quanti-ties. In what has been called the Northern turn of theSpanish model, social model developments in Spainalso included qualitative changes introduced by gov-ernments of both the Left and the Right (Moreno andSarasa 2008; Guillen 2010; Guillen and Leon 2011):

20



system from the early 1990s, turning it into a univer-sal system covering all types of workers. By contrast,in Italy where unemployment protection requiredparticularly long contribution periods it remainedrestricted to workers with indenite contracts andextensive work histories, and drops in labor demandwere also addressed through insurance and work-

time sharing schemes (Jessoula 2011; Sacchi 2013),creating a stronger bias in favor of old timers thanin Spain. In this regard, the regulation of the Span-ish labor market moved several degrees closer to thatof its Northern European neighbors in providing ameasure of exicurity that did not emerge in Italy.

The large share of xed-term employmentbecame an enduring and contentious feature of

Spains political economy from the late 1990s on.It has been cast as the cause of Spains high struc-

tural unemployment and a potential source of poorlabor productivity growth (Jimeno and Toharia 1993;Bentolila and Dolado 1994) and is blamed on a par-ticularly intense defense by labor market insiders(represented by the unions) in the form of high dis-missal costs for those on indenite contracts at theexpense of outsiders (Bentolila, Jimeno and Dolado2012). What is less often emphasized is the extent to

which this liberalization of the Spanish labor mar-ket at the margins through the liberalization of tem-porary contracts imbued the Spanish economy with

a type of hyper-exibility (albeit a segmented one)that has few counterparts in the EU15. The effectsof this particular mode of labour market exibility ontotal employment remain ambiguous. We do know,however, that it allowed for intense bouts of rapidemployment creation and employment shedding inresponse to macro-economic stimuli following mon-

etary union. Although Italys labour market was alsomarked by growing segmentation in employmentprotection combined with an utter absence of any

universal income of last support mechanisms (whatBerton et al. (2009) have referred to as a regime ofex-insecurity), the existence of non-market mecha-nisms for employment sharing among labor marketinsiders (at least until the Fornero labor market re-

form of 2012) appear to have helped stave off the de-gree of employment uctuation we observe in Spain

Lastly, in the area of labour relations, the roleof collective bargaining was signicantly strength-ened during the run up to EMU both through socialpacts and unilateral government action. In Italy, the

abolition of automatic wage indexation (scala mo-bile) agreed to by the main labor confederations in1992 helped the country overcome the economic andpolitical debacle of that year, when the existing partysystem collapsed and the lira was force out of theEuropean Exchange Rate Mechanism (ERM) (dellaSala 2004; Culpepper 2008). In Spain, the main cata-lyst for change was a 1994 government imposed abo-lition of the system of labor ordinances dating backto the Franco era (Perez 2000b). Though the inten-tion of that reform had been to liberalize the regula-tion of work conditions, its effect was that of rais-

ing the prole of collective bargaining as it forcedunions and employers to bargain on a wider rangeof issues than wage setting. Both countries also sawefforts to create a more articulated bargaining struc-ture in which national sectoral bargains took pre-cedence in setting wage minima, while productiv-ity premiums and other matters were left to lower

bargaining levels (Regini and Regalia 1997; Perez2000; 2010). This effort, turned out to be more de-cisive in Italy than in Spain where the national em-ployer confederation could not bring along its lo-cal afliates to accept national sectoral bargains inmany sectors. Coordination in wage bargaining to

maintain competitiveness, of which there is evidence

from 1994 on, was thus carried out principally bythe labor confederations until formal overarchingwage pacts with employers were renewed in 2001

On the other hand, in Italy the concertation processre-initiated in 1992 began to collapse in 2001 (justas Spains was being re-formalized) under pressurefrom the second Berlusconi government (Negrelli2011). Indeed, social pact making and concertationin Spain lasted much longer (well into the economiccrisis) than in Italy, leading some observers to speakof a virtual institutionalization of the process in thatcountry prior to the crisis (Natali and Pochet 2010)

20



The following section explores the combina-tion of political, institutional and economic condi-tions that made for disparate paths of social modelchange in the two countries, before we turn to their

role in the economic crisis.

II. The Evolution of the Spanish and Italian So-cial Models up to the Eurozone crisis

Welfare state and Varieties of Capitalismscholars in the past have posited that Italy and Spainare subject to self-reinforcing historical dynamics be-tween welfare state, labor regulation and productionregime features (Amable 2003 ; Molina and Rhodes2007; Hancke and Herrmann 2007). These analy-ses provide potentially important insights into thesources of observed continuities in the two countries.

Nonetheless, the notion that the social models of Ita-ly and Spain reect a similar long-range dynamic ne-glects crucial differences in the political contexts thatshaped social and employment policies. These differ-ences are important in understanding the divergent

ways in which social protection and labor market reg-ulation evolved in the period leading up to the crisis.

The social model that developed in postwarItaly was the result of a process of conict and ac-commodation between center-right governments

dominated by the Christian Democratic party, the

political Left (in particular the Communist Party ofItaly) and the main labor confederations (the Com-munist dominated CGIL, the Christian UnionistCISL, and the secular but anti-Communist UIL) inthe course of the Hot Autumn of 1969 and its after-math. This accommodation included the provision

of relatively generous old age pension benets, theempowerment of the national confederal unions atthe rm level, the participation of the confederationsin the control of contributory pensions, and the es-tablishment of a mechanism of automatic wage in-dexation (the scala mobile). By the time of the crisisof the so called First Republic due to the tangen-topoli scandals in the period 1991/92, this social

model had produced very signicant wage com-pression across regions with quite different charac-teristics (North and South) and a large public debtburden derived from the old age pension system

In Spain, by contrast, both the structure of

wage bargaining and the introduction of a PAYG sys-

tem, as well as some aspects of employment protec-tion, were all shaped in the context of the authoritar-ian right-wing dictatorship of General Franco. Thetransition to a liberal parliamentary regime occurredin the context of an acute economic crisis that went

some way towards containing demands for redistri-

bution by the political Left. The process of buildinga democratic social model was thus barely into itssecond decade at the start of the 1990s and this is alsothe way in which it was viewed by domestic actors. In

addition, the late transition to democracy also meantthat Spain enjoyed some of the advantages of a late-comer to democratic welfare capitalism. When pen-sion reform, for instance, was undertaken in the mid

