Few countries have undergone such dramatic changes in their political systems in recent years as Italy. With the collapse of the Christian Democratic Party, which had been a part of every governing coalition since World War II, and with the replacement of an extreme version of proportional representation by a system with 75% of seats elected by plurality in single-member districts, Italy's 1994 parliamentary elections provide a remarkable opportunity to study the impact of voting system change.

        Many Italians voted for a more majoritarian electoral system in the historic 1993 referendum out of a desire to create a two-party system with regular rotations of power. But the 1994 elections demonstrated that creating such a system requires more than changes in the electoral law. With geographically-polarized results and the continued representation of numerous political parties, the new parliament is arguably less stable than the notoriously complicated coalitions of the pre-reform era.
        Beyond the unprecedented victory of the coalitions formed by Silvio Berlusconi just months before the elections and the collapse of the parties of the old ruling coalition, the most striking feature about the outcome of the 1994 parliamentary election is the difficulty one has in saying with any precise detail what the outcome was.
        The difficulty stems in part from the facts that there were three separate ballots (Senate, Chamber of Deputies collegio and Chamber of Deputies circoscrizione list), that the voters were not constrained to cast their votes in a consistent manner and that choices presented to the voters generally differed among those ballots.
        The difficulty also stems from variability across the country in the identity even of the major contestants. Most notably, in the north Berlusconi's Forza Italia was allied with the Lega (Northern League) in the Polo della Libert�, which was opposed by the Alleanza Nazionale, but in the south and center, the Berlusconi coalition was the Polo del Buon Governo formed in alliance with the Alleanza Nazionale. A final difficulty was the failure of significant numbers of Senators and Deputies to join the parliamentary groups corresponding to the parties or lists on which they were elected.
        These difficulties are reflected in Table 1, which shows the coalition of the right (the Poli) to have won a strong absolute majority in the Chamber of Deputies, but to have fallen just short of a majority in the Senate. In fact, the Poli were a bit farther short of a majority in the Senate than the table shows, because in addition to the 315 elected senators, the Senate includes 11 life senators, none of whom adhered to the parliamentary groups of the Poli.

        Table 1 also shows the overall distribution of members among the parliamentary groups. Not directly reflected in the table is the fact that 24 deputies and one senator elected as members of the Progressisti, Patto per l'Italia, or Poli chose to join the Gruppo Misto in parliament, rather than one of the groups corresponding to their electoral alliance. Principal among these were 5 of the 18 deputies of Alleanza Democratica and all 13 of the deputies of the Patto Segni.
        Also not directly reflected in the table is the movement of 9 deputies who identified themselves in their biographies as candidates of Forza Italia into other groups of the right coalition, and the movement of 15 deputies who identified themselves as candidates of other right-coalition parties into the Forza Italia parliamentary group.
        Examination of Table 1 shows a significant difference between the blocks in the importance and operation of the proportional seats. While these clearly were intended to assure the continued representation -- and existence -- of the smaller parties whose votes had been required to pass the reform legislation, it was only in the block of the Patto per l'Italia that the PR seats played a major direct role in securing representation for smaller parties.
        Ironically, the principal beneficiary of the PR seats was the PPI itself, which as the successor of the DC was assumed in August of 1993 to be one of the major parties. The PPI won 18.7 and 32.5 percent of all the PR seats in the Chamber and Senate, respectively, amounting to 87.9% and 90% of its total representation in the two chambers.
        On the other hand, with the exception of a few seats in the Senate, the smaller parties appear to have won representation in spite of the PR seats, rather than because of them. This is especially true for the Chamber of Deputies, where the four percent threshold shut out parties that among them won roughly 15% of the vote.
        The impact of the threshold fell particularly hard on the left, which won more than half of the "wasted" list votes, and which might have expected to win 10 more seats had, for example, the threshold been one percent instead of four. (At the same time, the Poli would have won a decisive majority in the Chamber anyway.)

Regional Polarization and Easy Victories

        Breaking down the result of the 1994 election by region shows that even with regard to the Senate, for which the aggregate results were extremely close, the reforms have not given Italy a competitive bi-polar system. Rather, the projections from 1993 of "three Italys" have largely been realized, albeit in a form somewhat different from that originally expected.
        Using the single-member district seats as an indicator, as Table 2 shows, the north is a solid bastion for the right, which won roughly 90% of the collegi for the Chamber, and nearly as large a percentage (88%) for the Senate. Moreover, these contests were not close -- the average winning Senate candidate of the Poli in the north led his/her closest competitor by over 16% of the vote, while the average margin for his/her lower house counterpart was nearly 25%.
        This pattern is repeated, but in mirror image, in the center, which was a solid enclave for the left. The left won over 95% of the single member seats for each chamber in the center, with average margins of victory nearly 25%. Only in the south was there significant competition, both in the sense that both left and right won appreciable numbers of collegi, and in the sense that the average margins of victory on both sides were small enough to suggest significant numbers of seats could change hands in either direction.

        There were significant regional variations in patterns of competition, with regional differences not only with respect to which general groupings won seats, but also with regard to their patterns of competition. The most significant difference was within the right, where not only did the ally of Forza Italia differ across regions, but indeed the Lega (Northern League) had no candidates in the center or south and the Alleanza Nazionale ran against the Polo della Libert� in the north and north-center (Emilia Romagna and Toscana).
        A second difference was that Rifondazione did not have its own PR list for the Chamber in either of the Sicily circoscrizioni. Finally, assuming that the "real" competition in a first-past-the-post election is between the candidates in first and second place, the most common pattern was competition between left and right, with the center finishing at least third. Among those collegi in which a candidate of the center did manage to break into the top two positions, the pattern of right-against-center occurred most commonly in the north, while the pattern of left-against-center was overwhelmingly found in the south.

Less Voter Change than Meets the Eye

        The index of voter volatility is difficult to compute for the 1994 election because of the difficulty of identifying successor/predecessor parties. A reasonable approximation, however, is that the index stands at roughly 37.1, equivalent to saying that more than one voter in every three switched parties between 1992 and 1994.
        If, however, one looks only at what Bartolini and Mair have called block volatility, thus ignoring shifts among parties within the left or within the right and center, and thus also minimizing the consequences for the index of having Forza Italia rise and the DC collapse, but both within the non-socialist block, the index is only 5.5, high but no longer of earthquake proportions. Separating out the MSI-Alleanza Nazionale as a third block, increases the index to 8.1. In contrast, total volatility between 1987 and 1992 was only 15.5, while block volatility based on the two block model was 5.3, rising to 5.8 under the three block model.

Infusion of New Legislators

        Accompanying this high level of volatility was a radical renewal of parliamentary personnel. More than one third of the members of the Chamber of Deputies elected in 1992 had then served three or more terms in the Parliament and the overall rate turnover was about 44%; the corresponding figures for the Parliament elected in 1994 were 12% with three or more previous terms and over 71% having no parliamentary experience.

        Indeed, upwards of one third of the members of the new parliament report themselves to have had no previous experience in party or electoral politics at all. Naturally, the proportion of members lacking in parliamentary experience is highest for those parties that made great gains (the Lega, 69.5% new members; Alleanza Nazionale, 77.5 new members; and especially Forza Italia, 90.4% with no previous parliamentary experience), but even for the PDS, the PPI, and Rifondazione Comunista, over half the deputies elected in 1994 are new. Somewhat fewer senators (about 60%) are without previous experience in Parliament, with the partisan distribution of the inexperienced roughly the same as it is for the Chamber.
        Previous service in Parliament is not, of course, the only kind of political experience a member might have. Generally under one third of the members of each of the major groups report themselves to be political neophytes, a position which is least common among members of Alleanza Nazionale (under 20%). On the other hand, over 80% of deputies and senators of Forza Italia report, with some apparent pride, that they had never been involved in party or electoral politics before entering the lists in 1994.

Representation of Women

        The percentage of women representatives increased only in the Chamber of Deputies, although the decrease in the proportion of women in the Senate is trivial, and the increase in the Chamber of Deputies still leaves the representation of women in the Italian parliament well below that of most European countries. The increase in the Chamber of Deputies comes primarily from the PR seats, over one third of which were won by women (compared to women winning 9% of district seats), but this is still significantly below the 50% that one might naively have expected from the requirement of alternation of the genders on PR lists.
        The expectation that the increase in women's representation would come primarily in those parties that made the most significant gains finds only weak support. While it is true that 55% of the increase in the number of women deputies came in the three big parties of the right block, which collectively went from 80 seats in the old Chamber to 337 in the new, the remnants of the DC (PPI plus CCD) had two more women in the new Chamber even though they had lost nearly three-fourths of their seats, while women's representation on the left increased from 27 to 47 Deputies on a loss of roughly 20% of their seats.
        Even as the left was losing seats and the right gaining, the increase in women's representation as a proportion of the total number of seats won was greater for the left than for the right. While turnover clearly created opportunities for women to win election, the decisions of party elites to nominate women in winnable collegi and the required alternation of genders on the PR lists were more important.
        Representation of women clearly decreases as one moves from left to right along the political spectrum -- a notable change from the Eleventh Chamber of Deputies, and one potentially attributable to the demise of the voto di preferenza, which makes election now more dependent on the party's decision to place a candidate in a constituency where the party/alliance is strong, rather than on the strength of the candidate's personal appeal vis-a-vis other candidates of the same party.

        Richard S. Katz is professor of political science at State University of New York-Buffalo and author of numerous publications on electoral systems.