In a few weeks, France will vote for its more than 36,000 mayors. The main candidates for Paris city hall — the conservative Nathalie Kosziuscko-Morizet and the socialist Anne Hidalgo — are using the Internet to get voters, but in a special way. “You can vote twice !” says the socialist candidate’s Web site. How so ?
The conservative and the socialist candidates’ Web sites have sections where a potential voter can find a proxy and where a willing proxy voter can register. France does not allow mail-in voting or early voting : proxy voting (vote par procuration) is the only way to vote when physically prevented to vote. And parties are increasingly turning to social media to elicit proxy votes.
Proxy votes used to be a rounding error in polling results. That is no longer the case. During the 2012 presidential elections between 4 and 5 percent of all votes were cast by a proxy. In Paris, nearly 10 percent of all votes were cast by a proxy. They have long been seen as a means of fraud, especially in Corsica where in some municipalities more than 50 percent of all votes are proxy votes. They are now integrated in the repertoire of political actions.
Proxy votes are basic local political tools. This is visible in municipal elections where a few dozen votes can mean victory or defeat. The following map shows the frequency of proxy votes in the 2008 municipal elections in Paris. Paris is a city composed of 20 “arrondissements” (boroughs), and each borough elects a mayor. “Procuration” is usually low in municipal elections. However one arrondissement, the Fifth, stands out. The incumbent mayor, Jean Tiberi, was reelected with 225 more votes than his competitor. His team and his opponent’s team seem to have relied on more than 1,600 proxy votes. Tiberi was convicted in 2013 for vote rigging in several elections in the 1990s.
Would the result have been different without proxy votes? Is there a possibility that proxy voting biases results ? Do they tend to favor one political party? Looking at polling stations results (that is looking at a very small geographical units composed of around 1,000 registered voters) is enlightening.
First: In places where people vote at a high rate, they also vote more by proxy. Abstention and proxy votes are negatively correlated. It may be because, by definition, proxy votes are designed to reduce abstention. But places where people cast few invalid ballots (“votes blancs et nuls”) are also places with a higher proportion of proxy votes (see figure above). Invalid ballots (ballots that are empty, torn-out or struck out) are sometimes interpreted as a way to express one’s dissatisfaction with the candidate lists. So, many proxy votes may signify greater satisfaction with the candidate offerings.
Second: the geography of proxy votes in Paris is a familiar one. The proxy vote map and the Sarkozy map have many similarities. At the north and east of Paris, the polling stations register few “procurations”. At the center and the west, procurations seems much more frequent. The outskirts of Paris, often called the “belt of public housing” is also visible: proxy votes are at their lowest. Thanks to the Cartelec academic group, we have a series of social indicators at the level of the polling station. Proxy votes are more frequent in polling stations where senior citizens, managers, or owners constitute a higher proportion of the residents. The proportion of unemployed, of renters, of people without a high school diploma is negatively correlated with proxy votes. Ecological fallacy is still possible: One may imagine that the unemployed in upper-class neighborhoods are more likely to vote by proxy (although it is not clear why that would be so).
This can be replicated for several elections and several cities: The more bourgeois the neighborhood, the more frequent the proxy votes. In local elections, proxy votes favor a better organized party machine. In national elections, they favor the social groups who are already electorally mobilized. Proxy votes are designed as a neutral device: Their aim is to reduce abstention across the board. But in national elections, they are more frequent in places that vote for conservative candidates, and in places with a well-to-do population.
Unfortunately French vote registration is nonpartisan, and it is not possible to know the political affiliation of individual voters by looking at registered voters list. To better examine this issue, we would need surveys of individuals.