Electoral Systems in France

Transcription

1 Electoral Systems in France BY BYRON CRIDDLE DEBATE about electoral systems in the established liberal democracies is confined largely to those few cases where the system, for one reason or another, is held to lack legitimacy. Thus, in Britain in the 1980s, the 'first past the post' system came to be identified with the re-election, twice, on a minority (barely more than 40%) of the popular vote, of a parliamentary majority used contentiously for the introduction of radical policy changes. Meanwhile in France, a tradition of tampering with the electoral system as a means of marginalising opponents in a polity traditionally lacking in consensus has ensured that no electoral system has enjoyed complete legitimacy, but has been seen rather as an instrument for securing partisan advantage. Students of politics judge electoral systems essentially by two criteria: fairness, in terms of representativeness; and effectiveness in the provision of good (in the sense of 'strong' and 'stable') government. Advocates of electoral reform preoccupy themselves with the former criterion, seeing as important the election of a parliament as a microcosm of the nation, to the extent of securing a proportionality between votes cast for parties and seats won. Proportional systems are seen also as more likely to ensure greater representation, within parties, of women and ethnic minorities. Equally, the invariable outcome of proportional electoral systems is multi-party government, which may or may not meet the second criterion of strong or stable government. If such government is thought to be more likely from an executive comprised of one party, a non-proportional, or majoritarian, electoral system is held to be preferable. Majoritarian systems, involving variations on the 'first past the post' system as used in Britain and France and the election of a parliament from which smaller parties tend to be excluded, place a greater value upon 'effectiveness' (defined in terms of single-party government) than upon 'fair' representation. However, whilst it is clear that proportional systems do by and large ensure 'fair' representation, it is rather less clear whether majoritarian systems are significantly better than most proportional systems in the provision of effective (strong and stable) government, unless strong government is held to be incompatible with the inter-party bargaining required in multi-party governments. A further uncertainty surrounds the relationship between the electoral system and the party system. Majoritarian electoral systems appear to discourage those parties consistently underrepresented in seats and Oxford University Press

2 Electoral Systems in France 109 conversely confer greater legitimacy upon those parties whose governing function is consistently reaffirmed. Electoral systems, equally, may influence the way parties evolve alliance strategies and affect the extent to which they either present distinct ideological profiles or pursue blander catch-all strategies. 1 It is equally clear, however, that a party system may owe more to a nation's political institutions, traditions and social cleavages than to any given electoral system, and here a most salient example is the electoral system used in France for most of the parliamentary elections before 1940 and for all but one since The system comprises the election of deputies by two ballots in single-member constitutencies, with election secured either at the first ballot by a candidate obtaining an absolute majority of votes cast (provided that the votes received comprise 25% of the registered electors) or at the second ballot by a candidate polling a relative majority of the votes cast. The system was associated with very different political outcomes before 1940 and after Both pre-1940 and post-195 8, a multi-party system made the first ballot no more than a clearing house for a second-ballot run-off between, invariably, two candidates representing rival alliances. But whereas since 1958 second-ballot contests have been ideologically polarised into a left-right confrontation of national salience across all constituencies, before 1940 second-ballot contests were bipolarised electorally but not politically, with a large centre party, the Radical Party, representing the left in some areas but opposing it in others. The result was the election of many Radical deputies but with locallyvariable mandates, no parliamentary coherence, and unable to contribute to stable government. Bipolar in appearance at the second ballot, party politics remained multi-polar and fragmented in fact. Yet an identical electoral system in the Fifth Republic has been associated with the reduction of a still multi-party contest into a secondballot bipolarity of left and right, producing stable coalitions of parties cohering for the duration of a parliament; a nationally-salient and lasting bipolarity. The key ingredient in determing the contrasting outcomes of similar electoral systems is the presidentialisation of politics initiated by the de Gaulle presidency and by the institution of a directly elected de facto executive presidency as the regime's dominant institution with a two-ballot presidential electoral system involving a second ballot mandatorily bipolarised into a straight fight. With the de facto head of government elected by universal suffrage in a two-cornered contest bifurcating the country into two camps, it took little time for parliamentary elections to mirror the nationally bipolarised presidential confrontation. In considering the question of electoral system impact, it may thus be said that it is the presidentialisation of politics and the specific mechanisms of presidential election law which have determined the consequences of parliamentary electoral systems and set the tone for the electoral politics of the regime. Two-cornered presidential contests

3 110 Parliamentary Affairs imply a Downsian bid for the centre ground and presidential parties pursuing catch-all strategies; the re-election of President Mitterrand in 1988 was a victory for his pursuit of the median voter over his opponent Chirac's bid for votes not at the centre but at the (right-wing) extreme. It is the presidentialisation of politics, and the contribution to that of presidential electoral laws, that did most to bipolarize and de-dramatize political competition at other, parliamentary and local, levels. Even those who doubt the stability of bipolarity and detect a signficant reduction during the 1980s of the 'nationalising, moderating, disciplining and coalition-encouraging pressures of the first two decades of the Fifth Republic', 2 still locate the source of change in institutional forces, amongst which are changes in the electoral system in the 1980s. Electoral System Change In France, there is every evidence of the politician's belief in the importance of electoral systems as political weapons. As Yves Meny 3 has put it, deriving from a revolutionary tradition and strong ideological polarization, there has been a reluctance in France to accept the idea that defeat of a party in power leads ineluctably to accession to power by opposition parties, and thus, in consequence, a constitutional history punctuated by manipulated electoral systems and, if less frequently, by coups d'etat. Nor has there been an ease of relationship between the branches of government, but rather an imbalance strongly favouring the executive in its relations with both the legislature and the judiciary. In addition to a reluctance to implement the separation of powers, there has been a tradition of resisting any territorial division of powers that would amount to anything beyond purely administrative decentralization. In this context, tampering with the electoral system is merely one piece of evidence of a 'flawed' liberal democracy. Between 1870 and 1940 the parliamentary electoral system was changed seven times, and between 1945 and 1988 five times. The rules of the electoral game have been changed as frequently as this on the assumption that to do so would be to serve the interests of those political forces in the ascendant at the time. The Fifth Republic, however, is by far the most stable modern French regime, second only in longevity to the Third Republic, whose seven decades were in any case marred by serious poltical stress. The stability of the Fifth Republic has been mirrored in the essential continuity of its parliamentary electoral laws, which have alternated only once (in 1986) to a system of proportional representation from the two-ballot single-member majority system which, dating from 1852, has in any event been the most consistently applied electoral system since the 1870s. It was used in thirteen of the seventeen parliamentary elections up to 1940, and in eight of the thirteen between 1945 and Moreover, the presidential electoral system, already alluded to as formative, has also remained unchanged in the five presidential elections since Against this

4 Electoral Systems in France 111 background of relative continuity, the reversion for one general election to a system of proportional representation of the kind used in the 1940s is essentially an aberration, even if the motives behind the reversion to PR by President Mitterrand fell firmly within the tradition of partisan self-seeking. The objective was to minimise anticipated heavy Socialist seat losses and to slow down and balkanise the Right by putting the National Front onto the parliamentary map. Mitterrand's recourse to proportional representation duly achieved these ends, notably by giving 35 seats to the National Front, thereby helping to legitimate a party which would serve as an objective ally of the Socialists through its disruptive impact on the parties of the orthodox right. Recourse to proportional representation was rightly seen as 'a relatively successful exercise in electoral manipulation'/ Whilst not inflating electoral support for the National Front, by converting votes into seats, albeit insufficiently proportionately (with 10% of the votes it should have received 56 rather than 35 seats), PR gave it parliamentary status, in the same way that other elections held under various forms of PR at other levels in the mid-1980s (such as municipal elections in 1983 and European Parliament elections in 1984) had helped establish the party. The form of PR introduced in 1986 was similar to that used in 1945 and 1946, with departments forming multi-member constituencies and deputies being elected from ranked party lists in one ballot, provided the lists cleared a threshold of 5% of votes cast. It was a system which combined, through the principle of rough proportionality, a fairer representation of smaller parties (such as the National Front and the Communists) with a bias in favour of larger parties which benefitted from the mode of calculation by which seats were allocated (the highest average, or d'hondt, rule) and from the relatively small size of most of the multi-member constituences. According to one estimate, 61 (or 11%) of the 577 seats were misallocated in the sense that the Right was overrepresented with 45% of the votes delivering 50% of the seats and the Socialist Party overrepresented with 33% of the votes delivering 37% of the seats. 6 The short-term effects of PR in 1986 were clear: the Socialists were rescued from a serious parliamentary rout which would have put in doubt Mitterrand's continued presidential legitimacy and fundamentally destabilized left-right bipolarity. Whilst the left and right coalitions were disaggregated by the removal of the need to form alliances in order to win seats as in the two-ballot majority system, both the Socialists and the RPR-UDF alliance of the right retained their bipolar hegemony. The longer-term consequences of PR could have been more destabilising had the Chirac government not immediately reverted back to the traditional two-ballot majority system on taking office after the 1986 election. But as it is, elements of PR exist in electoral systems at other electoral levels in France and could, conceivably, over time, work as a disaggregating force.

5 112 Parliamentary Affairs National-Local Linkages Despite the traditional subordination of local government to centralized administrative control in France (albeit aleviated somewhat by the decentralisation reforms of the early 1980s), local politicians, and notably mayors, have traditionally enjoyed a prestige unknown in Britain where a certain disdain is reserved for cocked-hatted, chainwearing local worthies and low participation rates are recorded in council elections. In France, by contrast, local elections draw high turnouts; in 1989 a turnout as high as 70% in the municipal elections was the lowest participation rate since the war. Town halls have always 'been a prize for politicians, offering a platform for leadership, a source of patronage, a headquarters for a clientele'. 7 Certain aspects of this account may, however, now be out of date: the national standing of a politician, for example, now often serves to provide a base for local implantation, and not vice versa, as in the case of the Socialist parliamentarians, Jack Lang and Michel Delebarre, whose national careers were given local roots by their election as mayors in There has been talk of the 'presidential' nature of local government, with mayors and department and regional council presidents enjoying considerable scope for local initiative. 8 Following the 1985 law restricting to two posts the number of elected offices a politician may occupy simultaneously, it was not insignificant that scores of politicans chose to retain the combination of deputy and mayor. Mark Kesselman saw mayors as paternalistically above politics, with concern for communal harmony and often presiding over politically heterogeneous lists what he termed 'the ambiguous consensus'. 9 But Kesselman's study was made in the early 1960s, and it has become more fashionable in the seventies and eighties to talk less of the differences between communal and national politics than of the politicization, bipolarizatdon and nationalization of municipal elections, to see local voting as paralleling national political trends. Thus by the 1970s the left's national alliance was reflected in town halls and its gains in the 1977 local elections taken as presaging a swing to the left in the parliamentary elections of Equally, the defeat of the left in the 1983 municipal eleaions was a reflection of the unpopularity of the Socialist government and was seen as a portend for the approaching parliamentary election of 1986, in which the government was duly defeated. Local elections thus came to be seen as barometers of the national political mood and as vehicles for mid-term anti-government protest. This pattern of convergence may have owed something to compatible electoral systems at local and national levels. Traditionally, local electoral systems have not differed significantly from national ones; the Fourth Republic used PR for large towns; the Third and Fifth Repubics employed mostly variations of the two-ballot majority system. Munici-

6 Electoral Systems in France 113 pal elections, for example, have been conducted for most of the time since the 1880s under such a system, with the voter voting in townwide elections for rival lists of candidates, with the winning list occupying all the council seats in a municipal expression of the 'winner takes all' principle. From 1964 until 1982 all municipal elections (held every six years) were conducted under the two-ballot majority list system; only lists clearing a 10% vote threshold could contest the second ballot and, contrary to previous practice, there could be no changing of the composition of lists between the ballots in the larger towns (with over 30,000 inhabitants). This electoral law had a deliberately bipolarising motive, but in the municipal elections of 1965 and 1971 there remained many large town councils with heterogeneous 'centrist' lists, such as Louis Pradel's governing list at Lyon, which had no national salience at all, and many Socialist mayors continued to rely on centrist allies. The 1977 and 1983 municipal elections did, however, see a clearer penetration of municipal politics by the national alliances of left and right, with few large town contests escaping a nationalising process in the wake of two evenly-balanced close-run presidential confrontations (in 1974 and 1981) and (at least in 1977) a programmatic alliance of the left. It has been suggested that such a process of bipolarisation, extended into the town halls, owed more to the national strategies of the left parties and to the right's response to them, than to the electoral system as such. 10 The Fifth Republic's most radical change to municipal electoral law came in 1982 with the introduction of an element of proportionality in the allocation of council seats in all but the smallest towns (with under 3,500 inhabitants). Under this system, the list that wins (whether by an absolute majority on the first ballot or by a relative majority at the second) no longer takes all the seats, but initially only half. The remaining half are then distributed proportionately under the highest average rule amongst all lists that had won at least 5% of the votes at the first ballot, or 10% if the calculation was being made after the second ballot. Before 1982 only the three largest towns (Paris, Lyons and Marseilles) had councils including opposition groups, and that because elections in those cities were held on a sectoral basis, with different sectors able to elect different lists. Under the 1982 law the list polling 51% of the votes obtains half the seats plus half the remainder, giving it a dominant position on the council, but with the remaining quarter of the seats allocated proportionately to minority- lists. Lists with over 5% at the first ballot are also allowed to fuse with other lists between the ballots; in this way diverse elements may also find a way on to the council. Proportional representation reformers decry the crudeness of the proportionality of such a system, yet it does represent a significant step away from the 'winner takes all' practice. It is clear, however, that (as with the PR system introduced for the 1986 parliamentary election) the

7 114 Parliamentary Affairs degree of proportionality increases with the size of the council: the more seats available, the better the chance of lists other than the nearest runner-up list have of securing seats. The municipal election law of 1982 made special provision for Paris, Lyons and Marseilles, with lower-tier arrondissement (district) councils as well as the city council. The latter are elected from each of the arrondissements with the elected councillors sitting on both tier councils. In Paris 20 districts elected a total of 159 councillors, but the minimal degree of proportionality was confirmed in 1989, when the right, with 54% of the first-ballot vote, obtained 141 seats to the Socialists' 23% of the vote and 18 seats, and the ecologists' 8% and one seat. In Paris 20eme, no list won an absolute majority and a second ballot was held, at which the right polled 45%, the left 44% and the National Front 11%. This distribution in votes was reflected in a quite inequitable allocation of seats: right 10, left 3. If 1% of the voters had voted the other way, the result would have been an equally dubious landslide for the left. This sort of electoral system thus has capricious outcomes: with very small vote movements, councils can swing from left to right and back again. Thus in Nantes in 1983 the right polled 50% of the vote and took 48 (or 80%) of the 61 seats, while the left got 13 seats. In 1989 the reverse happened: the left, with 50%, took 48 seats and the right the remaining 13. n It is a long way from Kessleman's 'ambiguous consensus'. Yet the incorporation of even a modicum of proportional representation at municipal level does help sustain minority parties which in the 'winner takes all' system would have been entirely ignored. The 1989 elections, unlike those of 1983 which took place before the significant minor parties (the National Front and the ecologists) had got moving, saw such parties' lists being maintained into the second ballot. The ecologists, with a 9% share of the vote across the larger (20,000 inhabitants or more) towns, contested the second ballot in 22 of the 173 big towns where second ballots were held, and in only 4 other cases did they settle for a fusion with the Socialist list. Likewise, the National Front maintained its presence in 55 second-ballot contests and fused with the right in only two cases. Whereas there were very few threecornered contests at the second ballot in 1983, in 1989 they numbered 83 out of 173, some involving dissident Socialist or Communist mayors leading rebel lists into the second ballot against the official left and right, and winning as at Marseilles and Le Mans. This looked like the unravelling of bipolarity but probably reflected the uncertain situation in national politics where the left alliance had been undermined by Communist decline and by Mitterrand's pursuit of a centrist strategy in 1988, and where the right showed continuing disunity in opposition and lack of a credible presidential leader. It almost certainly reflected disillusionment with the failed consensus represented, in the wake of 'alternance' and 'cohabitation', by the major parties characterised by

