Media Briefings

PLEASE DON’T VOTE FOR ME: Evidence from Germany of elections with perverse incentives

  • Published Date: June 2015

Tactical voting is not confined to first-past-the-post electoral systems, according to research by Dr Jörg Spenkuch, published in the June 2015 issue of the Economic Journal. His analysis of a 2005 by-election in Germany shows that voters may behave tactically even under proportional representation – and the share of voters who do so may be substantially larger than previously believed.

The study explores a unique ‘natural experiment’ that occurred during the 2005 elections to the German Bundestag. Eleven days before the federal election, the sudden death of a far right-wing candidate necessitated a by-election in Saxony’s District 160. About 200,000 eligible voters were consequently asked to go to the polls two weeks after everybody else, and after the preliminary results in all other districts had already been announced.

Due to a flaw in the German electoral system, which combines elements of first-past-the-post with proportional representation, the Christian Democratic Union (CDU) would lose a seat in parliament if it gained more than approximately 41,000 votes. For comparison, in the last election in 2002, it had won 49,638 votes.

Conservative voters thus found themselves in the paradoxical situation that voting for their preferred party might, in fact, benefit its political rivals. In the end, the CDU won ‘only’ 38,208 votes, helping it achieve a narrow three-seat lead over the Social Democrats.

By comparing the results of the by-election with that in districts unaffected by the perverse incentives in the electoral system, Spenkuch shows that the Christian Democrats’ loss of votes is highly unlikely to be due to chance. While the CDU’s results in District 160 are statistically indistinguishable from those in other districts in prior as well as subsequent election years, in 2005 it lost 3.5 percentage points more in District 160 than in otherwise similar districts – exactly when additional votes might cost it a seat.

Moreover, the results suggest that supporters of the CDU substituted to the Free Democratic Party – its traditional coalition partner – while adherents of rival parties chose the CDU in an attempt to hurt it. In total, at least 10% of voters in District 160 cast tactical ballots.

In 2012, Germany’s High Court put an end to absurdities of this sort and declared the electoral system to be unconstitutional. Ironically, the Court had come to the exact opposite conclusion just four years prior: in 2008, it had concluded that situations such as the one in Saxony were theoretical curiosities, which were unlikely to have real-world consequences. Spenkuch’s results show otherwise.

UK voters have long been accustomed to considering candidates’ chances of winning before voting for them. When only the candidate who obtains the most votes in a given district enters parliament, citizens effectively waste their vote by supporting anyone who does not stand a chance of winning.

While first-past-the-post elections are well known for giving rise to tactical voting, other electoral systems, especially proportional representation, are commonly believed to be immune to such issues. The new research shows that this is not so.

ENDS


Notes for editors: ‘Please Don’t Vote for Me: Voting in a Natural Experiment with Perverse Incentives’ by Jörg Spenkuch is published in the June 2015 issue of the Economic Journal.

Jörg Spenkuch is at the Kellogg School of Management at Northwestern University.

For further information: contact Romesh Vaitilingam on +44-7768-661095 (email: romesh@vaitilingam.com; Twitter: @econromesh); or Jörg Spenkuch via email: j-spenkuch@kellogg.northwestern.edu