|In 500 Signatures’ debut guest blog, Martial Foucault and Sylvain Brouard provide a forecast of the FN’s likely performance in next week’s municipal elections. They look at the party’s performance across a series of elections in the 56 main cities in France, and find that the surge in the FN’s popularity and current cohesion suggest it will post an overall score well over 20% in Sunday’s first round - better than any previous election other than the 2002 earthquake presidentials.|
On 6th March 2014, the final number of electoral lists presented or supported by the National Front (FN) had been officially submitted to the prefectures in charge of organizing municipal elections. In municipalities of over 1,000 inhabitants, the Front National will field 596 lists– more than ever before. With nearly 450 lists presented in cities of more than 9,000 inhabitants, the party of Marine Le Pen is aiming to return to its 1995 strength, notwithstanding the new constraint of gender parity in the composition of the lists. For the record, the FN presented lists in 303 municipalities of more than 9,000 inhabitants in 1989, 456 in 1995, but only 183 and 106 in 2001 and 2008,respectively (on which there can be added, respectively, 114 and 5 cities in which no FN candidate ran for election, but where other extreme right lists of the extreme right).
Will such an increase in FN candidacies affect the outcome of the traditional left / right competition in these elections, whose issues are not only local but also national? The FN won its best election share in 1995, obtaining 13.1 percent of the vote in the larger municipalities, a score that enabled it to qualify for the second round in more than 290 towns, and to get nearly 1,000 municipal councillors. Is the FN likely to reproduce the same performance as in 1995?
Opinion surveys regarding vote intentions, as interesting they may be, have real difficulties in forecasting a realistic score for the FN due to the phenomenon of chronic under-declaration of its voters. Moreover, those surveys only report the estimated votes in a dozen or so cities, which is not always representative of the national trend, and therefore does not permit an extrapolation to the FN’s overall score.
To overcome such difficulties, researchers have successfully tested an alternative method, based on the modeling of previous elections results, to predict the national scores of the FN since 1984 as well as that of Marine Le Pen in the presidential election of 2012. To date, no approach has been implemented either at the disaggregated level of cities or for the FN vote in the municipal election. Accordingly, we have built a simple statistical model to identify the explanatory variables of the FN’s vote at the local level since 1998 (including all elections except legislatives) and then estimate the FN’s score in the first round of the municipal elections in March 2014.
This methodological choice deserves a little explanation.The results of the FN in the forthcoming elections have to be considered as a combination of local and national factors. If the incumbent mayor is accountable for her past performance, her notoriety and her policy promises, voters may nevertheless use this electoral opportunity to reward or punish the national government for its performance, as well as the opposition parties at the national level. From this perspective, it is therefore important to consider the FN’s popularity at the national level as a proxy of sympathy for third parties.Further, the electoral dynamics of each of the cities under investigation is a determining variable for the simulation of the future score of the FN. In Paris, the average score of the FN has been 6.6% between 1998 and 2012, whereas it has been about 25.5% in Carpentras in the south of France. The volatile results of the FN from 1995 to 2008 (see table 1) at the municipal level become part of a much more complex dynamic when we also consider its results in national, regional and European elections in each city. In Dreux, for example, where the FN enjoyed an average score of 14% of the votes between 1998 and 2012, the scores ranged from 4.3% in the municipal election in 2008 to 31.7% in the regional election of 1998.
Insofar as our aim is to forecast the FN vote at the forthcoming municipal elections, we select 56 cities in which the National Front and/or its dissidents have presented electoral lists in the municipal elections of 1995, 2001 and 2008. Based on the scores of the FN in the municipal, presidential, regional and European elections from 1998 to 2012 (as dependent variable), our statistical model aims at identifying the relative weight of diverse predictive variables of the FN’s vote.
Our working hypotheses are as follows:
- the higher the score of the FN in the former election as well as in the former election of the same type in each city, the higher the score of the FN in the next election;
- the more the FN’s popularity grows, the more the FN vote share increases;
- the more the crime rate rises, the more the FN vote share increases;
- when the FN is divided, its score declines;
- the type of the election affects the score of the FN.
Based on 616 observations(56 cities x 11 elections), our model reports statistically significant coefficients for each of the independent variables. One methodological point: a forecast model has to rely on a parsimonious and robust statistical model, that anyone can replicate, based on explanatory variables known before the election. Our model is only based on variables known before the election, like for example the popularity of the FN, measured in the TNS-Sofres Figaro-Magazine barometer, held in December of the year before the municipal, presidential and regional election as well as in March before the European elections. The model accounts for a large proportion of the FN’s score variance (R-sq.= 0.7).
As shown in Table 1, the average estimated score of the FN for the different elections in the 56 cities is very close to the real electoral outcome. The largest discrepancy between prediction and reality (2.59%) is observed for the presidential election in 2007. For the municipal elections of 2001 and 2008, the difference between the estimated and the observed score is small, respectively 0.37 and 0.24. Overall, the model reliably estimates the variations in election results for the FN since 1998. It can therefore be used to predict the score of the FN in the forthcoming municipal election of 2014.
Table 1:Performance of the model
Source: Ministry of the Interior, EDEN (CEVIPOF), authors.
Based on the variables known in December 2013, the model forecasts an average score of 22.8% for the FN in the 56 cities under investigation with a margin of error of around 0.45%.
Our estimate for the first round predicts an historic municipal surge for the FN. In 1995, the average score of the FN in this sample of cities was around 17.17%. If such a tendency was confirmed or even exceeded next week in all the municipalities the FN is presenting or supporting a list, this would allow the Front National to be qualify for the second round in numerous cities and, consequently, to weigh on electoral competition, but also obtain an unprecedented number of municipal councillors and even win a number of municipalities. Against this background, a high number of FN municipal councillors elected in March at the local level could give the party a real chance to influence the campaign for the senatorial elections in the autumn.
About the authors:
Sylvain Brouard is an Associate Research Professor at the Centre Emile Durkheim, Sciences Po Bordeaux
Martial Foucault is a Full Professor at CEVIPOF, Sciences Po, Paris
A French version of this blog can be found at Slowpolitix