Italian politics

Italy has a new electoral law. It repeats one old mistake.

Italy has a new electoral law. I was excited about it until I started to write this post. Now I'm not so sure anymore.

Last fall, Renzi famously lost the referendum on a reform of Italy’s parliament. Part of this reform was the abolition of the Senate. Renzi was so sure about it that a new electoral law from earlier in 2016 had only covered the Lower House. When his referendum failed, as a consequence, the Senate was forced to use a pre-existing electoral law. The supreme court intervened and declared that this mess was in violation of the constitution and a harmonized electoral law was needed. Over the past months, several alternative laws had been drafted and failed to pass through the legislative process. One law that closely mirrored the German electoral system almost made it. Still failed. That’s maybe also okay. The German law has its own very big problems (key word: leveling seats).

This upcoming week, the senate will vote on the newest proposal, the Rosatellum 2.0. The Lower House, Italy’s other parliamentary chamber, already approved of it with a confidence vote last Thursday. It had support from the ruling PD, Silvio Berlusconi’s Forza Italia, and the right-wing hardcore populists from Lega Nord. The anti-establishment populists from M5S (the “Five Stars Movement” of Beppe Grillo), currently leading the polls by a hair length in front of the PD, aggressively oppose it. For them, it’s purely designed to keep them out of government.

As a quick reminder: under the current electoral law (of the Senate? – I might remember this wrong), the winning party would get an absolute majority of seats in parliament, even if it did not proportionally receive enough votes. Grillo is confident that he would take that absolute majority and rule. Renzi was confident about that in the past, but for obvious reasons is not anymore. Hence, Grillo is certainly right that the motivation of any other party to support this electoral law is to improve their chances by decreasing his.

But what’s in it. I found that foreign newspapers reported rather superficially. What I explain in the following is based on my understanding of some Italian news sources.

One electoral law

This is the obvious thing. The electoral rules for Senate and the Lower House would be the same. What the electoral law does not change is the perfect symmetry in power of both chambers of the parliament. That to change was the aim of the constitutional reform that lost in the referendum.

A mix of first-past-the-post and proportional democracy

Both chambers will reserve one third of their seats for candidates that win the majority in a specific constituency. Read: that receive the highest number of votes, even if it’s less than 50% of votes in that constituency. The other two thirds are proportionally assigned to party lists based on their national average vote share. So far, it does sound like it’s strongly reflecting the German system.

Just one vote

However, in Germany, a voter gives one vote to a candidate and another vote to a list. Vote-splitting is possible. I’ve done it. Last time I voted, I gave my first vote to a candidate of the Social Democrats and my second vote to the list of the Liberals. In Italy, each voter will exert only one vote. For instance, let’s say there are only six seats in parliament. There are also only two constituencies: Milan and Rome. They take two of the six seats. The other four are assigned proportionally. The girl might vote for Francesco Milanese (a fictional character) from the Pasta Party (a fictional party). Francesco might receive 20% of votes in Milan’s constituency, so also the Pasta Party received 20% of votes in Milan. Alberto Romagnolo, also from the Pasta Party, might receive 30% of votes in Rome. The average of 20% and 30% is 25%. Francesco did not get the seat in Milan (he came in second to Diana Volontaria from the Pizza Party), but Alberto did. So the Pasta Party gets one of two seats from those that are given first-past-the-post. And because they have received on average 25% of votes across all constituencies, they also get one out of four seats from those seats that are proportionally assigned. This is extremely simplified, of course. Hence, very popular candidates drive the result of a party as a whole. Very unpopular candidates can mean that a party loses votes from voters that sympathize with the party, but not with a specific person.

One candidate, up to five lists

How weird can it be? Francesco Milanese can only be elected in one constituency. Take the example with just six seats, for of which are assigned proportionally. However, he can be on the list of up to five different parties (or predefined list alliances). Remember that he lost in Milan. But for the proportional part, he is on the List of the Pasta Party, the Pizza Party, the Panzerotto Party, and also the Piadina party. He’s actually first on the list of the Piadina party. The Piadina party also got 25% of the votes on average, so wins one out of four seats. Francesco will take it.

One vote for a guy, or one vote for a party

Okay, imagine that Francesco from above is really not your kind of guy. You still like the Pasta Party. You actually have two options. You can vote for Francesco. Your vote will then help the Pasta Party with the proportional seats. Or you vote directly for the Pasta Party. Your vote will still be counted for Francesco as well. If I got it right. I’m really not sure about this part. Here’s the original Italian description from

Barrando solo il candidato uninominale il voto va ai suoi partiti in modo proporzionale. Barrando solo il partito il voto va sia al candidato uninominale, sia al partito.

Remember that Francesco can be on more than one list. If you vote for Francesco, your vote will count for all the parties that he is affiliated with. So one vote for Francesco means that for the proportional part there is one more vote for the Pasta Party, the Pizza Party, the Panzerotto Party, and also the Piadina Party. However, if you vote for the Pasta Party, that vote also counts for Francesco, but not for these other three parties. Basically, this is a very weird way to express pluralistic political sympathy.

A 3% threshold and a 10% threshold and it’s still a mistake

Individual parties need at least 3% to obtain any seat in parliament from the proportional seats. Else, their seats will be proportionally divided among those parties that got more than 3%. But parties can also form lists ahead of the election. These lists need to win at least 10% to receive any seat. That’s an improvement from before, when also lists needed not more than 3% to get into parliament. Hence, this is touted as a step to overcome the super-splintered setup of many Italian parliaments.

And I think it’s not enough.

In fact, it still means that weak parties can rest on the support of big parties, then enter a coalition agreement with another party as part of a list, and half-way through the legislative period break free, switch sides, and bring down the government. I think that predefined list alliances for the purpose of elections are generally a weird thing. If one really wants to insist on them, the 10% threshold should flexibly respond to the number of parties. 10% sounds fair for a list of two or three parties. If it’s a list of ten parties, it still is just a convenient way to bypass the minimum threshold.

These list alliances are also a great way to accumulate votes for the first-past-the-post system. I can foresee a mechanism in which small parties get awarded with high positions on the lists of list alliances for the proportional part, since the big parties of these alliances will expect to win more constituencies by means of the first-past-the-post share. Now imagine that this gamble doesn’t work and it’s only the list that comes into play. The consequence is a gross over-representation of small parties.

Hence, I agree with the Five Star movement here in that I think it’s undemocratic. And given that I don’t like these guys, please take this as a sincere statement. One fix that I see is to allow list alliances only to parties who currently have no seat in the parliament – or not more than a minimal number. Say five. That way, it’s not a tool for big parties to better their chances in local constituencies. It would indeed turn into a mechanism to help new parties gain parliamentary recognition.

Votes for abroad

Previously I said that one third of the seats are based on first-past-the-post and two-thirds are assigned proportionally. However, there are more than 100% of seats in any Italian parliament, since a few are reserved for Italians living abroad. They cannot vote for a local candidate, but they can express their preference for a specific list. Those 12 seats are then assigned proportionally, but based only on the vote preferences of those living abroad. Hence, if the Pasta Party won 75% of the votes of expatriate Italians, she will get nine of these twelve seats – even if the Pasta Party’s national average was only 25%. (I assume that expatriate Italians crave more for really good pasta.)

Oh, and there is one reserved seat for the autonomous region of Val d’Aosta.

It’s still proportionally more complicated

So far, I explained that the proportional seats would be assigned based on national lists. That’s actually not quite correct. In reality, there will be 231 local constituencies for the first-past-the-post part. On top of that, there will be 65 local constituencies for the proportional part. Those 231 and 65 obviously overlap. It just means that the Pasta Party will get some proportional seats based on the 20% they received in the proportional constituency of Lombardy (Milan is in Lombardy), and some seats based on the 30% they received in the proportional constituency of Lazio (Rome is in Lazio).

Despite pre-election alliances, Italy will need post-election coalitions

Roberto D’Alimonte, professor at LUISS in Rome, simulated what the new electoral law implies for forming majorities. In his simulations, he assumes that there are basically three blocks: the Five Star movement, a center-left alliance under the leadership of the Social Democrats, and a center-right alliance between Forza Italia and Lega Nord. (The latter one is an alliance of intimate enemies.)

A simple graph visualizes the simulation outcomes. First, there is the share of seats gained from the first-past-the-post part (maggioritario) on the y-axis. Second, there is the share of seats from the proportional part on the x-axis. The non-surprising point to be made is that in order to gain the absolute majority of seats in parliament, one party has to have at least 40% of the proportional seats (if it also has 70% of the first-past-the-post seats), or at least 55% of the first-past-the-post seats (if it also has 50% of the proportional seats). Since at present no list alliance that forms before the election is likely to get as many votes by itself, the consequence will ultimately be a post-election coalition of list alliances.

The Five Star movement is unlikely to be part of any of them. One of their key promises to their voters is that they will not enter coalitions with other parties in order not to have to dilute their political agenda. But if the Center-Right forms a list alliance to improve their performance in the local constituencies (first-past-the-post), it implies that Italy’s next parliament will have a governing coalition of many and exactly one opposition party. Or re-elections. And re-elections at latest once enough minor members of these list alliances drop out for the government to loose its majority.

Again, this is based on my likely incomplete understanding of this new electoral law. I recommend to read the original Italian news sources carefully (I think I managed that part) and to interpret them with a proper background in political sciences (I’m not going to give myself too much credit here). There are some other very critical opinions (this one is one from Südtirol – it’s a commentary in German), which more or less support my understanding.

Ultimately, I appreciate the underlying basis of the reform (a combination of first-past-the-post and proportional voting), but these list alliances are just nuts. I can’t really see how they add to more stability. They seemed gone in many of the previous drafts that had entered parliament this Summer. Hence, I was really excited that there was finally a new electoral law. Ironically, that ended when I started writing this post and understood the details.

To say it in Italian: uffa.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: