Kluwer, ELECTORAL COMPETITION IN MIXED SYSTEMS OF REPRESENTATION

Transcription

1 Encyclopaedia of Public Choice, Charles Rowley and Friedrich Schneider (editors), Kluwer, ELECTORAL COMPETITION IN MIXED SYSTEMS OF REPRESENTATION Ram Mudambi, Temple University Pietro Navarra, University of Messina and London School of Economics Over the last decade electoral engineers have become increasingly cognisant of three important facts that affect worldwide electoral institutions. First, several established democracies such as Israel, Italy, Japan, New Zealand and Venezuela undertook major electoral reforms. Second, the process of democratisation in formerly communist countries brought about constitutional reforms where the design of new systems of representation is the central concern. Third, mixed-member electoral (MM) systems have been the choice of the vast majority of countries implementing electoral reform. What are MM systems? How do they work? Bogdanor (1983; p.16) pointed out that although the main electoral systems are probably already known, there are undoubtedly many ingenious ways of breeding new combinations. MM systems are one set of such ingenious combinations. These systems are hybrids using both proportional representation (PR) and plurality (PL) formulae for the election to a single representative body. Massicotte and Blais (1999) argue that such hybrids eliminate many of the weaknesses of pure PL and pure PR systems. Their aim is to provide representation to small and diffuse political groups as well as providing incentives to form stable governing majorities. In practical terms, such systems incorporate some preference-aggregating mechanisms based on PR and others based on PL. The PR mechanisms guarantee that all societal preferences, no matter how diffusely held, are represented. The PL mechanisms ensure strong incentives to form mass political formation that can prevent government instability. MM systems are operationalized by allowing voters to cast two simultaneous votes. The first ballot for the single-member district (SMD) candidate a Plurality (PL) ballot. The second ballot is cast for a party for the proportional allocation of seats a Proportional 1

2 Representation (PR) ballot. Shugart and Wattemberg (2000a) describe an MM system as a particular type of two-tier system. In one tier seats are allocated nominally to candidates, and in the other tier seats are allocated to party lists. In other words, in the first tier votes are cast for candidates by name and seats are allocated to individual candidates on the basis of the votes they receive. On the other hand, in the second tier votes are cast for party lists and seats are allocated on the basis of the votes received by the listed candidates. Since each voter has two simultaneous votes available, he or she can express a double preference for a party, by voting for it in the PR ballot and for that party s candidate in the PL ballot. Alternatively, the voter can express split preferences by voting for a party in the PR ballot, but not for that party s candidate in the PL ballot. This split-ticket phenomenon is a unique feature to multi-ballot elections and can significantly influence overall election strategies and results. This phenomenon has been extensively studied in the MM systems literature. Jesse (1990) and Schoen (1999) describe split-ticket voting in Germany as a product of strategic voting to support the formation of a given political coalition. Klingeman (2001), analysing elections in the same country, views the split-ticket phenomenon as a result of voters choosing candidates based on performance rather than party affiliation. According to Banducci et al. (1999), similar strategic motivations seem to guide voting behaviour in New Zealand. Navarra (1997) gives a different explanation for split-ticket voting in the Italian elections following electoral reform. Voters express split preferences in order to diversify the risk of making a wrong choice in a changing environment. Benoit et al. (2002) present a strategic model of voting behaviour in MM systems. They use Italy as a case study and estimate the relative numbers of strategic and non-strategic voters in Italian SMDs for the 1994 and 1996 elections. McAllister and White (2000) found that the individual motivations of voters are the major cause of split ticket voting in Russia principally as a consequence of weak partisanship. However they also 2

