I have now had a chance to review the performance of the Italian electoral systems (Chamber of Deputies and Senate) in the recent election, compared with the 2006 election. These two elections had very different results–alternations in government from Right to Left and back to Right being the most notable–but were held under the same rules. Thus they offer a good opportunity to begin to isolate the effects of the rules from other factors.
As I have noted many times, the Italian electoral system put in place by the previous government of Silvio Berlusconi shortly before the 2006 election is fundamentally majoritarian. For instance, in the Chamber, it rewards the alliance with the plurality of the nationwide votes a qualified majority of 55% of the seats. The Senate rules are similar, but with a crucial distinction: the majoritarian bonus is calculated on the basis of several regions rather than nationally. (The Senate also has various lifetime senators, which is hardly a parenthetical point, given that it was in the Senate that the last government fell.)
Nearly all the news accounts of the last two Italian elections refer erroneously to the electoral system as proportional. Fundamentally, it is not. However, there is a key proportional component: Once a given alliance is determined to have the plurality and thus minimally 55% of the seats, the allocation of those seats is proportional to the parties comprising the alliance, and the remaining seats are similarly divided proportionally among parties within losing alliances (as well as any parties not running in an alliance but qualifying for seats).
So, we have in Italy since 2006 a true mixed electoral system. Not a mixed-member system,1 such as the one in place from 1994 to 2001. But a mix of principles inasmuch as pre-election alliances are rewarded with a bonus (an unmistakably majoritarian feature) while small parties are also rewarded for surpassing a quite low threshold (an obvious feature of proportional systems).
Which element of this odd mix will prevail in terms of the shape of the party system? Will we see high fragmentation because of the proportional element (again, calculated only after the “bonus”), or low fragmentation because of the majoritarian element?
The answer is, it depends. On which election we look at. The 2006 result was indeed fragmented, which presumably is what caused so many observers to comment on the electoral system as if though it was proportional, full stop. But the 2008 election produced an outcome that is far more consistent with the overarching majoritarian logic of the system.
I will use here the effective number of competitors (parties or blocs).
For the Chamber, the effective number of competitors was as follows (given as votes/seats):
2006, by bloc: 2.03/1.98
2008, by bloc: 2.72/2.19
2006, by party: 5.48/4.91
2008, by party: 3.74/3.07
Now isn’t that interesting! The effective number of competitors when measured by alliances went up in 2008, while when measured by parties it went down–a lot. The change in the parties is obviously much greater than the change in the blocs. Some parties (notably the Union of the Centre and the Rainbow Left) chose to go it alone, and not join a pre-election bloc in 2008.
The really big change in the number of parties reflects a substantial reduction in the internal fragmentation of each bloc. Here are the effective number of vote-winning parties within each bloc, again in the chamber:
2006, right: 3.17
2006, left: 2.35
2008, right: 1.49
2008, left: 1.26
By number of parties, the left had nine parties that won at least one seat in 2006 and only two in 2008, whereas the right had seven in 2006 and only three in 2008. Measured by the percentage of the alliance’s total vote that was concentrated on the largest party, the change is more dramatic:
2006, right: 47.8
2006, left: 63.5
2008, right: 79.9
2008: left: 88.4
It is really hard to overstate just how much this electoral system is majoritarian. It just looked less majoritarian in 2006 than it would in 2008. The majoritarian nature of the system was simply somewhat masked in 2006 by the presence of many small parties on the left that had their seats shares boosted by the fact that they were part of the alliance that earned the majoritarian bonus. For instance, four parties had from 2.0 to 2.5% of the national vote in 2006 yet an advantage ratio of around 1.2. There are not many electoral systems–majoritarian or proportional–that will give such a large bonus to such small parties.
In the Senate, the impact of the system is less dramatic, because of the regional calculation of both the majoritarian bonus by bloc and the intra-bloc proportionality. However, in the Senate, too, the outcome is much more majoritarian than it was in 2006. In the previous election, the votes result was so close that there was actually split outcome between the two houses: whereas the left alliance won the vote in the Chamber, 49.7% to 49.5%, the right alliance won the votes in the Senate, 50.2% to 49.0%. Thus, whereas the left benefited nationally from the bonus provision in the Chamber, the regional bonuses did not all accrue to the same alliance in the Senate. Still, the right won the Senate seats in 2006, 153 to 148 (before inclusion of the unelected Senators).
With a more decisive result in both houses this time, the winning alliance will have around 55% of the seats in each house.
Finally, we would expect the Senate to be more majoritarian (or “less proportional,” if you prefer) than the House, due to the regional allocation. Yet that did not appear to be the case in the close elections of 2006, as noted above. In 2008, it is clear. The votes hardly differ between the two houses, but the seat result shows more concentration on the two large blocs in the Senate result than in the Chamber, precisely as we expect from a majoritarian system.
Chamber, 2008, votes percent, seats percent, advantage ratio
Right, 46.8, 55.0, 1.18
Left, 37.5, 38.7, 1.03
Centre, 5.6, 5.8, 1.04
others, 10.2, 0.6, 0.06
Senate, 2008, votes percent, seats percent, advantage ratio
Right, 47.3, 54.9, 1.16
Left, 38.0, 42.5, 1.18
Centre, 5.7, 1.0, 0.18
others, 9.8, 1.6, 0.16
Notice the significantly worse performance in the Senate than in the Chamber of the Union of the Centre, the one significant party not to join either major pre-election bloc. Most of the difference was picked up in the Senate by the Left, the second largest bloc nationally.
Italy’s current electoral system is one of the strangest and most complex to be found among the older established democracies. But in its second run, it produced a decisive result in both houses and the closest thing Italy has seen yet to the vaunted “two-party system.” Despite its warts, is this the electoral system Italy has been waiting for? Stay tuned.
Meanwhile, the discussion at the previous planting on Italy continues…
Source note: I used primarily Adam Carr. However, I discovered a significant error in his reporting of the 2006 Senate result. So I turned to the Italian Ministry of the Interior election archive. There were minor discrepancies between Carr and the results there for 2008 and for the 2006 Chamber, but not enough to change the calculations I reported above (and had already performed before catching the 2006 Senate error in Carr).
- That is, two tiers, one nominal/plurality and the other list/PR, but also majoritarian; that system was also mis-labeled (mixed-member) proportional in many sources. Although there was some “vote linkage” (partially compensating from the list tier those parties that did not perform well in the nominal tier) that made the system less majortarian than it might otherwise have been, it was much more MMM than MMP. [↩]