8 Electoral Systems in France 115 the National Front as 'the gang of four'. Encouraged by a voting system in local elections which offered to turn votes into seats, electors could register their disaffection from the parties of the failed consensus by voting in an unimportant local election for parties unidentified with that failure. Thus the 1989 municipal elections were less divergent from national politics than might be supposed; they reflected accurately a malaise at national level, whose long-term threat to bipolarity, notwithstanding the potentially disaggregating aspects of a moderately proportional local electoral system, could not be gauged over a short time span. As long as proportional representation was limited to elections at sub-national level, its impact should neither be discounted nor exaggerated. The system of proportional representation introduced for the regional elections of 1986 did, through the larger number of seats allocated to each multi-member consituency (department) offer more hope to smaller parties, and indeed the most novel outcome of those elections was the emergence of the National Front holding the balance in eight of the twenty regional councils. But even proportional French electoral systems have traditionally inclined to disporportionality. Thus the parliamentary electoral system of 1986 was one designed to safeguard the interests of the larger parties, with a majority of seats being awarded to an alliance of right parties securing only 45% of the vote. Nor have motives behind the introduction been disinterested. Apart from the introduction of proportional representation of the 'fairest' kind with party lists competing in a single nationwide constituency for the first European Parliament elections under Giscard d'estaing's presidency, experiments with PR are a feature of Mitterrand's tenure in office. Yet such innovations owe far more to political manipulation (Mitterrand's desire to divide and rule) than to 'Socialist idealism' and have been confined the 1986 parliamentary elections excluded to sub-national level contests. As long as a non-proportional electoral system is retained for the most important presidential election, and provided further use of PR is avoided in parliamentary elections, the existence of proportional representation at less decisive regional and local sites is an affordable risk to the dominance of the larger parties within a bipolarised system of competition. Such a system has, in any event, been associated with the traditional two-ballot electoral system which combines both pulling and pushing characteristics: pulling in the sense of indulging a multi-party tradition at the first ballot but pushing parties into second-ballot alliances in order to maximise their representation, the centripetal forces of the second ballot correcting the centrifugal ones of the first. French electoral systems have been characterised as the least proportional in the established democracies, and yet their main offence has been rather to overrepresent whichever party or coalition of parties

9 116 Parliamentary Affairs happened to be in the ascendant at any given time rather than to misrepresent the electoral mood in a serious way. What French electoral systems have not been is 'proportional' in the manner desired by critics of majoritarian electoral systems, though it is at least debatable in general terms what level of proportionality is either acceptable or necessary. The German 'additional member' system, for example, with its way of topping up from a list system to ensure overall proportionality, has been associated with a party system where, for all but a few years in the history of the Bonn republic, a small hinge party the FDP held the balance between the two larger parties and thereby determined with which of them it would decide to govern. Where the overall goal of the polity is consensus, such a practice is well-suited; proportional representation thus serves, arguably, to create a condition in which the parties, or proportions, come essentially to be indistinguishable one from the other. Such a condition might, however, be thought preferable to that in which a party, under the British majoritarian system, with barely 40% of the vote converted into a large majority of seats, is elected three times consecutively and proceeds to dismantle the consensus. The French may well be fortunate to be in neither condition, nor yet, rhetorically at least, to value consensus. The help of Paul Wilder of the Electoral Reform Society is gratefully acknowledged. 1 V. Bogdanor in D. Butler & V. Bogdanor, Democracy and Elections (Cambridge UP, 1982), pp H. Machin, 'Stages and Dynamics in the Evolution of die French Party System', West European Politics, October Y. Meny, Government and Politics in Western Europe (Oxford UP), pp Significant modifications were however made to the threshold below which candidates were excluded from the second ballot: 5% of the votes in 1958; 10% of registered electors in 1967; and 12Vi% of registered voters in Since the second ballot has been the decisive ballot in 81% of all constituencies, on average, since 1958, a high threshold is an important barrier to small parties and an incentive to forging alliances before the first ballot. 5 A. Cole & P. Campbell, French Electoral Systems and Elections (Gower, 1991), p P. Cole, Representation (Electoral Reform Soceity), Spring P. Williams, French Politicians and Elections (Cambridge UP, 1970), p J. Frears, Parties and Voters m France (Hurst, 1991), p M. Kessdman, The Ambiguous Consensus (Knopf, 1967). 10 A. Cole & P. Campbell, o/>. at., p P. Cole, 'Liberti and fraterrute, but not much egaliti', Representation, Autumn, 1989.