1990s under the leadership of the rst post-transitionconservative government, Spanish unions had fewer

incentives to block that reform than Italian unions didin the 1990s because they were not as heavily domi-nated by older pensioners as was the case in Italy

Constitutional and political developmentsalso took the two countries in different directionsduring the 1990s. In Spain the role of parliament wasstrengthened (Prez Daz 1999) while in Italy thecollapse of the postwar party system in 1990-1992produced a shift from consensus oriented politics toa more majoritarian democracy, the emergence of a

new, bi-polar party-system and a strengthening of theexecutive. Up until its great crisis, the post-war Italian

Republic had allowed for far greater proportionalitythan Spains post-Franco electoral law. The Span-ish electoral system, although a party list system,involves a strong bias in favor of the two largest par-ties in any given electoral circumscription, favoringthe two major national parties (the Social-democrat-ic PSOE (Partido Socialista Obrero Espaol), andCenter-right catch-all PP (Partido Popular)), and the

20



of its internal statutes, the CEOE exercises verylittle authority over its afliates, making it a par-ticularly weak organization in comparison to othernational employer associations with lower member-ship rates (Costas and Nonell 1996; Nonell et al2006). While Italys Conndustria was affected byimportant tensions between its small and large en-

terprise factions, the ultimate dominance of the lat-ter has generally given it greater capacity of action

On the labor side, the Italian confedera-tions enjoyed a (still relatively high) 34 per centunion density rate in 2007 (down from 38 per centin 1990), and this membership was based largelyin manufacturing industry, providing the nationalconfederations a great deal of leverage both at therm and national policy levels. Yet while the Span-ish labor confederations had a low membership rate(16 percent), they drew their strength far more fromworkplace elections (in which about 80 per cent ofworkers take part) and from the legal frameworkwhich provides that agreements signed above the

rm level by any union meeting the criteria of mostrepresentativeness (based on work place elections)automatically applies to all workers at the sectoral/territorial level at which it is reached and that agree-

ments remained in effect until renewed (the principleof ultra-activity), at least until the 2012 labor marketreform. Thus, while unions did not have strong rep-resentation within rms (with the exception of verylarge rms and the public sector), their representa-tiveness (as measured by workplace elections) andtheir negotiating capacity far outstripped their for-

mal membership levels. This rendered them capablesocial partners. On the other hand, while Spanishunions did not have the role, played by their Italian

counterparts in co-managing the main public pil-lar pensions system , as well as the more recentlycreated closed second-pillar voluntary occupa-tional pension funds, reforms in the early1990s didstrengthened their role in managing the public em-ployment service, the continuing vocational trainingsystem and health and safety at work (Guillen 2010)

mainstream regional nationalist parties where theseplayed a central role (Catalonia, the Basque Country,Galicia and the Canary Islands). However, in 1993the Italian electoral system was changed to a predom-inantly SMDP system, producing the coalescence ofthe countrys multiple political parties into left andright-wing blocs. While Spain mostly had minority

governments by one of the national parties support-ed by the mainstream regionalist parties (which es-poused positions ranging from Christian Democratic

to market-liberalism) in the 1990s and 2000s, Italysaw a steady alternation between the Right-wing co-alition headed by Silvio Berlusconi and the re-com-bined center-Left. Spain thus arguably had a politi-cal scenario more propitious to consensus orientedsocial reforms than did Italy in the post 1994 period.

In addition to these very different politicalcontexts, there were also key differences in the socialmodel institutions of Italy and Spain that have tendedto be overlooked by comparative VOC and welfarestate scholars and that, when taken together, gaveSpanish social actors and governments a greater ca-pacity to produce policy coalitions in favor of socialmodel reform than their Italian counterparts duringthe 1990s and 2000s. These differences included 1)

the cohesiveness of union versus employer organiza-tions, 2) the administrative role of unions in criticalareas of the welfare state, 3) the legal framework ofcollective bargaining, which gave unions in Spainfar more inuence than implied by membershiprates, and 4) the extent to which the wage bargain-ing system caused cross-regional wage compression.

The organization of employers, and in par-ticular the degree of encompassingness of em-

ployer associations, has by now been widely rec-ognized as a fundamental point of difference in thepolitical economies of European states (Soskice1990; Hall and Soskice 2001; Martin and Swank2011). The Spanish labor confederation (CEOE)has a particularly high membership rate at 72 per-cent, and the Italian employers association (Conn-dustria) a respectable 51 per cent. However, because

20



Lastly, while both countries had what arecommonly labeled intermediately centralized andcoordinated wage bargaining systems, the Italiansystem produced a far greater degree of wage com-

pression across sectors, regions, and categories of

workers rst as a result of the scala mobile, whichfrom 1977 on increased compensation across sec-tors by a common denominator for ination (Lockeand Baccaro 1999), and later by the important roleof national sectoral bargains. The Spanish bargain-ing system, in which more workers are covered by

bargains at the provincial-sectoral level than at thenational-sectoral level and in which ination revi-sion clauses are dependent on collective bargains,always allowed more wage dispersion across re-gions, sectors, and categories of workers. Coordi-

nation in wage bargaining, on the other hand, was

principally achieved in both countries through peak-level incomes policy guidelines that covered Italianwage setting since the early 1990s and wage bargainsin Spain from 1977 to 1988 and again after 2001.

The different sources of power and incentives

faced by Italian and Spanish employers, unions, and

governments are likely to have had an important im-pact on the course of social model reform in the twocountries. One way in which this can be observed isin the relative success of social pacts aimed at pro-ducing social model change in the period prior to thecrisis. As the following sections describe, that pro-cess appeared far more successful in Spain than inItaly (where it repeatedly collapsed during periodsof conservative government in 1998 and again after2001), undergirding many of the changes in the Span-

ish social protection regime described above. Thepersistence of social bargaining in Spain prior to thecrisis bespeaks differences in the reform dynamics

of the two countries which suggest that the degree of

similarity attributed to the two cases in the 1990s mayhave been more the effect of a particular historicalvantage point than of persisting historical dynamics.

2.1 Italy: From grand pacts to reform stalemate

The run up to monetary union began in Italywith a major political crisis, the result of a series ofpolitical trials for corruption and abuse of power tha

affected a wide gamut of the Italian political classThe crisis climate was reinforced when the Italianlira was forced out of the European Exchange RateMechanism a few months after the 1992 general elec-tion. This prompted a period of social reform focusedon wage moderation and pension cost containment

with some, albeit limited, attention to labour mar-ket issues (contractual exibility and social shockabsorbers, i.e. income maintenance for the unem-

ployed) forged through social pacts between the ma-jor labor confederations and the technocratic govern-ments that oversaw the transition to a new electoralaw in 1993. The process of negotiated social reformcontinued up to the end of the 1990s (Locke andBaccaro 1999; Regini and Regalia 1997; della Sala2004), but it broke down once Italys participation inEMU was secured. Starting in the early 2000s, Ital-ian governments unilaterally introduced labor mar-ket reforms that increased the segmentation between

labor insiders and several categories of outsiderswith varying levels of entitlement to social protectionor none at all (see Breton et al. 2009; Sacchi 2013)

The dual economic and political crisis ofItaly in the early 1990s created a context in whichthe national union confederations emerged as keystrategic partners to the technocratic governments

of the period seeking to reestablish Italys credibil-ity (Regini and Regalia 1997; Negrelli 2011). Thecollapse of the dominant Christian Democratic Par-ty due to the tangentopoli scandals, the crisis ofthe Italian Communist Party, and the prospect thatEMU membership might elude Italy following theERM crisis all created a perception of emergencyamong unions and employers. This facilitated twoimportant tripartite agreements: the rst in 1992abolishing the scala mobile system of automatic

20



wage indexation; the second in 1993 creating a newand stable architecture for incomes policies and col-lective bargaining (Regini 1996: 726; Perez 2000;2002). The negotiation of the pacts also cementeda new period of inter-confederal union coopera-tion, which had been a critical feature of the earli-er period of political exchange from 1970 to 1984.

The abolition of the scala mobile representeda critical change in the Italian system of wage setting.Since the late 1970s, the mechanism had becomea driver of both ination and of wage compressionacross sectors and regions in Italy. Because the wageadjustments dictated by the scala were so signicant,they had come to dominate wage setting in the prior

decade and therefore had rendered collective wagebargaining largely meaningless (Baccaro and Locke1996). The subsequent 1993 agreement created anew, two-level architecture for wage setting -- withnational sectoral bargains (guided by overarchingcross-sectoral incomes policy guidelines) settingminima and so called proximity agreements at ei-ther the company or local level (depending on thesector) setting productivity-linked upgrades. Thisnew, more orderly bargaining structure allowed bothgreater wage coordination and a rise in wage differen-

tials. Wage dispersion, however, continued to be lim-ited by the fact that minimum wages continued to be

negotiated at the national central level by employersand unions (Hancke and Hermann 2007; Molina andRhodes 2009; Birindelli and dAloia 2008: 29-34).

Following the transformative pacts on wagebargaining of 1992/93, the technocratic and Center-Left governments of what came to be known as theSecond Italian Republic sought to turn what had been,

initially, a mechanism of emergency governance intoa routine mode of policy-making based on concer-tation with the unions. This model proved success-ful in allowing Italy to gain inclusion in EMU, but itrepeatedly broke down in the context of subsequentRight-wing governments. The rst major pension re-form (following the rst Berlusconi government un-successfully attempt to impose a unilateral reform)

was passed by the government of Lamberto Dini in1995 and was largely based on the advice of unionand Ministry of Labor experts. Although its impacton costs and liabilities was modest, it did bring abouta shift from an earnings-based to a more contribution-

related PAYG system, standardized contributionsacross categories of employees (private, public and

self-employed) and incentives for the use of fully-funded supplementary pension schemes. Tripartitenegotiations also produced a number of labor reformmeasures that sought to address the unprecedentedlyhigh level of unemployment observed in the South(at 26 percent in ) and among youth (at 30 percentthroughout the 1990s) during the 1993-1995 reces-sion. These included the 1996 Patto per il lavoro andthe 1997 reform of the labor market (Law 196/1997)also known as the pacchetto Treu. The pacchettoTreu represented the rst major departure from anItalian tradition of labor law heavily focused on lim-iting contract exibility, as it allowed new forms ofinternal labor market exibility (in particular work-ing time adjustments to address cyclical uctuations)as well as external exibility, by expanding the scopefor atypical (temporary and xed-term contracts). Itwas subsequently reinforced by the 2003 Biagi Re-form - the rst to extend the use of temporary con-

tracts throughout the private and public sectors

The 1990s and 2000 also saw two attemptsto create a minimum income of last resort schemeto address the problem of the working poor, a mea-sure that might have shifted the traditional bias insocial spending in Italy. Means-tested minimum in-come schemes (reddito minimo di inserimento) wereimplemented in an experimental way from 1998through 2003 and again, for a subset of municipali-

ties across Italy, between 2007 and 2009, in bothinstances under governments of the Left (Boeri andPerroti, 2002; Sacchi and Bastagli 2005; Colombino2012) Yet the Right repeatedly set back the clock onthese effort when it returned to power in 2001 andagain in 2008. The last iteration of these repeated at-tempts to extend social protection to cover labor mar-ket outsiders prior to the Eurozone crisis occurred

when the center Left government in the region of

20



Lazio tried to extend the minimum income programat the regional level without central government sup-port in 2009, but the effort was abandoned againafter the Right won the regional elections in 2010.

The electoral defeats of the Left by the Belus-coni coalition (in 2001 and, again, in 2008) also paved

the way for the breakdown of the social concertationprocess. The last major tripartite pact on collectivebargaining of the 1990s reached under the DAlemagovernment, the so-called Christmas Pact of 1998,maintained the principle of wage concertation basedon the two-level bargaining structure agreed in 1993(Negrelli and Pulignano 2007). But the second Ber-lusconi government rejected the policy principle ofconcertation replacing it with a less binding notionof social dialogue. Though in the rst tripartite pactreached under its tenure - the 2002 Pact for Italy-the government backtracked from its position by ac-

knowledging the role that concertation had played inallowing Italy to participate in EMU, it subsequentlyfailed to deliver on the policy objectives agreed towith the unions which included new income sup-port measures for the poor and unemployed in returnfor more monitoring of work searches and increased

public investment in the Mezzogiorno in subsequent

budget laws. The new social policy provisions of the2007 Pact for Welfare signed between the unionsunder the Centre-left Prodi government in 2007 -which included exibilization of eligibility criteriafor pensions so as to cover more outsiders (Colomboand Regini 2011) - were also rescinded after the cen-ter-right coalition came back to power in April 2008.

2.2 Spain: Social model reform in the context of

a highly segmented Labor Market

In the two decades leading up to the nan-cial crisis, Spains social model underwent a numberof important changes, many of which were agreed

as part of bipartite and tripartite social pacts. So-cial pacts had played a critical role during the demo-cratic transition of the late 1970s and early 1980s.A founding pact among political parties (Pactos de

la Moncloa) served to contain the wage price spirathat took hold during the last stages of the Francoregime and set the base for a series of subsequentripartite wage pacts. The effectiveness of this nego-

tiated incomes policy was enhanced by the provisionin the 1980 Labour Statute that extended the application of any bargain signed by any union meeting

representativeness criteria in works council electionsto all rms and workers in the sector or territoriallevel at which it was signed. One of these pacts, the1983 Acuerdo Econmico y Social (AES) - signedbetween the Socialist government of Felipe Gonzalez, the CEOE and the UGT (but not the CC.OO.)and covering wages for 1984-1986 proved particularly important. In it, the government and theCEOE agreed to increase the coverage of Spains

underdeveloped unemployment benets system andto raise minimum wages for youth in return for theUGTs acceptance of the legalization of temporarywork contracts (a concession that would lead to theprofound dualization of the Spanish labour markethereafter) (Guillen 2010). The PSOE governmentsrefusal following its second electoral victory in1986 - to compensate the Socialist confederationfor these concession with further social policy measures led negotiations on incomes policy to unrave

in 1988 (Perez, 2000). During the following yearsthe UGT which had delivered wage moderation tosupport the PSOEs adjustment policies only to seeunemployment rise above 20 percent and had suffersignicant losses in works council elections - wouldjoin the CC.OO. in a far more militant wage bargaining stance to recoup its credibility. The socialpact process that had underpinned the political transition would not recover until the PSOEs electoraldefeat by the right-wing Popular Party (PP) in 1996

Several factors conspired to allow for a re-launch of the negotiated reforms process followingthe 1996 elections. First, the two major labor confederations suffered a crisis of legitimacy as a resultof the even more dramatic rise in unemployment thafollowed the ERM crisis of 1992 (Polavieja 1998)Both confederations came under new leadership tha

20



recognized the negative fall-out of the militant bar-gaining positions taken in the 1989-1994 period andwere willing to take a more compromising stance.Secondly, the new government of Jose Maria Aznarneeded to strike a centrist tone (it was the rst timethe Right had come to power since the transition to

democracy and it did not have an absolute majority)

and embraced social pacts as a way to legitimize it-self. Third, as the 1993/1995 recession came to anend and the Bank of Spain began to lower interestrates to converge with those of other EMU coun-tries, Spain experienced a rapid economic recovery

and a dramatic improvement in the governments

scal position, allowing the PP government to reapthe benets of the orthodox policies pursued by itspredecessor and to back union concessions with

policy rewards in a manner that the PSOE gov-ernment had failed to do. In 1996 it thus signed alandmark agreement (the Pacto de Toledo) with theunions that set in motion the gradual revision of theold age pension system, allowing Spain to achievea far more sustainable pensions balance than mostof its eventual Eurozone partners had at the time.

In April 1997, it took further measures to sup-port a pact between employers and the unions com-

mitting both sides to seek national sectoral pacts insectors not yet covered at that level and to establishprocedures for the horizontal extension of collectivebargains (across sectors within a territory). This ef-fort to seek greater coordination in wage bargaining

through its centralization had limited success, as thepeak associations (in particular the CEOE) were un-able to bring along lower level afliates to consoli-date bargaining at the national level in most sectors.But from 2001 on in what may be seen as an effort

compensate the failure of bargaining reorganizationat the national-sectoral level the unions agreed toreturn to national bi-annual wage negotiations withemployers and to set global wage guidelines (Acu-erdos Interconfederales sobre Negociacin Colectiva(ANCs)) under which lower level negotiators wereinvited (though not bound) to increase wages in linewith ination and productivity (Herrmann 2005).

By contrast to Italy, where both pension re-form and labor market reform (in particular fol-lowing the 2003 Biagi reform extending the use oftemporary contracts) remained mired in conictin Spain contention between governments and the

unions centered more squarely on the latter, and spe-

cically on the issue of employment protection forthose on indenite work contracts. Seeking to recti-fy the extensive use of xed-term contracts, both theconservative Aznar government and the subsequentSocialist government of Rodriguez Zapatero thatcame to power in 2004 sought to reduce the resortto temporary contracts. Their strategy was to create

new categories of indenite labor contracts carryinglower dismissal costs and offering tax subsidies asincentives for employers to use such contracts. Re-ductions in labor taxes were also agreed (rst as partof the 1997 social pact and again in 2002. None ofthese reforms, however, managed to put a signicantdent in the predilection of employers (not just privatebut also public administrations at various levels ofgovernment) for temporary work contracts. The ex-traordinarily high proportion of employment on suchcontracts (illustrated in Figure 4, section 3 below)would only show a signicant decline from 2009 on,

when non-renewal of workers on temporary contractswas used as the rst line of employment shedding

2.3 Explaining the Contrasting Reform Dynamics

The political dynamics of social model re-form in Italy and Spain produced some convergencein the area of labor market regulation with the exten-sion of temporary work contracts in Italy during the1990s and 2000s. But the political process did not

allow for signicant change in the structure of socialspending in Italy, while it did do so in Spain. Asnoted, pension reform coupled with increased spend-ing on programs beneting the working age popula-tion allowed Spain to move its welfare state closerinto line with the distribution of expenditure in otherEU15 countries. And Spain had developed a far moreencompassing unemployment benet system which

20



protected all categories of workers and provided aminimum benets to workers on xed-term contracts.

A number of factors help explain the contrast-ing pattern of social model reforms in Spain in thedecades prior to the crisis. Many of the changes im-

provements in social protection introduced in Spain

were part of the social pact making process, which,as noted, broke down repeatedly in Italy. This in it-self can be linked to other factors. We have alreadypointed out the opposite political contexts: the pre-dominance of minority governments in Spain versus

the majoritarian turn in Italian politics following the1993 electoral reform. Three further differences werealso important. The rst is the sharply different scalposition of the state, which allowed Spanish govern-ments to support agreements between employers andhamstrung Italian governments, in particular those ofthe Left, in their efforts to reform the Italian socialmodel. After the economic adjustment effort thatthe last Gonzalez governments imposed to keep thepeseta in the ERM, Spanish governments were ableto support pacts between employers and unions withscal resources, a signicant contrast with continu-ous pressure to bring expenditures under control onItalian governments in their efforts to scale back Ita-

lys persistently high public debt. Thus, while Span-ish social pacts focused (as elsewhere in Europe)on regulatory trade-offs, there were also signicantexamples of distributive bargains - including publicinvestment, tax relief, renegotiation of the minimumwage (by more than 30 per cent in 2004), as well aspension reform concessions to the unions (in 2001and 2007), in return for their support for a strongercontributory element in the PAYG pensions systemand for private complementary pensions. Such quid

pro quos helped Spanish governments secure ongo-ing union commitment to wage moderation and facil-itated agreement on changes in labor market reform(Rodrguez Cabrero 2002; Molina and Rhodes 2007).

Although the divergence in social model re-forms in Italy and Spain can be dated back to theearly 1990s (or even the 1980s, with the revamping

of the unemployment insurance system), the leewaySpanish governments had to support distributive

bargains relative to their Italian counterparts wasvastly augmented following monetary union. Thetwo countries joined the Euro with very different

macro-economic fundamentals and, given the Eurozone economic constitution which emphasized pub-

lic debt and decits without creating mechanisms tocompensate asymmetric growth and nancial owsin the Eurozone, this set off very different dynamics

in the two countries. Spain entered the union with a

particularly low public debt burden and a regulatoryframework that allowed for a very high level of labormarket dualism and, implicitly, exibility of employment creation. It thus experienced a decline in interest rates without any demand for compensatory scarestraint coming from the Eurozone (given that its

public debt position was far better than that of coreEurozone countries prior to the crisis). This spurreda rapid expansion of credit by Spanish nancial institutions and investment concentrated in non-tradablesectors (notably construction) that was nanced indirectly by large private savings ows from the Eurozones core economies via the Eurozones inter

bank lending market and purchases of Spanish loanbacked bank securities by other Eurozone banks

The result was a real estate bubble that eventually proved disastrous for the Spanish economyYet while it lasted, the boom in domestic demandencouraged a very strong pace of labor force activation (in particular of women) and spectacular employment growth. This facilitated the pattern of social policy change which moved Spain away fromthe traditional familial male-breadwinner modewith which it was associated in the VOC literature

and considerably closer to the social policy regimesof more employment friendly Northern EU economies. By contrast, in Italy, which joined the Eurowith a particularly high public debt burden, governments struggled to meet the Eurozone Growth andStability Pact objectives after entry. Although EMlowered the risk premium on Italian public debtand allowed some employment growth (as well as

20



price rises in non-tradables sectors), the ongoingneed to reduce public debt in compliance with theGSP underpinned a period of limited government in-vestment and extended low growth through the rstdecade of the currency union. This general contextundercut repeated efforts by governments of the

center-Left to shift the balance of welfare spending

in ways that might have encouraged further laborforce activation and more signicant pension reform.

Secondly, while Spanish unions managed toconverge on a strategy of moderation and accom-

modation after the mid-1990s and to maintain theirunity in action, the main Italian labor confederationssaw deepening divisions in the 2000s. This can part-ly be attributed to the fact that the pay-offs to unityin the form of government quid pro quos were someager in Italy and to the fact that Spanish govern-ments became more committed to social pacts afterthe early 1990s when attempts by the PSOE govern-ment to go it alone proved to be counter-productive.But unity between the labor confederations andtheir commitment to a strategy of accommodation

in Spain also t the unions need to broaden theirconstituency to the growing and very large num-ber of workers on temporary contracts, which con-

stituted the majority of workers in some sectors,such as construction and in the public sector. Thisaltered the nature of insider-outsider dynamics inSpain more than it did Italy, allowing employmentregulation to become a topic for social pact bargain-ing. Meanwhile, in Italy, the return to power of theBerlusconi coalition in 2001 against the backgroundof majoritarian electoral reform and party-systemtransformation brought with it a more confronta-

tional style by governments of the Right and this

increased tension among the labor confederations.

Lastly, to some extent the far-reaching re-form of the wage bargaining system in Italy in theearly-to-mid 1990s and the subsequent institutionalembedding of wage moderation took care of one

of the most critical reforms for both unions and

employers (the former gained greater control ofthe system, while the latter secured the end of in-ationary wage claims, even if they wanted furtherreforms to enhance wage exibility). Thus, a majorbone of contention and the most compelling issuefor social bargaining by Italian governments andemployers was removed. In Spain, as noted, the

collective bargaining centralization foreseen in the1997 Pacts remained incomplete. This, in additionto other factors (including the unions goal of pre-venting a further decentralization to the rm levelwhere their presence, at best, was very weak) helpedproduced the annual bipartite ANCs (Acuerdos parala Negociacin Colectiva) which set national cross-sectoral pay parameters from 2001 on. This bipar-tite overarching wage setting framework helpedto sustain the broader process of social bargainingwell past the onset of the economic crisis in 2008

3. Social Models and Economic Performance

under EMU

The austerity cum labor market liberalizationcentered approach pushed on Spain and Italy by theirEuro partners have been premised on the assump-

tions that social model characteristics played a major

role in landing both countries in the external nan-cial positions that rendered them susceptible to con-tagion during the sovereign debt crisis. Even when

as in the case of the IMF (2013) and more indirect-ly the EU Commission (Bontout and Lokajickova2013), there have been acknowledgements that thepace of budgetary restriction may be undercutting

the utility of liberalizing labor market reforms andcontributing to the very outcomes they are meant

to address (economic and labor market stagnation

public decits and nancial sector vulnerability)these same institutions (and in the case of the EU,the Eurogroup) have insisted that the pace of labormarket and spending reform be kept up or even tak-

en further as a way to facilitate the internal devalua-tion that is offered as the only way out of the crisis

What role social model features (i.e. socialprotection and labor market regulation patterns)

20



played in contributing to the situation of the twocreditor countries, however, remains very unclear.The argument (when it is spelled out) usually refersto excessive borrowing and loss of competitivenessas measured by the evolution of unit labor costs inrelation those of Germany, linking these phenomenato the deteriorating current account positions the two

countries in the run up to the crisis but without muchanalysis of the sources of this deterioration. Labormarket failures such as the phenomenally high levelof unemployment in Spain since the outbreak of thecrisis and that of youth unemployment in Italy aretypically attributed to persisting rigidities that keepthose markets from clearing. Although concessionshave been made that there are other market failuresat play (most notably the faulty transmission of mon-etary policy in the Eurozone to creditor countriesas reected in the much higher cost of credit facedby rms in the latter) the fundamental demand thatthe pace of labor market liberalization and spend-ing reform be altered so as to allow adjustment viainternal devaluation has been maintained in inter-governmental negotiations within the Eurogroup.

This section takes up the question of the roleof social model features in the economic performance

of the two countries following the launch of the Euroand in the course of the crisis. I rst review variousperformance indicators before turning to the questionof the likely contribution of social model features inlight of other factors impacting the economic perfor-mance of the two countries. Looking rst at growthrates, Figure 2 illustrates the dramatically differentexperience of the two countries in the decade leadingup to the nancial and sovereign debt crises. WhileItaly, with the brief interlude of 2000-2001, experi-

ence consistently low growth in the decade leadingup to the crisis (averaging 1.6% from 1999 to 2007compared to the Eurozones 2.3 average), Spainjoined the Eurozone at a time when it was begin-

ning to experience an economic take-off and grew

consistently well above the Eurozone at an averageannual rate of 3.7% over the same period. After2008, Spain experienced a considerable contraction

Figure 2: GDP Growth rates (Source: Eurostat)

-8.0

-6.0

-4.0

-2.0

0.0

2.0

4.0

6.0

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

Euroz

Spain

Italy

(adding up to 5.6% of its GDP from 2008 through2012). Italy experienced a short acute recession from2008 to 2010, falling back into recession in 2011 anda cumulative contraction of 6.9% from 2007-2012

Turning rst to the most common indicator of labormarket performance the unemployment rate thecontrast between the two countries could not bestarker. Spains has been marked by extreme uctua-tion over the last two decades, from a rst high of 24per cent in 1994 down to 8 (right around the Euro-zone average) just prior to the crisis, and up again toa record 27 percent at its height in 2013, before eas-ing back to 26 per cent. This rise, as some have sug-gested, appears quite disproportionate to the actual

contraction in Spanish GDP over the period. Italysunemployment rate also varied considerably but farless so than Spains, from 12.3 percent in 1998 downto 6.8 percent in 2008 and up to 12 percent by theend of 2013. On the other hand, a common feature ofboth countries is their persistently high youth unem-ployment rates (age group of 15-24), which has beenwell above the EU15 average throughout the period

20



Nevertheless, it is noteworthy that in the case of Spainyouth unemployment has conformed to the generalrule in the EU (with the notable exception of Germa-ny) in that it tends to double the overall unemploy-ment rate and, thus, appears to be a function of that

overall rate. In Italy, by contrast, youth unemploy-ment has tended to exceed general unemployment by

a higher multiple (a more than 3 to 1 ratio in 2012).

However, because both Italy and Spain alsoexperienced very signicant demographic changeduring the decade prior to the Eurozone crisis - through

1) immigration (of around 2 million into Italy and 5million into Spain according to ofcial gures) and2) change in female labor force participation -- unem-ployment rates are not in fact a very good measure of

labor market performance. It is far more informativeto focus on employment rates which show the numberof job holders relative to the working age population.

As illustrated in Figure 3, both Italy and Spainexperienced very considerable employment growthduring the rst decade of EMU right up to the eco-nomic crisis. That employment growth, nonetheless,was particularly pronounced in Spain where jobsincreased by almost 6.5 million from 1999 to 2007

(compared to 2.5 million in Italy). Indeed, employ-ment growth in Spain was so strong up until the crisisthat it accounted for more than 40 percent of all jobscreated in the Eurozone to 2007. The rise in overallemployment rates also reected signicant increasesin the employment rates of women in both countries(Figure 4). Again, this phenomenon was much moremarked in Spain, which saw a rise in the employmentrate of working age women from 41 percent in 1999to 56 percent in 2008 (compared to a rise from 40percent to 47 percent for Italy over the same period).The dispersion of regional employment rates (espe-cially for women) also declined in Spain consider-ably more than in Italy during this period, leadingsome sociologists to argue that Spain had undergonea denitive turn from the family/kinship to the du-al-earner family model (Naldini and Jurado 2013).

igure 3: Employment to Population Ratio (OECD Labour Force Statistics iLibrary, May 2013)

igure 4: Employment rates by Gender (OECD iLibrary, May 2013)

45

50

55

60

65

70

Spain

EU15

Italy

Figure 4: Employment rates by Gender (OECD iLibrary, May 2013)

35

40

45

50

55

60

65

70

75

80

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Italy Men

Spain Men

EU15 Men

Italy

Women

Spain

Women

EU15

Women

20


20/58



in poverty reduction than the Italian welfare state.Thus, while the at-risk-of-poverty rates after trans-fers in Italy was, at almost 20 percent in 2012, only4 percent lower than the pre-transfer proportion, inSpain it was, at 22 percent, 7 percent lower than thepre-transfer proportion that year (not very different

from the level in Germany for instance) (Eurostat).

Lastly, there is the performance of the twoeconomies in terms of competitiveness measures.

Much attention has been paid to the relative rise ofunit labor costs of the Eurozone periphery coun-tries in the period leading up to the sovereign debtcrisis and in the role this played in the rise of thesecountries current account decits during the period(see for instance Dadush 2012; Sinn 2013). Bothphenomena, and the connection between the two,

are central to the view that the fundamental prob-lem of adjustment in the Eurozone is one of com-petitiveness and that social model changes are criti-cal to rectify this problem. This view also servesas an implicit justication for the position that in-ternal devaluation by way of regulatory reformsintended to reduce labor costs was the only way toresolve the crisis and that such adjustment via inter-nal devaluation had to be achieved as a precondition

for macro-institutional changes in the Eurozone.

Figure 6 illustrates the evolution of the RealEffective Exchange Rate (a measure that takes into

account developments in each countrys main trad-ing partners) of Italy and Spain based on their unitlabor costs. It shows that, but for a brief period ofdepreciation up to 2001, the Eurozone as a wholeexperienced a signicant appreciation of its REER,with Spain and Italy mostly trailing the Eurozones

overall development until 2005 and then experienc-ing somewhat higher ULC growth in Spain, followedby a considerable depreciation (internal devaluation)in Spain following the start of the crisis that wasclearly not matched in Italy. Indeed, that internal de-valuation of ULC is likely to have been much greaterthan the graph indicates, as aggregate compensation

gures in Spain were articially inated after 2007

by the fact the share of low wage jobs declined farmore dramatically than that of higher wage jobs. Theactual reduction in wages behind the gures is thusestimated to have been approximately twice as muchas the such aggregate compensation gures indicate(Bank of Spain, Economic Bulletin, February, 2014)

Figure 6: Index of Real Effective Exchange Rate, ULC based (2005 = 100 ), IMF International F

Statistics

40.000

50.000

60.000

70.000

80.000

90.000

100.000

110.000

120.000

130.000

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Italy

Spain

Eurozone

The role that social model features played in thisgrowth, competitiveness and labor market perfor-mance picture, however, is far from clear. The mat-ter is complicated by the fact that monetary unionimplied a major macro-economic shock for bothcountries because it altered the course of monetarypolicy by bringing interest rates down to those ofcore countries and, at the same time, encouraged a

signicant intensication of nancial integration inthe Eurozone area (Obstfeld 2013). At the same timeother factors affecting domestic demand, such as the

leeway for public investment or for growth of pri-vate credit differed greatly between the two countriesgiven their very different public debt positions andthe Growth and Stability Pacts exclusive focus onpublic debt.

20



economies starting as early as 1998 (when their central banks began to lower their prime rates to converge with the Eurozone average), the need for continued scal consolidation to address the public debin Italy meant that the effect of the monetary policychange was much larger for Spain because it fed intoa process of economic recovery that was already un

derway. As lower interest rates in Spain coincidedwith a continuing decline in public debt (and hencecreated greater room for private borrowing), they seoff a prolonged demand-driven growth cycle centeredin the construction and real estate sectors. With employment rising, more people were able to move intothe labor market and therefore also into the housingmarket. Women, long punished by decades of highunemployment (due, in part, to Spains earlier orthodox scal and monetary policies) had acquired highlevels of educational qualications and were particularly well primed to enter the labor force in large numbers (ergo the 20 percentage point rise in their labouforce participation rate in just a decade). Spains rapidly expanding labor market also attracted large numbers of immigrants (many of whom found work in the

construction sector) producing a demographic boomDemographic growth, low interest rates and risingemployment drove housing prices and, by way of the

wealth effect, further fed private borrowing. In Italyby contrast, lower interest rates appear to have madesome contribution to domestic demand in the imme

diate period after 1998. But this effect was quicklymuted by the continued need to reign in public sector debt following an initial, short post EMU boom

The main difference in the demand proleof the two countries was on the side of investmen

(rather than consumption), specically in commer

cial and residential housing investment and in capital formation. Housing investment in Spain roseat a very fast rate so much so that housing starttopped those of the three largest economies in the Eurozone (Germany, France and Italy) taken togethein the last years of the boom (Estrada et al., 2009p.35 ). Investment in in other construction waalso considerably higher (by over two percent) than

The run-up to monetary union involved largescal adjustment efforts in both Italy and Spain. Yetthe fundamentals emphasized by the Growth andStability Pact with which the two countries joinedthe Euro were far apart, in particular their publicdebt positions. One reason for this had to do withthe 1992/93 ERM crisis. In the case of Spain, whichhad relatively low public debt even at its height in the1990s, the central bank and the Socialist governmentof the time chose to pursue all necessary measures tokeep the Spanish peseta in the ERM and avoid the fate

of the Italian lira and the British pound, which wereforced out of the mechanism. One consequence ofthis was the rst dramatic surge in unemployment re-ferred to above, to almost 25 percent in 1994. In the

case of Italy, which faced the ERM crisis with a farworse public debt position, the employment conse-quences of that crisis were more limited because theliras exit and large devaluation afforded relief for thecorporate sector. The orthodox course of the Socialistgovernment in Spain coupled with a series of con-trolled devaluations of the peseta within the systemand the much lower accumulated public debt burden- allowed for a very successful scal turnaround. Theprimary decit was cut to 2 percent as early as 1997

and would continue a steady decline from there on,leaving Spain in primary budget surplus territory formuch of the 2000s. Italian governments imposedlarge budget cuts from 1996 to 1999, lowering thegovernments annual decits from 6 percent to a lowof 0.8 percent in the year of EMU entry. But theirfailure to bring the large public debt most of it ac-cumulated in the pre-1993 First Republic period- into line with the Stability and Growth pact meant

that scal policy remained continuously constrainedfollowing monetary union, a fact that only exacerbat-ed the standoff over social model reform in the 2000s.

The different public debt positions with whichItaly and Spain joined the Eurozone meant that theeffect of the move to a single monetary policy ondomestic demand was of a different order. Though

EMU brought a reduction in interest rates for both

20



in these three other economies. The boom in Spain

also had a self- reinforcing character in that it bothdrove and was fed by the large scale demographicimpact of immigration - itself driven by job growth.It may have been further facilitated by a liberaliza-tion of land zoning laws carried out by the rst Aznargovernment, by tax incentives encouraging home

ownership, and by the failure of the Bank of Spain toreign in the lending of regional savings banks to thereal estate sector. However, in evaluating the role ofthese other factors it is important to keep in mind that

without the lower interest rates and the rise in loan-able funds coming through the Eurozones interbankmarket following monetary union, the Spanish realestate bubble could never have developed as far asit did. Indeed, it would not have been possible with-out two much larger developments in the Eurozone:1) the mismatch between ECB policy (set largely tomatch the different output-gap scenarios of core Eu-

rozone countries, notably Germany and France) anddomestic demand conditions in Spain (or, for that

matter, in Ireland, which experienced very similarconsequences), and 2) the nancial imbalance cre-ated by the rise in the German national savings ratefrom 2002 onwards, which played a key role in -nancing the Spanish bubble through the interbank

Eurozone market (on this point see Pettis 2012).

Figures 7a and 7b show the evolution of out-put gaps for Spain and Italy in comparison to theirtwo largest Eurozone trading partners (Germanyand France) before and after the crisis. As 5a illus-trates, Spain and Italy joined the Euro with differentoutput gap positions (quite negative in Spain, posi-tive in Italy), largely because Spain had suffered amuch stronger recession after 1992. Between 1999

and 2001, all Eurozone countries experienced a shortboom bust cycle, and France and Germany enterednegative output gap territory in 2002 lasting to 2006.

Figure 7a: Output Gaps up to crisis (percentage of GDP)

Source: IMF, Word Economic Outlook Database, May 2013

-3

-2

-1

0

1

2

3

4

France

Germany

Spain

Italy

Figure 7b: .and from 2007 on

Source: IMF, Word Economic Outlook Database, May 2013

-6

-5

-4

-3

-2

-1

0

1

2

3

4

5

2007 2008 2009 2010 2011 2012

France

Germany

Spain

Italy

20



During this period the ECB renance ratebecame very accommodating (see Figure 8), in par-ticular as most Eurozone economies (including Ger-many, France, and Italy) were unable to meet theStability and Growth Pact decit objective. Whilethis scenario persisted, Eurozone banks in slow-er growing economies (principally Germany and

France) directed excess savings towards other Euro-zone countries (much of it through interbank lend-ing and purchases by German and French banks of

Spanish bank issued covered bonds). Although otherfactors restrained investment in Italy, the output-gapgures suggest that ECB policy was also quite ac-commodating for the Italian economy, which alsosaw considerable nancial inows, although thesewere directed largely to Italian public debt. Then,as Germany and France re-entered positive output

gap territory in 2006, the ECB raised rates (Figure8) and kept them up until 2009 coinciding preciselywith the Global Financial Crisis. This step precipi-tated the very fast collapse of the Spanish real es-tate market (which was based on variable rate loanslinked to the Euribor) and the crisis of the Spanishsavings banks with the heaviest exposure to that

market. In Italy, the dynamic was different as easymoney had owed mostly into government bonds

prior to the crisis, lowering the public debt burdenbut without producing the private investment boom

observed in Spain. Nonetheless, up until the crisisinows of funds from surplus countries helped easethe nancial burden of the still large public debt that

Figure 8. ECB Refinance Rates

Source: ECB

would prove Italys Achilles heel during the sover-eign debt crisis.

It is critical to keep in mind this macro-econom-ic backdrop in considering what role social modelfeatures may have played in the economic perfor-mance of Italy and Spain and in particular that of

their labor markets - before and after the crisis. Thedifferent demand scenarios not just between Italyand Spain, but between Spain and most of the rest

of the Eurozone up to 2007 is particularly impor-tant in understanding the economic boom and out

of the ordinary employment growth in Spain. By2007, the construction sector accounted for almosttwice the share of GDP (11 %) as in the rest of theOECD (5.6%), and a much larger share of employ-ment (13%) than the rest of the Eurozone (7.7%)

By contrast in Italy, the size of the constructionsector closely reected the Eurozone averages, ac-counting for 5.4% of GDP and 7.7% of employment

With this macro-economic backdrop in mindwe can consider arguments that are commonly of-fered for how social model features may have con-tributed to the economic outcomes Spain and Italyexperienced following monetary union and to the vul-

nerable position they found themselves in during thesovereign debt crisis. Chief among these outcomes

are 1) the much larger labor activation and employ-ment growth experienced in Spain prior to the cri-

sis, in particular by women, 2) the far more dramaticemployment destruction in Spain in the course of thecrisis compared to relative employment stability in It-aly, 3)the growth in ULC both countries experiencedup to the crisis, and 4) the much faster internal deval-uation in Spain compared to Italy and 5) the failure

of this devaluation to stem the fall in employment.

Focusing rst on the different employmentand growth performance in the years leading up tothe crisis, we have already noted the key differencein macroeconomic fundamentals with which the twocountries joined the Euro and the extraordinary de-

mand boom to which the single monetary policy con-tributed in Spain. Employment growth was also sup-

20



ported in Spain by the fact that, as we shall see below,labor compensation growth remained quite modestconsidering the high GDP growth rates, though this

was also true in Italy. These factors are likely to ac-count for the bulk of the difference between the twocountries. Nonetheless, one difference in the twocountries social model features the ease with which

employers could expand temporary employment inSpain could also have played an ancillary role.

First, it is clear that the Spanish employmentregime allowed for very fast employment creation assoon as monetary policy took a less orthodox turn.The fact that the proportion of temporary employ-ment to overall employment nonetheless did notrise very much suggests that, at least during the eco-nomic growth period, temporary employment did ef-fectively work as a stepping stone towards indeniteemployment. The years (2003-2007) in which theratio did increase as a result of the particularly in-tense expansion of the construction sector were ones

in which the wealth effect of rising housing priceshad already taken hold (so that people invested inreal estate for its expected appreciation) and duringwhich the mismatch between ECB interest rate poli-cy and conditions in Spain was particularly acute (as

suggested by the output gap data). We can thereforeventure the conclusion that the exibility affordedby temporary contracts in Spain is likely to haveinteracted with the very particular macro-economiccontext so as to feed, rather than restrain, the impact

of those macroeconomic conditions on employmentgrowth and on the allocation of resou

ces open forum 20: eurozone crisis and social models: what we can learn from italy and spain

Documents