Bürgerschaftswahl in Hamburg 2011
Following the very positive forecasting experience with the presidential election in Portugal on the 23/01/2011, we decided to make a step further and try, for the first time, to forecast the outcome of a political election for the single parties.
Our first experiment in this regard is represented by the German state elections in Hamburg to be held on the 20/02/2011. Even though the internet penetration in Germany is high (79.1% of the population as of June 2010), nevertheless this election still represents a "tough environment" for three reasons:
- The number of voters involved is quite low: less than 1 million voters, compared to 4.5 million for the Portuguese presidential elections (and over 9.5 million potential voters).
- Modelling and forecasting single parties is more difficult than single individuals.
- In case of German elections, online searches start to give significant indications only in the last weeks before the election date, while older data are much more noisy, if not completely non-existent.
Following the same logic used with the Portuguese presidential elections, we decided to start our test with single parties in a tough environment such as the Hamburg state election, because any eventual problems in our methodology would be much more evident and simple to detect ex-post.
Interestingly, when we started working with the Hamburg election data, we noticed immediately that the official political poll houses, when doing their calibrations, tend to give more weights to the recent German national elections held in September 2009 for some parties (like the CDU), while for other parties (like the FDP, Grune and partially the SPD) they tend to give more weights to the past state elections held in Hamburg in 2008.
Since we do not possess the experience about the Hamburg local political dynamics that official political poll houses (should) have, we considered in this case the following two possibilities:
1) Forecasts using GTI data and historical electoral results for the German national elections in 2009;
2) Forecasts using GTI data and past historical electoral results from 2008 onwards (including the past State election in 2008, European election in 2009 and German national elections in 2009).
Our forecasts using data up to 18/02/2011 are the following ones:
1) CDU: 24.38 SPD:36.73 FDP: 7.61 GRUNE: 20.88 LINKE: 8.98
2) CDU: 29.37 SPD:38.32 FDP: 5.84 GRUNE: 18.32 LINKE: 6.79
For sake of simplicity and interest, I considered only the five most important parties. Finally the latest trends in last 90 days reported by GTI are reported below:
