The Italian elections: a forecast

This evening I’ll be participating in a panel at Cambridge about the Italian elections. Also there will be Cardiff-ites Paul Furlong and Mark Donovan, and Bathite (Bathist?) Anna Bull.

At the event, I’m going to give a forecast of the results of the Italian elections. You can read the full version of what I’ve written on SSRN (The 2013 Italian elections: a poll-based forecast), or read the abstract below:

I forecast the results of the Italian general election using aggregated polling data. According to this forecast,

  • there is an extremely high probability of the centre-left winning an overall majority in the lower chamber (p ≈ 100%); the probability of the centre-left winning an overall majority in the Senate, however, is extremely low, at 4.5%.
  • The centre-left is very likely to lose in Lombardia and the Veneto. It has a two-in-five chance of winning in Sicily.
  • The radical left list, Rivoluzione Civile, is unlikely to win seats in the Senate (p = 20.7%). By contrast, the two extreme-right parties which form part of the centre-right coalition (Fratelli d’Italia and la Destra) seem to have a good chance of acquiring representation in the Senate (though not necessarily in the Camera).
  • A coalition between the PD and the list led by Mario Monti would be extremely likely to have a majority in the Senate (p = 94.2%). However, this depends on considerable cohesion within the Monti-led list. A coalition which drew only on ‘definite’ Montiani (i.e., not including those placed on the list by the UDC or FLI) would only have a 50:50 chance of a majority, even with support from minor parties on the centre-left.
  • Consequently, almost irrespective of the government that forms, fresh elections are likely to be held within the next two years.

As always, I’d appreciate any comments you have. I’ll be picking up some of the issues in this forecasting exercise over the next couple of days.

Update [14/2/2013]: Duncan McDonnell quite rightly points out that Fratelli d’Italia doesn’t have a very strong claim to be considered an extreme right wing party, given that they’re to the left of what was Alleanza Nazionale, and that people like Ignazi had removed AN from their lists of ERWPs.

10 thoughts on “The Italian elections: a forecast

  1. How would FdI/LD gain representation in the Senate, but not the House? Because of the different election thresholds?

  2. Yes: although the implicit threshold caused by the “district magnitude” in each constituency is quite severe, it can be passed by relatively small lists in large regions as long as they are part of the winning coalition. So, the seats all come in the big three of Lombardia, Sicilia and Veneto.

  3. You guys don’t understand a squad about Movimento 5 Stelle. Abandon statistics and mathematics and connect any evening at 21:00 on YouTube channel lacosachannel and see how many people turn up every time to listen to Beppe Grillo.
    But that’s ok you will learn more about us next Monday.

  4. I think I’m definitely less confident (in the ordinary language sense, rather than the statistical sense) about making predictions about M5S, because it’s a new party. I think it’s amazing that M5S has won this much support given its absence from traditional media. But I don’t know how much followership on Youtube counts in elections. That might say more about the intensity of people’s preferences than the number of people with such preferences. But thanks for the comment, in any case!

  5. Hej Chris, that’s the point! You see you are “surprised that M5S has won this much support” shows that you don’t understand it and cannot treat it the standard way.
    The YouTube thing is not about how many people follow on YouTube, you misunderstood me, the YouTube streaming show how many people turns up to listen to Beppe. Evey city, every square, twice maybe 3 times a day. Squares that can host 5,000/30,000 people and don’t forget Rome on the 22 Feb.
    Must know that M5S gets its vote with “passaparola”, on the street, on the internet, teens asking grandpa and grandma to vote, the sense of community that is going beyond the political aspect. It is a cultural revolution and it is not antipolitcs as described but quite the opposite, it exists because people wants politics, the real one for the people.
    We will shuffle the cards and play a new game after 70 years, that’s what people want.
    Don’t be surprise if Monday we will soar 20% and more.

  6. Yep, 20% is not impossible; the trend certainly was upwards. But for the moment, I’m going to stick to my estimate of 16%. Come Monday, all will be revealed!

  7. Hi Chris, all is revealed now and as I said last Tuesday we were expecting more then 20%. In fact 25% fits us and we will grow even further the next election probably in autumn this year.
    As I wrote, I think you should consider some sort of element in the statistic prediction model that takes into consideration the number of people in the street attending Beppe’s speech and the possibility of an extraordinary historical happening like the entering a new era in the way politics is done.
    We will set new rules, a new political model which is closer to the original concept of ancient Greek democracy, that is direct democracy (thru internet).
    We are not antipolitics we are the new politics.

  8. Hi M5S_Lombardia — yes, you were quite right! I’m not quite sure how to “account for” the tendency of the M5S to be under-estimated by the polls. All of the things that you mention are important, but difficult to include in a quantitative model. Polls have many disadvantages, but they are at least easy to interpret.

    Good luck with what comes next – it’s certainly going to be interesting…

Leave a Reply

Your email address will not be published. Required fields are marked *