3 claim that there is evidence of voters strategic behaviour based on their socio-economic status. MM systems are currently used to govern about one-fifth of the world s population (Massicotte and Blais, 1999). Countries using MM systems differ in the percentages of seats allocated on the basis of the PL and PR ballots respectively (i.e., the PR/PL seat ratio). In democratic countries the PR/PL seat ratio ranges from 64/36 PR/PL split of Georgia to 20/80 PR/PL split of the Philippines and Thailand. However, across countries it is generally true that the PL component is greater than the PR component. There are actually only four countries (Israel, Georgia, Hungary and Venezuela) with a greater PR component in the distribution of seats in the national legislature (see Shugart and Wattenberg (2001) for a definition and typologies of MM systems in modern democracies). The vast majority of countries currently using MM systems adopted them as a result of recently implemented electoral reforms. Three broad results emerge in terms of electoral competition: 1) an increase in the effective number of parties in former communist countries or in countries governed by PL systems of representation before the reform (see for example Boston et al., 1997 and 1998 for New Zealand; Moser, 2001 for Russia and Schieman and Benoit, 1999 for Hungary); 2) no dramatic changes in the effective number of parties in countries previously governed by PR electoral systems (see for example Reed, 1999 for Japan; Mayorga, 1997 for Bolivia; Bartolini and Dalimonte, 1995 for Italy; Crisp, 2000 for Venezuela) and 3) the formation of broad-based electoral coalitions in all countries adopting MM electoral systems (see for example Mudambi et al., 2001 for Italy; Hazan, 2000 for Israel; Kulisheck and Crisp, 2001 for Venezuela; Mayorga, 2001 for Bolivia). These three general results seem to confirm that the countervailing tendencies in both the PL and PR dimensions lead to electoral competition between two opposing coalitions in a 3

4 multi-party political system. Mudambi et al. (2001) and Mudambi and Navarra (2002) provide the first attempt to formulate a theory of electoral strategies in MM systems characterized by an electoral competition between two competing coalition in a multi-party context. They argue that the PR/PL seat ratio provides parties with two sets of incentives, drawn from the PL and PR components of the system. Neglecting the narrow concerns of their core supporters harms a party s performance on the PR component. Thus, narrow political formations must retain their unique identities. However, success on the PL ballot requires broad-based support. Dealing with both sets of incentives requires narrow political formations to form coalitions. Within this framework, political party strategies in SMDs are aimed at two ordered objectives. The first is the PL-driven coalition focussed objective, under which each party seeks to ensure that its coalition gains control of the government (inter-coalition competition). The second is the PR-driven party focussed objective, under which each party strives to dominate its coalition (intra-coalition competition). The power of a party within the coalition is important because it determines the importance of the party s agenda in the agenda of the coalition. Assuming that the intra-coalition game is played non-cooperatively with Nash strategies, each coalition party is expected to exert maximum effort (and elicit maximum support from its adherents) only when the primary objective of a coalition victory is threatened. This will occur in marginal SMDs where the ex ante belief in a coalition victory is weak. In such a constituency inter-coalition competition drives the electoral behaviour of parties. On the other hand, in safe SMDs where the ex ante belief in a coalition victory is strong, the primary objective is not in doubt, so the secondary objective (intra-coalition competition) determines party behaviour. Here parties exert maximum effort (and elicit maximum support from their adherents) only if the candidate in the PL ballot is one of their own. If the coalition candidate is from one of their allied parties, they put virtually all their energies into fighting the PR ballot. 4

5 In order to illustrate electoral strategies in MM systems let us describe the two pathological extreme types of constituencies. Without loss of generality, suppose that the coalition candidate belongs to party i. At one extreme, the candidate receives no support from any party except for his own. Such a constituency, where coalition parties pursue their own intra-coalition objectives at the expense of coalition prospects may be termed an internecine constituency. In this case, the coalition vote share V = v i, where V is the coalition PL vote share and v i is the vote share of party i in the PR ballot. At the other extreme, all the parties, which adopt the approach of 'all for one and one for all', embrace the coalition candidate. Such a constituency, where the constituent parties act for all intents and purposes like one party may be termed a Dumas constituency. In this case, the coalition vote share V = v k, i.e., the sum of the shares of all the parties in the coalition. Party behaviour in most constituencies is likely to fall between these two extremes and the actual PL vote share will also be affected by such factors as the candidate s personal popularity (or lack thereof). Thus, the coalition vote share will tend to exceed the vote share of party i, but will tend to fall short of the maximum potential vote share. But how short will it fall? This will depend on expectations of the constituent parties with regard to coalition victory in the constituency. In constituencies where coalition victory is expected, i.e., safe constituencies, parties will feel free to pursue party-focused objectives at the expense of coalition-focused objectives, causing the shortfall to be large. However, parties do not have this luxury in constituencies where victory is in doubt or unlikely, and will exert themselves to obtain the best possible showing for the coalition. Thus, in such constituencies the shortfall is expected to be small. The electoral strategies described so far can have important implications on the determination of competitive electoral equilibria. Duverger s law and Duverger s hypothesis state respectively that PL based elections with SMDs generally result in a twoparty polity, while PR systems generally give rise to multi-party systems (Duverger, 1954). 5

6 However, Duverger s propositions have generated controversy and have been questioned. One question has been the generality of the propositions: do we expect PL systems to result in two-party polities and PR systems to generate multi-party systems, always and in all circumstances? (See, for example, Riker, 1982; Bogdanor, 1983; Sartori, 1994). Reed (2001) analyses the 1994 and 1996 Italian national elections and demonstrates that, although there is little evidence of convergence toward a two party system, more than 80% of SMDs moved closer to bipolar competition between two candidates. Therefore, he claims that Duverger s law works as expected at the district level. The analysis of MM system electoral strategies by Mudambi and Navarra (2002) suggests that this electoral mechanism generates incentives to create two-sided coalition-based competition with a multi-party system. They point out that MM systems have a significant PL driven component and demonstrate that this is generally sufficient to ensure that the polity can be controlled by broad-based political formations. However, the PR component ensures the persistence of many narrow focus parties, whose support can often be crucial in government formation and continuation. Thus, MM systems are multi-party systems in which the formation of coalitions leads to two-sided politics. This analysis can also be interpreted using the spatial theory of electoral competition (Hinich and Munger, 1997). The predominant view in this literature is that candidates maximise votes or plurality and therefore see policy as a means to winning elections (Downs, 1957; Kramer, 1977). Some scholars, however, have modelled policy-oriented candidates who fight elections to achieve their own policy goals (Wittman, 1983; Calvert, 1985). Such candidates are interested not only in winning elections, but also in implementing their policy preferences. The protagonists in an MM system are the constituent parties of coalitions, rather than individual candidates. However, they face the same trade-off of balancing their individual policy positions with the coalition s overall 6

7 policy position. The former affects the party s performance on the PR ballot, while the later affects its performance on the PL ballot. Are MM systems expected to generate convergent or divergent competitive equilibria? On the basis of the discussion so far, it may be hypothesised that the answer to this question depends on the relative importance of the PR and PL ballots. If the PR ballot is much more important, intra-coalition competition is more important than inter-coalition competition. In this case parties focus on their individual policy agendas and the outcome is divergence. Conversely, if the PL ballot is more important, inter-coalition competition is more important than intra-coalition competition. This means that parties focus on creating a common coalition policy agenda to compete for the median voter and the outcome is convergence. The relative importance of the PL versus the PR ballot would typically be measured by the PL/PR ratio. For example, in Italy the PL/PR ratio is 75/25, while in Germany it is 50/50. Since Italy has a much higher ratio, the PL ballot is relatively more important, and ceteris paribus, we expect more convergence than in Germany (Navarra, 2002). Ram Mudambi Associate Professor of Strategic Management, Temple University (USA) and Reader in International Business, University of Reading (UK) Pietro Navarra Associate Professor of Public Economics, University of Messina (ITA) and Research Associate, CPNSS, London School of Economics (UK) REFERENCES Banducci, S., J. Karp, J. Vowles and T. Donovan (1999). Strategic voting in the 1996 New Zealand Election: Implications for a new mixed system. Working Paper, University of Waikato. Bartolini, S. and R. Dalimonte (eds.) (1995). Maggioritario ma non Troppo: le Elezioni Politiche nel Bologna: Il Mulino. 7

8 Benoit, K., D. Giannetti and M. Laver (2002). Strategic Voting in mixed-member electoral systems: The Italian case. Mimeo, University of Bologna. Bogdanor, V Conclusions: Electoral systems and party systems, in Bogdanor, V. and Butler, D. (eds.) Democracy and Elections: Electoral Systems and Their Political Consequences. Cambridge University Press: New York and London. Boston, J., S. Levine, E. McLeay and N. S. Roberts (1997). From campaign to Coalition: The 1996 MMP Election. Palmerston North, NZ: The Dunmore Press. Boston, J., S. Levine, E. McLeay and N. S. Roberts (1998). The New Zealand Parliamentarians: Did electoral system reform make a difference? Legislative Studies, 12: Calvert, R. (1985). Robustness of the multi-dimensional voting model: Candidate motivations, uncertainty and convergence. American Journal of Political Science, 29: Downs, A An Economic Theory of Democracy. Harper and Row: New York. Crisp, B. F. (2000). Democratic Institutional Design: The Powers and Incentives of Venezuelan Politicians and Interest Groups. Stanford, CA: Stanford University Press. Duverger, M. (1954). Political Parties, their Organisation and Activity in the Modern State (Translated by B. and R. North). Wiley: New York. Hazan, R.Y. (2000). Parties, Elections and Cleavages: Israel in Comparative and Theoretical Perspective. London: Frank Cass. Hinich, M.J. and Munger, M.C. (1997). Analytical Politics. New York: Cambridge University Press. Jesse, E. (1990). Elections: The Federal Republic of Germany in Comparison. (Translated by Lindsay Batson). New Yok: St. Martin s. 8

9 Klingeman, H.D. and B Wessels (2001). The political consequences of Germany s mixed-member system: Personalization at the grass-root? In M.S. Shugart and M.P. Wattemberg (eds.) Mixed-Member Systems: The Best of Both Worlds? Oxford: Oxford University Press. Kulisheck, M. R. and B. F. Crisp (2001). The legislative consequences of MMP electoral rules in Venezuela. In M.S. Shugart and M.P. Wattemberg (eds.) Mixed-Member Systems: The Best of Both Worlds? Oxford: Oxford University Press. Kramer, G. (1977). On a class of equilibrium conditions for majority rule. Econometrica, 41: Mayorga, R.A. (1997). Bolivia: Electoral reform in Latin America. In the International IDEA Handbook of Electoral System Design. Stockholm: Institute for Democracy and Electoral Assistance. Mayorga R.A. (2001). The mixed-member-proportional system and its consequences in Bolivia. In M.S. Shugart and M.P. Wattemberg (eds.) Mixed-Member Systems: The Best of Both Worlds? Oxford: Oxford University Press. Massicotte, L. and A. Blais (1999). Mixed electoral systems: A conceptual and empirical survey. Electoral Studies, 18: McAllister, I. And S. White (2000). Split-ticket voting in the 1995 Russian Duma elections. Electoral Studies, 19: Mudambi, R., P. Navarra and G. Sobbrio (2001). Rules, Choice and Strategy: The Political Economy of Italian Electoral Reform. Aldershot: Edward Elgar Ltd. Mudambi, R. and P. Navarra (2002). Electoral Strategies in Mixed Systems of Representation. Discussion Paper, Jan., Research Program on Democracy, Business and Human Well-being, London School of Economics. 9

10 Navarra, P. (1997). Voting Diversification Strategy: A Risk-Bearing Model of Voter Behaviour in 1996 Italian National Election, in Economia delle Scelte Pubbliche, Journal of Public Finance and Public Choice, 1: Navarra, P. (2002). Spatial electoral competition in mixed-member system of representation. Mimeo, University of Messina. Reed, S.R. (1999). Political reform in Japan: Combining scientific and historical analysis. Social Science Japan Journal, 2: Reed, S. R. (2001). Duverger s law is working in Italy. Comparative Political Studies, 34: Riker, W.H. (1982). Two-party systems and Duverger s law. American Political Science Review, 76: Sartori, G. (1994). Comparing Institutional Engineering: An Inquiry into Structures, Incentives and Outcomes. Columbia University Press: New York. Schoen, H. (1999). Split-ticket voting in German federal elections, : An example of sophisticated balloting. Electoral Studies, 18: Schieman, J.W. and K. Benoit (1999). How thick is the veil of ignorance? Uncertainty and the electoral system origins in Hungary. Mimeo, Columbia University. Shugart, M.S. (2001). Electoral efficiency and the move to mixed-member systems. Electoral Studies, 20: Shugart, M.S. and M.P. Wattemberg (2000a). Mixed-member electoral systems: A definition and typology. In M.S. Shugart and M.P. Wattemberg (eds.) Mixed-Member Systems: The Best of Both Worlds? Oxford: Oxford University Press. Shugart, M.S. and M.P. Wattemberg (2000b). Are mixed-member systems the best of both worlds? In M.S. Shugart and M.P. Wattemberg (eds.) Mixed-Member Systems: The Best of Both Worlds? Oxford: Oxford University Press. 10

11 Wittman, D. (1983). Candidate motivations: A synthesis of alternatives. American Political Science Review, 77: