MORE POWER TO THE EUROPEAN PARLIAMENT? Abdul Noury, ECARES-ULB and Gérard Roland, UC Berkeley, ECARES, CEPR and WDI.

Transcription

1 MORE POWER TO THE EROPEA PARIAMET? Abdul oury, EARE-B and Gérard Roland, Berkeley, EARE, EPR and WDI

2 Today, the E i no longer a mere union of tate, but more and more a union of citizen everthele, Euroean deciion are till taken almot ecluively by the tate The role of the elected Euroean Parliament a a ource of direct legitimation i underdeveloed Thi role ha to be further trengthened if we are to overcome the democratic deficit of the nion- through more deciion-making ower for the Euroean Parliament and the creation of a econd chamber [] Diviion of cometence, more weight for the Parliament, making the treatie imler- never ince the early day of the Euroean integration were there o many fundamental quetion on the agenda and never were there o many momentou deciion to be made 1 Jochka Ficher 1 Introduction Prooal by German chancellor chroeder and Foreign Miniter Ficher to increae the ower of the Euroean Parliament relative to thoe of the Euroean ommiion and the Euroean ouncil have been increaingly at the forefront of olicy debate in Euroe uch rooal have triggered eriou debate at the highet level of Euroean olitic The iue of increaing the ower of the Euroean Parliament i likely to be at the heart of rooal for reform that are being dicued at the onvention that ha been etablihed at the December aeken ummit and at the 2004 intergovernmental conference to revie the ice Treaty What effect can one eect from increaed ower to the Euroean Parliament on deciion-making in the E? What doe the at eerience of voting behavior in the Euroean Parliament teach u about how a Parliament with enhanced ower would behave? When aking about how the Euroean Parliament (EP) would behave if it had more ower, one ha everal quetion in mind: 1) How are voting coalition likely to be 1 From the eech by Jochka Ficher in Dublin on 30 Aril 2001 The eech can be found at: 2

3 formed? What will be the role of normal left-right olitic and coalition baed on cro-national ideological familie and what will be the role of country coalition where grou of countrie coalece on the bai of their national interet? 2) How table and redictable will voting attern be? In other word, what degree of diciline can one eect from coalition forming in the Parliament? Giving an anwer to thee quetion i crucial in evaluating the effect of increaed ower to the Euroean Parliament In articular, more ower to the EP would have ubtantially different imlication deending on whether legilator vote more along ideological arty line or according to coalition of national interet Indeed, if coalition are cro-border and baed motly on the logic of the left-right divide, then giving more ower to the EP hould reinforce integrationit tendencie in the E and encourage Euroean-wide olitical debate along the traditional left-right dimenion On the other hand, if coalition are baed motly on national interet, then increaing the ower of the Parliament i likely to make deciion-making in the E le redictable and more chaotic Indeed, coalition will more likely reemble thoe that form in the Euroean ouncil but the high number of Euroean rereentative and the higher diverity of interet involved i likely to make it more difficult to form majority coalition Increaed ower to uch a body i very likely to lead to aralyzing deciion-making and make it le table and redictable The anwer to thoe quetion i even more imortant in the light of the future enlargement of the E to countrie from entral Euroe If the EP vote motly along ideological arty line, enlargement will mainly increae the ize of the EP and oibly affect it ideological comoition but it would not fundamentally threaten it ability to oerate Giving more ower to the EP would then be beneficial for deciion-making within the E becaue it would mitigate the danger of aralyi in deciion-making that are inherent to enlargement, eecially in the light of the ice Treaty (ee eg Baldwin et al, 2001) On the other hand, if the Euro-deutie vote mainly on the bai of national htt://wwwauwaertige-amtde/www/en/uche/eingabe 3

4 interet, more ower to the EP i likely to further contribute to the aralyi of deciionmaking in the E ouncil The literature tudying intitution of the E and i raidly growing and the number of contribution devoted to the EP i becoming large (Hi, 2001; Kreel and Tebeli 1999; Raunio 1997) The eiting academic reearch on the EP doe not rovide traightforward anwer to uch imortant quetion mainly due to the lack of the data In articular, the imact of increaed ower to the Parliament or of enlargement ha not yet been analyzed We are in a oition to give a better anwer to quetion about the EP than the eiting literature We have ut together a unique databae including all roll call vote, ie legilator recorded vote, from all individual deutie between 1989 and 1999 Thu, our databae cover in the comlete Third and Fourth arliament, with reectively 2733 and 3740 roll call vote regitered for the former and the latter 2 uch a databae allow the mot comlete analyi to date of voting behavior The uefulne of uch a databae goe well beyond the uroe of the aer at hand Very little i known in the large ublic about the E Parliament Yet, the information made available to the ublic about roll call allow to crutinize very carefully the behavior of Euro-deutie, much more o than for mot national Parliament in Euroe where uch information i le readily available Thi information will become more and more valuable over time and will allow to undertand the dynamic evolution of voting behavior in the Euroean Parliament in it early year Turning back to the objective of our aer, the analyi of the dynamic of voting behavior in the EP i clearly ueful to ae the effect of increaed ower to the EP and of enlargement The ower and reonibility of the Euroean Parliament ha teadily 2 Thi i art of a larger reearch roject jointly organized with imon Hi of the ondon chool of Economic utting together a databae on voting in the Euroean Parliament from 1979 to

