Electoral change and its impact on the party system in Italy

Abstract

The importance of electoral change as a factor of party system transformation in post-1992 Italy is evaluated by observing two distinct components of electoral change: changes in electoral behaviour and changes in the electoral law, and their impact in the different arenas (electoral, parliamentary, etc.) in which parties compete as individual independent actors or as components of more or less organic coalitions. The analysis of numerous party system indicators shows that electoral factors are not only responsible for most of the changes which occurred in the party system after the effects of Tangentopoli were exhausted, but also for the creation of a structural divergence between the electoral and the parliamentary party systems.

Notes

1. The 1987 Italian election was the last to exhibit ‘normal’ volatility values and is an ideal starting point.

2. The two reforms were approved in 1993 and 2005 and became effective respectively in 1994 and 2006. They were preceded by a minor reform of preference voting (1992). Reforms of the regional and municipal electoral laws had an indirect impact on the Italian party system at the national level as well.

3. The reform resulted from a rather complex process that began in 1992 with a popular referendum. Because of the cumbersome and somewhat haphazard procedure under which it was written and approved, the law is very complicated (Katz 1994). Here a consideration of its dominant characteristics is sufficient.

4. Chiaramonte and Di Virgilio (2006) are clearly of this opinion. The law guarantees the coalition with the highest number of votes at least 340 of the 617 seats assigned in the national distribution. This means that unless a coalition obtains at least 55 per cent of the vote, the amount theoretically necessary to obtain 340 seats, the winners obtain a seat bonus the size of which varies according to the actual amount of votes they obtain. The losing coalition and/or independent lists get the remaining 277 seats. Within the two coalitions seats are distributed by PR amongst their respective party components that have obtained at least 2 per cent of the vote nationwide or 20 per cent at regional level.

5. Of the several characteristics identified by Giovanni Sartori (1976) in his discussion of party systems, these two are implicitly acknowledged as constituting the system's structure. This choice, which we will follow here for obvious reasons of space and practicality, privileges economic left-right dimension and overlooks other relevant dimensions in Italy's political space, such as church–state or centre–periphery.

6. As important and dramatic as they may appear, changes that occurred in the nature of the basic units of the system, that is in the parties themselves, did not concern the structural variables of Sartori's model – the number of parties and polarisation. Only if these two variables undergo significant change (i.e. if the number of relevant parties drops below six or left–right self-placement differences drop to values close to four points), can we say that the party system has changed (Sartori 1982). Polarisation also reflects and is a consequence of the direction of competition. If the direction of competition remains centrifugal the system is still polarised.

7. Many authors have argued that Italy's polarised pluralism is a thing of the past and have attempted to provide their own definitions of the new system (see for example: Morlino 1996; Pappalardo 1996). Here I will not define the new system on the basis of Sartori's classic categories, but rather observe possible variations in its structural variables.

8. The Berlusconi II government exhibited exceptional stability; but this happened at the expense of numerous substitutions of ministers and a sizeable increase in the number of vice-ministers in the course of its term. Notwithstanding these attempts to preempt some of its coalition parties' potentially destabilising tactics, the Berlusconi's second government also came to a premature end and was replaced by Berlusconi III in 2005.

9. Italy's mean TV between 1953 and 1992 was 7.9, a value very close to the norm (7.7) for systems with 6 to 10 parties (Bartolini and Mair 1990: 158). Mean BV for the same period was 2.6, a value near the bottom of the country distribution.

10. Total volatility was 5.3 in 1972, 9.1 in 1976 and again 5.3 in 1979. Block volatility was a low 1.3 in 1972, a very high (by Italy's standards) 4.0 in 1976, and an average 2.6 in 1979 (Bartolini and Mair 1990: Appendix 2).

11. Newspaper accounts of the 2006 elections the day after indicated an increase in turnout at 83.6 per cent. This figure, however, excludes Italians abroad, who were always included in the statistics of previous elections. Turnout for this category of voters was about 39 per cent and contributed significantly to lowering the overall percentage to 81.2, as reported in Table 2.

12. The greater incidence of spoiled ballots in 1996 than in 1994, when the new electoral rules where introduced and voters should have been, logically, even more unprepared, can be blamed on the fact that some post-DC voters were disoriented by the DC's successor parties having presented separate lists in opposing coalitions, after having kept their unity under the PPI's banner in 1994.

13. The PPI's split forced us to revise our BV calculation criteria and to include the Popolari per Prodi in the centre-left bloc. This accounts for the still high 1996 BV value. The same criterion was used in 2001 and 2006 for the Margherita.

14. Such data however does not justify in my opinion the further conclusion of the near irrelevance of centre party groupings and moderate voters in determining electoral outcomes. In the UK, for example, alternation in government is decided by presumably centrist voters' lateral moves whose size is smaller than those observed in Italy (Bartolini and Mair 1990: 111).

15. The Casa delle libertà has so far managed to secure an undisputed leader, Silvio Berlusconi, even when in opposition, something with which the Unione seems to have constant problems. The parliamentary political game, however, is still dominated by individual party, or even faction, leaders.

16. An analysis of such continuity and pre-1976 values for most of the Table 4 indices can be found in Morlino (1996).

17. It should be noted that the 1993 and 2005 laws introduced thresholds (respectively 4 and 2 per cent) which made them in theory more capable of curbing fragmentation.

18. It is worth noting that high PwS scores in 1994, 1996, and 2001 are determined by the distribution of plurality seats among the parties making up the electoral coalitions, as few parties managed to obtain seats in the PR part of all three elections: 7 in 1994, 8 in 1996, and only 5 in 2001. In 2006, 11 parties passed the national or regional thresholds and obtained seats.

19. Even individual non-attached senators were relevant in the election of the President of the Senate that was decided in favour of the centre-right's candidate, Carlo Scognamiglio, by one single vote.

20. Polarisation is calculated by dividing ideological distance values by 9, its maximum (10 − 1 = 9) theoretical value (see Sartori 1982, especially fig. 10.1).