5 increaed over time Before 1987 the EP had only a conultation ower The ingle Euroean Act of 1987 increaed the ower of the EP by introducing the cooeration rocedure The 1990 Maatricht Treaty granted further ower to the EP by initiating the co-deciion rocedure Finally the 1997 Amterdam Treaty ha trongly etended the co-deciion rocedure nder thi rocedure, which require the agreement of both the ouncil and the EP on legilation, the EP ha eentially the right to amend and veto legilation over many olicy area uch a internal market, ocial olicy, conumer rotection and the environment Thi increae in ower ha not only changed the tatu of the Euroean Parliament a an actor within the Euroean nion but ha alo had an effect on the voting and articiation behavior of it memberhi Moreover, under the Fourth arliament, the EP wa enlarged to rereentative from the new entrant Autria, weden and Finland We can thu evaluate the effect of that enlargement on voting behavior in the EP The emirical analyi of roll call vote in the EP ugget clear anwer to the quetion we ak: - Ye, giving more ower to the EP hould reinforce an already very viible tendency toward traditional left-right olitic tyical of national legilature; - Ye, giving more ower to the EP in the contet of enlargement can enhance the functioning of E intitution and mitigate aralyi in deciion-making that rik to lague the E ouncil after enlargement Thee are imortant olicy imlication in the enlargement contet The ice Treaty made all the imortant intitutional deciion relative to enlargement, not only by etting the weight of the 12 otential new entrant in the Euroean ouncil (and alo in the Parliament and in the ommiion) but alo by deciding on the majority rule in the ouncil Baldwin et al (2001) howed that with enlargement to 27 E countrie, the aage robability 3 of a deciion in the ouncil would go from 78% under the current rule with 15 E countrie to 21% under the ice rule, a clear deterioration The ice 3 The aage robability i the hare of winning coalition among all oible coalition of countrie in the E ouncil It i an abtract indication of the difficulty of aing a bill 5

6 Treaty ha thu failed to addre atifactorily the iue of oible aralyi of deciionmaking within an enlarged E The reult of our analyi ugget that uch a aralyi of deciion-making i not likely to haen in the EP Moreover, debate and coalition in the EP will take lace along the traditional left-right dimenion Even if no change i made in the voting rule in the ouncil after enlargement, we think that giving more ower to the EP hould imrove deciion-making in the E under the ice rule There are two reaon for thi Firt, more ower to the EP hould increae coheion of arty grou a our reearch ugget and thu generate rather redictable voting attern in the EP Thi would alo increae the legitimacy of the EP econd, an increaed legitimacy of the EP i likely to create reure on the ouncil ountry rereentative in the ouncil would tend to ue their veto right more arely if they feel the reure of the EP, eecially in cae where their veto right would be ued to defend ecial minority interet grou at home A more owerful EP having at heart an-euroean interet (along the left-right dimenion) would doubtlely hel reveal uch veto ractice to the general ublic, thereby leading ouncil member to refrain more from uing their veto right to defend ecial interet which conflict with the more general interet of the ublic Our aer roceed a follow ection 2 eamine the role of the EP in the E deciionmaking roce a a whole and dicue why the iue of coheion and coalition formation in the EP are relevant in E-deciion-making in the enlergement contet ection 3 eamine whether Euro-deutie vote rimarily according to their ideology or to national interet To do that we meaure the coheion of arty and country vote and analyze coalition formation for both arty grou and national delegation Then we focu on Euro-deutie voting behavior in a multi-dimenional olicy ace In ection 4 we ue regreion analyi to eamine the effect of enlargement and increaed ower on voting diciline The behavior of new entrant and the Member from the oorer countrie are analyzed reectively in ection 5 and finally ection 6 conclude 6

7 2 THE EROPEA PARIAMET WITHI THE EROPEA ITITTIO I THE PERPETIVE OF EARGEMET ince the Foundation of the Euroean ommunity, the two main deciion-making bodie have been the Euroean ommiion and the Euroean ouncil The ucce of Euroean integration owe a lot to the Méthode ommunautaire devied by the Founding father of the E The Méthode ommunautaire i a deciion-making method where the Euroean ommiion ha more or le ecluive right to draft legilative rooal The ommiion lay thu the role of advocate of Euroean integration while the Euroean ouncil, rereenting the head of government of E countrie, lay the role of a veto chamber, regrouing national interet The agenda-etting role of the ouncil ha remained limited Wherea it can reject rooal by the ommiion, a unanimity in the ouncil i needed to amend a ommiion rooal The ucce of the Méthode ommunautaire relie on the fact that the ommiion ha contantly rooed new legilation for Euroean integration, thu reventing a collective action roblem at the level of initiation of legilative rooal, wherea the ouncil, which baed it deciion motly on unanimity rule in the firt decade of Euroean integration, rereented overeign national interet ntil 1979, the Euroean Parliament wa eentially comoed of delegation from national Parliament conulted on a narrow range of legilative rooal ince 1979, the EP i elected by univeral uffrage and i the only major E olitical body directly elected by Euroean citizen It ower within the E were initially virtually nil a it layed the role of a conultative body It ower have neverthele gradually eanded over the year The ingle Euroean Act introduced the cooeration and the Aent rocedure The ooeration rocedure alied to roughly one third of the legilation brought to the EP, and in articular the bulk of legilative harmonization related to the ingle Market rogram The Aent rocedure alie to the admiion of new member, to aociation agreement with Third countrie and a few other tak like modification of the rocedure of Euroean election, modification of the rotocol of the EB, anction againt eriou breach of the Treaty by a member tate Here, a ouncil oition mut be aroved by a majority (majority of Member for acceion and 7

8 electoral reform, otherwie imle majority) The cooeration rocedure i comoed of two reading where the firt one i like the conultation rocedure At the econd reading, the ouncil deciion in the firt reading i ent to the EP who ha 3 month to either arove, reject or amend Rejection require a majority of EP member The EP oition i then ent back to the ommiion and to the ouncil The ommiion can either withdraw the rooal, accet or reject the EP amendment Amendment by the EP acceted by the ommiion require unanimity to be rejected by the ouncil nanimity i alo required to reject the ommiion rooal Amendment rejected by the ommiion can be adoted by the ouncil again by unanimity While the increae in ower entailed by the cooeration rocedure wa not huge, it wa neverthele ignificant Deciion that go through a econd reading are uually deciion the ouncil want to a The EP may thu ue the threat of imoing, in the econd reading, a unanimity requirement that may not a in the ouncil It may alo threaten the ommiion with rejection if it doe not adot it amendment Thu, in cae of vote with qualified majority, the EP ha ome clout under the cooeration rocedure IERT FIGRE 1 The o-deciion rocedure, etablihed in the Maatricht Treaty and imroved in the Amterdam Treaty give more ower than before to the EP Firt, it allow rejection of legilation by the EP at the econd reading econd, if the ouncil doe not agree with an amended rooal by the EP at the econd reading, the rooal goe to a conciliation committee comoed of an equal number of member on each ide reonible for carving out a comromie If a comromie fail, the rooal i rejected Otherwie, it i adoted rovided it i acceted by the EP (imle majority) and by the ouncil (Qualified majority) The co-deciion rocedure thu give effective bargaining ower to the EP a it can ue it right of rejection to negotiate comromie with the ouncil ote that the role of the ommiion ha been weakened under the co-deciion rocedure a it i not imortant any more for the EP to ecure uort of the ommiion for it amendment o-deciion now cover a great deal of E legilation with the imortant ecetion of EM, agriculture, fiherie and fical harmonization Moreover, in ome co-deciion area (citizenhi, mobility of worker, elf-emloyed, culture), unanimity i till required in the ouncil which make bargaining with the EP articularly difficult 8

9 IERT FIGRE 2 Deciion in the EP are uually taken by imle Majority but a majority of member i required for rejection under cooeration and co-deciion Vote are of 3 tye: by how of hand (rare), electronic vote or roll call vote A roll call vote can be requeted if at leat 32 MEP or a olitical Grou ak for it MEP do not know in advance whether a roll call will be requeted or not Roughly a third of all vote are by roll-call but their hare ha been increaing over time In 1988, the number of roll-call vote er hour of lenary eion wa about 11 In 1998, it wa about 15 In our databae, 2291 out of 6473 vote were directly legilative and ued one of the eiting rocedure Among the legilative vote, 46% ued the conultation rocedure, 27% the cooeration rocedure, 25% the co-deciion rocedure and 1% the aent rocedure MEP are organized in olitical grou The eating in the hemicycle are like in national Parliament with artie ranked from left to right according to their ideology In the current Parliament, the eating i a follow At the etreme left, one ha the radical left which regrou many of the communit, former communit or etreme left artie lu the ordic Green eft artie from candinavia Then, from left to right there i the Party of the Euroean ocialit (PE), the Green and allie regrouing regional artie motly from ain, but alo from Wale, the liberal EDR (Euroean iberal Democrat and Reformit), the Euroean Peole Party (motly hritian Democratic and conervative artie)(epp-ed), Eurocetic gaullit and rightit grou and non affiliated The name of the grou change over time acro legilature and ometime within legilature reflecting defection from ome national artie Table 1 give the arty grou with the denomination they had over time In ubequent table, we will ue the common abbreviation lited in the econd column to the left of Table 1 IERT TABE 1 It i not by coincidence that MEP are ranked according to ideology The arty grou to which they belong have a real eitence and internal organization The EP allocate budget to arty grou Each grou ha a hair, a ecretariat and taff working 9

10 for them Member of Grou meet in Bruel and during lenary eion in trabourg to make joint voting deciion Grou have whi who check the attendance and voting behavior of grou member even though grou have limited mean to anction their member uch tructure do not eit at all for the national delegation in the EP The quetion of whether voting in the EP haen along ideological or country line i very imortant in term of undertanding both the role of the EP within the Euroean intitution and alo in term of thinking about the imact of enlargement If coalition formation in the EP were to haen along country line, then the EP would be irrelevant for olitical deciion-making within the E Indeed, the majority rule in the ouncil (qualified majority or unanimity) are more tringent than thoe in the Euroean Parliament If coalition formation in the EP would take lace on the bai of country interet, like in the ouncil, then vote in the EP would never rereent a contraint in deciion-making in the E A meaure that can be acceted by the ouncil would alway be acceted by the EP and a meaure that would get a majority in the EP would not necearily a the hurdle of the ouncil (ee eg Bindeil and Hantke, 1997; Turnovec, 1996) In a Euroe of 27 countrie, if EP coalition are baed on country grou, coalition formation in the EP would become very unredictable and oibly chaotic but the ame combinatoric of oible coalition hould alo aly to the ouncil, and even more o given the higher majority threhold in the ouncil However, if coalition formation in the EP occur along the traditional left-right dimenion, even after enlargement, then deciion-making (in term of danger of aralyi in deciionmaking) in the EP hould not uffer from enlargement wherea deciion-making in the ouncil will, at leat under the ice rule Thi hould have a oitive effect on the legitimacy of the EP A well-functioning EP may eercie ubtantial reure on the ouncil to adot certain deciion uorted by a majority of MEP, and oibly enjoying uort in the oulation If on to of that, the EP i given more ower (etenion of co-deciion, oibly more ower in the nomination of the ommiion), then thi hould have a tabilizing influence on E intitution after enlargement and mitigate the danger of aralyi in deciion-making An imortant normative quetion i however the following: how imortant i it to revent aralyi in deciion-making? Indeed, much of the literature on Euroean 10

11 integration take the oint of view that too tringent majority rule are negative (ee alo the analyi in Baldwin et al, 2001) From the oint of view of economic theory, unanimity rule ha motly virtue Even before Buchanan and Tullock (1969), economit undertood that unanimity rule enure areto-imroving deciion, ie deciion where no interet i hurt comared to the tatu quo Thi inight would eem at bet artial for oberver of the E deciion-making roce Aghion and Bolton (1997) have made an imortant te in imroving our undertanding of the trade-off involved under variou majority rule and their analyi alie very well to the Euroean deciion-making roce The tarting oint of their analyi i that contitution are incomlete contract They cannot foreee e ante all oible contingencie Relaing the contraint of unanimity voting ha the imortant effect of reducing the holdu ower of veted interet who tand to loe from efficient deciion adated to change in the tate of the world, eecially when comenation of thoe grou entail ubtantial inefficiencie On the other hand, by relaing the unanimity contraint, one increae the likelihood of tyranny of the majority hurting minority interet In the contet of the E, it i very difficult to tell what the otimal uermajority hould be Mot oberver would robably ay that majority rule in the ouncil give too much holdu ower to individual countrie wherea we cannot really think of clearcut eamle of tyranny of the majority in the E contet In any cae, whatever reform will be decided in relation to majority hurdle in the ouncil, a more owerful EP in an enlarged Euroe would likely to function reaonably well rovided coalition occur on a left-right dimenion and rovided voting behavior i more or le redictable due to a well-functioning Euroean arty ytem We now turn to the evidence 3 PARTY OR OTRY OYATY? ince it i very imortant to determine whether Euro-deutie vote more according to their ideological loyalty (ie with their arty grou) or according to national interet (ie with other Euro-deutie from the ame country, we ued different meaure arty and country loyalty from different angle 11

12 21 The coheion inde The main meaure we ue i the coheion inde I i for a given grou i of voter We ued thi meaure in order to comare the coheion of voting within a arty grou veru within a country grou ote by Y i the number of Ye vote ereed by grou i on a given vote, i the number of o vote and A i the number of abtention The coheion inde i defined a follow: I i = ma{ Y, i i, A } i [( Y + + A ) ma{ Y,, A }] i ( Y i + i i i + A ) i i i i (1) Thu, for eamle if deutie of a Party (country) cat 100 ballot on a given vote and if all vote Ye, then the coheion inde will be equal to 1 However, if they are comletely divided and 50 vote Ye and 50 vote o, then the coheion inde will be equal to 0 4 The coheion inde i a variant of what i uually called the Rice Agreement inde following Rice (1928) Table 2 and 3 give the mean coheion inde of artie and countrie and their tandard deviation (in arenthee) reectively for the third and for the fourth arliament IERT TABE 2 and TABE 3 We clearly ee that coheion i much tronger for artie than for countrie The coheion inde for artie ha an average of 842% in the third arliament and of 822% in the fourth arliament while the average for countrie i reectively 744 and 688% There i thu a 10% oint difference in coheion Moreover, the tandard deviation i alo lower for artie than for countrie howing that the variability in coheion i much lower for artie than for countrie The coheion of countrie remain generally omewhat higher than that for the EP a a whole (58% in the third arliament and 52% in the fourth arliament) Thi reflect to a certain etent the fact that many national delegation are 4 The coheion inde varie between 0 and 1 in almot all cae ot however that if there are many of abtention vote, the inde may be negative Indeed, in a few cae we have negative indee for two artie (PE and Anti-Euroean): 19 out of 6747 cae for ocialit and 250 out of 6474 for anti- Euroean 12

13 dominated by one arty 5 imilarly, the large coheion for Greece and uembourg can be elained by the fact they are mainly affiliated to the two larget olitical grou ote however that the coheion inde for France i lower than that for the EP! French rereentative are thu the mot divided country in the EP, a fact that maybe run counter to the intuition of many ote alo that coheion i the highet among the arty familie reflecting the uual olitical cleavage in advanced democracie: ocialit (PE), chritian-democrat and conervative (EPP), green (GR) and liberal (EDR) ote however that the Green and the Radical Alliance (RAD) have the highet coheion in the Fourth arliament Thee coheion indice may not give an adequate reflection of overall coheion becaue of the relatively large hare of unanimity or near unanimity vote Many vote in the EP reflect the oition of the Parliament relative to ommiion rooal or ouncil deciion or even Parliament initiative Thee vote a often by unanimity or near unanimity Indeed, deite rereenting different ideologie and countrie, MEP generally hare the objective of increaing the ower of the EP There i thu a concerted effort in the EP, whenever oible, to aear united in front of the ouncil and the ommiion, eecially in cae of a conflict between thee three bodie Figure 3 give the ditribution of the majority ize of the EP for the third and fourth arliament The ditribution i bimodal with a local mode around 5 in the third arliament but a high concentration toward 1 In the fourth arliament, there i a high mode of 085 and much more weight at lower majority ize howing le conenu, and uggeting a move toward more claical legilative dynamic IERT FIGRE 3 In order to mitigate the effect of thee unanimity vote on our average coheion inde, we calculated a weighted coheion inde where the coheion inde i divided by the oberved majority in the Euroean Parliament multilied by two Thu a vote with a 5 The firt at the ot electoral rule ent motly K abor arty rereentative to the EP 13

14 narrow majority of 50% get a weight of one and a unanimity vote get a weight of ½ When comuting the average of thoe weighted coheion indee, we divide them by the average inde a erfectly coheive arty would obtain o that a erfectly coheive grou would till get an average of 100% IERT TABE 4 AD 5 A one can ee from Table 4 and 5, the change with reect to Table 1 and 2 are not very big Two thing are worth noting however Firt of all, the difference between coheion of artie and of countrie i now more ronounced The coheion of countrie i clearly lower wherea the coheion of artie remain baically unchanged Moreover, in contrat to Table 2 and 3, the coheion of artie tend to be tronger in the fourth arliament a comared with the third wherea for countrie, the ooite can be oberved The lower coheion for countrie how that the latter ick u the effect of unanimity or near unanimity vote wherea the coheion of artie remain trong even when lacing more weight on vote with narrower majoritie till another way of looking at the ame roblem i to look at coheion for divided vote in the EP, that i vote where the effective majority wa not cloe to 100% Table 6 and 7 how the coheion indee for the vote when only vote with majority lower than a certain threhold We lowered the threhold gradually from 100% to 65% IERT TABE 6 AD 7 Here the reult are even more triking We ee that the coheion of artie remain roughly the ame and hardly fall The Green in the Fourth Parliament are even more coheive for divided vote! wherea the coheion of countrie fall harly, eecially in the Fourth Parliament, thu reinforcing the concluion that the coheion of countrie move with the coheion of the EP Interetingly, in the Fourth Parliament, the K and Ireland remain more coheive than other countrie when we throw away vote with high majority Germany ha the lowet coheion for vote with le than 65% majority both in the Third and in the Fourth Parliament 14

15 One immediately wonder whether the coheion of artie i the ame acro iue We thu divided the vote in 11 large categorie of iue uing the title of the vote a reorted in the Minute of the lenary eion To give an eamle, the rooal on ovember 17, 1998 include Emloyment in Euroe, ocial action rogramme , linical trial, and ombating fraud /Protection of Euro among many other iue We made ue of thee ubject title to claify roll-call vote in 11 large categorie together with one category that contain 378 vote difficult to claify Each large category wa in turn divided in ubclae A difficulty here i that a large number of vote can be claified in many different way For intance the lat vote on ombating Fraud can be claified in egal or in other categorie uch a ecurity Another difficulty relie in the fact that many different iue and conflict may be hidden under a ame heading A we will ee below, even a articular voting equence on one reolution or bill, entail many different iue and correonding voting coalition With a very large number of vote, much of the noiine related to vote-ecific idioyncracie i filtered away Thi i le the cae with diaggregation into maller number of vote, even under one homogeneou heading Table 8 lit thee iue a well a the number of vote covered in each category Table 9a and 9b give u the coheion er broad iue in the Third and Fourth Parliament IERT TABE 8 and 9a,b Table 8a and 8b how that there i no triking difference in coheion when dividing vote according to iue The table for the third Parliament nonethele tend to indicate that while ocialit become more coheive on Human Being iue (human right, refugee, ), conervative (EPP) and liberal (EDR) are coheive on ecurity iue and green and leftit are coheive on nature and environmental iue The table alo indicate that radical (RAD) and Gaullit (GA) vote coheively on iue related to ervice For the fourth Parliament, the above finding remain true for conervative, green and radical but not for ocialit and liberal who become now alo more coheive on ervice ote that the mean coheion inde i the lowet for Drug and the egal ytem in the Third and Fourth Parliament wherea it i the highet for ervice and ecurity, again in both Parliament 15

16 It i ueful to comare the evidence roduced in thi ection to that of a national Parliament where both ideology and regional identity lay a role Belgium fit that definition retty well Among national arliament the Belgian arliament i the mot comarable arliament becaue on the one hand there are two communitie or nation and everal olitical familie There are ditinct regional Flemih and French-eaking ocialit, hritian Democratic, liberal and Green Partie and there i no ingle national arty The ocialit family i thu comoed of two ditinct regional ocialit artie Party familie in Belgium are thu concetually cloe to arty grou in the EP ike arty grou in the EP, thee olitical familie alo include different tendencie going from left to right and including regionalit Of coure, the Belgian arliament i much maller than the EP Moreover, unlike, the Euroean nion, Belgium ha a genuine arliamentary ytem oheion of arty familie hould reflect coheion of government coalition ince coalition are baed on arty familie, never on artie only 6 According to recent theorie of olitical regime, artie forming a coalition in arliamentary regime are redicted to be very coheive (Huber, 1996; Diermeier and Fedderen, 1998, Peron et al 2000) IERT TABE 10 Given the relevance of the comarion with Belgium, we were able to collect more than 2000 roll call vote from the Belgian arliament during the 49 th legilature (ie from 30/06/1995 to 18/12/1997) 7 Our ummary reult are lited in Table 10 We can immediately ee from Table 10 that arty coheion i much tronger among Belgian arty familie, confirming the rediction about coheion of coalition artner in arliamentary ytem Ecet for the iberal, all other arty familie have a coheion inde ranging from 95% to more than 98% The average coheion inde i 93% The coheion inde for 6 Thu for eamle, if the Flemih Green are in the coalition, the French-eaking Green mut alo be art of it even though they could be diened with to form a minimum-winning coalition, which i the cae today oheion of familie conit mainly of that commitment to articiate together in a coalition or otherwie to be in the ooition 7 urriingly, it i very difficult if not imoible to collect roll call vote for national Parliament like Germany or France uch data are more readily available in newer democracie uch a Poland and the zech reublic ee eg oury et al (1999) and Mielcova and oury (1998) 16

17 iberal i only 73%, ie 20 ercentage oint lower than the average coheion inde ote that the iberal were then in the ooition and that the coalition wa comoed of the ocialit and hritian democratic familie The very trong arty coheion tand in tark contrat to the coheion on a regional or linguitic bai which i very low The average coheion inde for region i one third of that found for olitical familie ote alo that the coheion for the arliament a a whole i higher than the coheion of Flemih MP ote however the lower coheion of the Parliament comared to the EP IERT TABE 11 It i alo intereting to comare the Belgian Parliament and the EP on ome key tylized fact Table 11 how the average ye, no, abtention and reence rate in both Parliament The mot triking difference i the generally high abentee rate in the EP ote alo that the average ercentage of no vote i higher in the Belgian Parliament than in the EP ooking a bit further, we comared the ercentage of rooal ut to vote that were acceted and rejected We found an accetance rate of 40% in the Belgian Parliament comared to 587% in the EP Thi fact i urriing becaue it contradict the view one can eaily derive from game-theoretic model that only vote that are likely to a are ut forward Thi i not true in the EP, but even le true in Belgium which i a full-fledged Parliamentary ytem! IERT TABE 12 We alo comared coheion in the EP with coheion in the ongre Table 12 give the coheion and relative coheion inde for ongre between 1991 and 2000 At firt ight the coheion among democrat and reublican in ongre aear lower than that of Party grou in the EP Indeed coheion i uually in the 70% while it i in the 80% and higher for EP grou, and above 85% for the two main artie, the PE and EPP However, vote are alo much more olarized in the ongre If one look at the relative coheion indee (ie coheion of artie relative to the coheion of the legilature), then one find that coheion of artie in the ongre i omewhat higher, on average above 17, wherea PE and EP have reectively a relative coheion of 143 and 14 for the third Parliament and of 157 and 156 for the fourth Parliament Moreover, coheion ha been increaing in the EP wherea in the ongre, it ha been declining ince 1996 and reached level comarable to thoe of PE and EPP for the 17

18 fourth Parliament oheion in the EP and in the ongre are thu quite comarable The comarion with the i eecially relevant becaue the i a reidential ytem The theorie of democratic olitical regime cited above redict that elected rereentative in a reidential ytem vote le coheively than in a arliamentary ytem ince the E i not a arliamentary ytem (but not a reidential ytem either), it coheion i not redicted to be a high a in arliamentary ytem However, it i overall reauring to ee that arty coheion i quite comarable to that oberved in the Houe of the mot mature reidential ytem in the world Overall, the evidence from the EP how u that coheion among artie i relatively trong and i getting tronger over time while coheion among countrie i not trong and i getting weaker a the ize of majoritie in the EP tend to decreae Thi indicate that arty grou tend to act coheively like in normal democratic legilature Thi alo ugget that coalition among artie hould form according to ideological affinity We now turn to that quetion 32 oalition Given that the E i not a arliamentary ytem, there i no coalition government We hould therefore not eect the tyical attern of legilative voting of arliamentary ytem where artie of the coalition vote together and artie of the ooition likewie oalition are likely to form on a vote by vote bai everthele, to the etent that artie are coheive a een above, it i intereting to ee which artie tend to vote together more often or le often I there in articular a left-right attern in coalition formation acro vote? If thi i the cae, then voting of left wing artie hould be oitively correlated with each other and negatively correlated with right wing artie To do thi, we meaured for each vote the intenity of uort within a arty grou for a majority deciion, ie the ercentage of member who voted in favor of (againt) the deciion when it wa adoted (refued) and when a majority of the grou voted in favor (againt) We call thi the intenity of uort In Table 13, we reent the correlation coefficient of the intenity of uort for bill acro artie 18

19 IERT TABE 13 We clearly ee a oitive correlation between left-wing artie, a oitive correlation between right wing artie and no correlation or negative one acro the ideological ectrum In the third arliament, vote of PE are motly correlated with thoe of the EFT and RAD and o are the vote of GR Vote of EPP are motly correlated with thoe of EDR and two other right-wing olitical grou, GA and ATI-E One thu ha the broad left (ocialit, Green, eft and Radical) on one ide and the broad right on the other (conervative, Gaullit, liberal and nationalit) ote however the very centrit role of the liberal whoe vote are correlated to thoe of ocialit but not to the other more left-wing artie A imilar icture emerge in the fourth arliament, ecet that now the vote of PE are more trongly correlated with the vote of GR and le with the EFT Thi reflect recent tendencie of ocialit artie to move from labor artie to rainbow-tyle artie ote alo that correlation are not tranitive Thu, PE vote are more correlated with thoe of GR and the latter more with thoe of EFT However, vote of PE are not more but le correlated with thoe of EFT Thi ugget that coalition formation i not unidimenional, a oint we ick u below when analyzing the dimenionality of coalition formation Vote of EPP are now alo more trongly correlated with GA and le with EDR The centrit oition of the latter i confirmed ote alo that vote of GR are more correlated with thoe of EFT, confirming tendencie toward rainbow coalition on the left For comarion, Table 14a and 14b looking at country correlation do not how any ecific attern tending to ugget table grou of coalition o negative correlation can be oberved nor i there tronger correlation between ecific grou of countrie The correlation ick u motly the change in the ditribution of the ize of the majority ote that our meaure of coalition formation (the correlation coefficient) are baed on all vote A already mentioned, for a large number of the EP vote the oberved majority i large When electing the vote with narrower majoritie (eg for vote with majority le or equal to 70%) the reult of table 13are reinforced 19

20 IERT TABE 14 Another way of looking at coalition i to ak what are the mot frequent majority coalition that emerge By majority coalition, we mean a ituation where a majority of member of a arty or a country voted in favor of a deciion and it wa adoted or againt it and it wa refued The mot frequent coalition, not urriingly given the high number of unanimity vote, wa the coalition of all artie or countrie everthele, a triking difference can be oberved in Figure 4 and 5 Figure 4 look at the frequency of the number of artie in the majority We ee that in the third arliament, the mode i 7 while in the fourth arliament it i 6 While thi number include all oberved coalition of 6 artie, the figure neverthele ugget ome tendency toward arty olarization in the fourth arliament Figure 4 tand in contrat with Figure 5 where one ee that the mot frequent coalition of countrie i the one that include all countrie The difference between Figure 4 and 5 once more clearly ugget that the dividing line in Parliament i ideology and not country interet IERT FIGRE 4 AD 5 33 Eamle of vote The tatitical analyi give a comrehenive icture that i more accurate than the inference one may draw from looking at ecific vote It i neverthele ueful to comlement the tatitical analyi by ecific eamle to get a feel of coalition formation and coheion with articular vote or voting equence We firt tart with ome of the mot dramatic vote in the EP that were given wide ublicity in the Euroean media, ie attemt to bring down the ommiion by a vote of cenure The EP ha alway had the right to cenure the ommiion but it i only in recent year, under the anter commiion that it ha tried to ue that right The firt attemt at bringing the ommiion down wa in relation with the BE or mad cow dieae in February 1997 After voting in Augut 1996 a reolution very critical of the ommiion attitude toward the BE crii, the EP reconvened in 20

21 February to dicu the reult of the official inquiry of the ommiion handling of the BE crii A reolution wa tabled uorted by PE, EPP, EDR, GR and GA The reolution intruct the ommiion to follow u eriouly the concluion of the ommittee by ovember and threaten otherwie with a vote of cenure after that The reolution wa adoted by near unanimity (422 for, 49 againt and 48 abtention) The vote of the motion had been ut on the agenda before the vote of cenure rooed by Belgian ocialit Haart The vote of cenure neverthele attracted a urriing 117 vote in favor, 326 vote againt and 15 abtention Thi wa neverthele a very lit vote without a majority in any grou or country coming u in favor everthele, there wa a ubtantial uort for the motion, with EDR and EFT voting more or le one to one in favor while PE, EPP and GR voted more or le one to three in favor ote that 15 out of 47 K rereentative voted in favor Only the deutie from uemburg voted againt by olidarity with anter Thing became much more eriou in January 1999 with the reon-marin crii, following the reort of mimanagement and diverion of fund in the office of thee two ocialit ommiioner The occaion wa the refual of dicharge of the 1996 budget by the Parliament becaue of thoe candal Two reolution were on the table, one ut forward by EPP (Marten), EDR (o) and GR (Aelvoet), the other one wa ut forward by the ocialit, the left and the radical The former elicitly aked for the reignation of reon and Marin, the latter called for a committee of eert to eamine quetion of fraud and call for change in the adminitrative culture of the ommiion and other meaure The firt vote taken wa on a requet of GR to vote on the motion of cenure before voting on the variou reolution The Green (Telkämer) argued that the vote wa the mot imortant one and that vote on the reolution firt would divide the EP The ocialit (Green) oke againt by invoking at rocedural eerience, a in the BE motion of cenure The motion wa rejected by 429 vote againt 99 and 14 abtention The Green voted en bloc in favor, the non affiliated and anti-euroean voted with big majority in favor and EFT voted the motion with a narrow majority It wa rejected by a large majority of member of other artie By contrat, one ee that there i no clear attern in the country vote A majority of all country grou rejected the motion, following roughly the attern of the EP a a whole 21

22 IERT TABE 15 The firt reolution (aking elicitly for the reignation) wa then ut to vote aragrah by aragrah and wa rejected in 10 vote The attern of thoe vote (available from the author) wa quite clear The EDR, GR and EPP voted in favor with uort from other ecet for PE and RAD who were able to block the vote by a narrow majority In ome vote EFT ided with PE The country attern of vote i le clear but neverthele omewhat indicative A narrow majority of deutie in all countrie voted in favor but a majority of MEP from the K, Italy, Greece, ain and uemburg voted againt Thi robably motly reflected the arty comoition of vote ecet for ain and uemburg where olidarity with Marin and anter layed a role There wa no French olidarity vote with reon The Italian and Greek vote may indicate a different olitical culture where the criticim to reon and Marin eemed benign and more art of the day-to-day realitic adminitrative culture in thoe countrie The econd reolution wa then ut to vote in three te Here the arty comoition of the vote wa ubtantially different A majority coalition voted in favor with the PE, EFT, RAD and GA The EPP wa lit and o were the liberal Here again, the country comoition reflect motly the arty comoition but alo with a Mediterranean uort and a ordic ooition The reult can be een in table 16 IERT TABE 16 After the adotion of thi reolution and an intervention by anter, the vote on the motion of cenure then took lace eader of arty grou ereed the oition of their grou Marten, rereenting EPP aid that EPP deutie hould vote with their concience, de facto not imoing any diciline o for EDR, Aelvoet for GR, Bonde for ATI-E unambiguouly called for cenure Paty for GA announced a lit vote from hi grou with the French uorting cenure and the Italian, Irih and Portuguee againt Puerta for the EFT announced that a majority of their grou would vote the cenure alumière for RAD announced they would vote againt The vote followed eactly the announcement of the grou leader but coheion wa le than erfect in the 22

23 cae of the liberal The motion wa voted down with a narrow majority The reult of the vote can be een in table 17 IERT TABE 17 All in all, we ee with one of the mot dramatic voting eion in the hitory of the EP that arty grou olitic layed the dominant role everthele, arty diciline wa clearly lower than i the cae in normal arliamentary olitic One hould emhaize that thi dramatic eamle i not one where the traditional left-right dimenion i eected to emerge Thu, the Green and the iberal are generally very trongly uortive of tranarency and clean government, which i much le the cae for the ocialit ountry interet neverthele did not lay a ubtantial role ecet robably for the MEP from uemburg and to a maller etent the MEP from ain and Italy MEP from Italy voted with large majoritie againt all motion howing aarent ditate for thi kind of ublic criticim of mimanagement of fund ote that no country olidarity from France wa oberved We alo looked in detail at many articular event of voting equence which we only briefly reort here We looked at imortant vote in 1998 on coheion fund, reform of common agricultural olicy, tructural fund and on fiherie one of thee vote were fundamental in the ene that they rereented fundamental change to eiting olicie They tended to be conenual and were mot often the reult of comromie worked out between the imortant arty grou Interetingly, in thoe cae, one ee many amendment rooed by maller artie that are tyically rejected Many vote alo concern articular formulation of entence and are hardly conequential Quite a few vote have a low arty and country coheion, eecially in the area of AP Thi imreion i however not validated by the broader tatitical analyi Another caual imreion i that MEP tend to amend ommiion rooal by adding roenvironment entence, and emhaizing general goal like emloyment creation and other ocial objective without really changing the ubtance of ommiion rooal A articularly heated, though not very conequential debate took lace when the Renault lant wa cloed in Vilvoorde in Belgium to be relocated in France and when the worker came to the EP to rotet Aart from the taunch uort received by worker on the left and the outright hotility on the right of the hemihere, there were vote on a 23

24 reolution recommending the ommiion to rooe legilation on job tranfer by multinational enterrie and recommending and aking for an evaluation of the E directive on collective redundancie everal right wing amendment to the initial reolution were voted down Moderate amendment were acceted with GR, RAD and EFT voting againt The final reolution wa uorted by the left and ooed by the right The EPP voted for the reolution with 10% defection in it rank Mot vote in thi eiode were clearly on the left-right dimenion and artie howed relative coheion o recie attern could be detected in the country vote Another cae of a articularly divided vote wa a roject of a reolution to evaluate the Een ouncil After variou amendment where vote were clearly between the left and the right, the rooed reolution wa rejected when the ocialit droed their uort to the EPP and the liberal The bone of contention wa a difference in evaluation in the imortance of the iue of fighting unemloyment We looked at the debate on the 1999 budget which wa aroved by a large majority The EP ha limited ower in changing the budget rooal of the ommiion everthele, our imreion i that it ue the little ower it ha to focu on iue like emloyment, education, the environment, regional develoment and convergence rather than agriculture or fihing orbett et al (1999) who are thorough inider to the EP make the ame obervation It eem alo that the EP lay more a role of a check on the ommiion ending lan rather than that of a big cah-ender Variou ending item get ent beck to the ommiion to ak for clarification before getting EP aroval 34 atial analyi The evidence we howed o far only allow to ick u a ingle dimenion to the olitical ace in the Euroean Parliament However, one of the main finding of the emirical tudie of the E olicy making i that there are at leat two dimenion to the E olitical ace (eg Hi, 2001) Thee tudie ue the atial framework develoed by Hotelling (1929) and Down (1957) for ue in the electoral arena and by Black (1948) for 24

25 ue in legilative analyi The atial model ha become the workhore theory of modern legilative tudie In articular, the atial model of voting ha been etenively alied to the analyi of roll-call vote in the ongre (Poole and Roenthal, 1997; Heckman and nyder, 1997) The goal of the atial model of roll-call voting i to etimate the ideal oint of each legilator in a multi-dimenional olicy ace a well a the ye and no location or the cutting line for each vote on different dimenion of the olicy ace The atial model aume that each voter (legilator i) ha an ideal oition or bli oint ( i ) on different dimenion of a multi-dimenional olicy ace It alo aume that in thi ace each roll call (vote j) i rereented by two oint ( z, and z j, no ) correonding to the olicy conequence of the voting outcome (ie the ye and no outcome) The legilator utility ha two comonent, a determinitic and a tochatic comonent For eamle, legilator i' utility when voting on roll call j can be rereented by u ( i, z j ) = f ( δ ( i, z j )) + ε i, j where δ z i ) i the ditance between the ideal oint of legilator and the vote outcome (, j The firt term on the right-hand ide of equation (1), the determinitic comonent of the utility, i a negative function of the ditance between the legilator ideal oint and the ye and no' outcome of the vote Therefore, the legilator i aumed to chooe the vote (ye or no) cloet to hi or her ideal oint alternative The econd term on the righthand ide of equation (2) i the tochatic comonent or the error term It i a random variable and cature the effect of idioyncratic hock ecific to legilator i when voting on roll call j, (ie the omitted attribute affecting the legilator choice) There i a big debate among olitical cientit about whether the method urely ick u legilator reference or whether arty oition are alo icked u in the analyi We need not be concerned with that quetion here a we are uing the Poole and Roenthal methodology mainly to cale voting behavior in a multi-dimenional ace j ye (1) To etimate the atial model, one mut define a ecific functional form for the determinitic comonent of the utility function and a robability ditribution for the error 25