A study by Valentino Larcinese examines the channels through which voters keep politicians accountable. Using the expenses scandal as a source of data, he finds that it was scandal-related press coverage which affected negatively the 2010 performance of standing MPs. Punishment in the ballot box, however, was relatively small and generally not sufficient to remove MPs involved in the scandal from their seat. One of the reasons why in general elections the punishment of corrupt politicians can be modest is because ideology matters and many voters may still prefer a corrupt politician to an ideologically more distant alternative.
An important function of democratic institutions is to make public officials accountable to citizens. This control works through incumbents’ fear of the next election and by offering voters the opportunity to “throw out the rascals”. As we approach a general election characterised by party fragmentation and low trust in politicians, it is useful to assess whether elections do in fact purge the political class by removing corrupt politicians. In an attempt to answer this question we have studied the impact of the 2009 expenses scandal on the outcome of the 2010 general election in a paper co-authored with Indraneel Sircar of Queen Mary College (University of London).
The scandal offers an ideal setting to study accountability channels in some detail and to identify some of the causal links at play. First, the scandal involves a well-defined set of political actors, namely the members of parliament who were in office in May 2009, who all faced the same rules and constraints regarding their expenses. Second, the scandal erupted within a short time frame and focused on the same issue for all MPs involved, namely abusing the allowance system. These two features make scandal involvement comparable across MPs and provide important advantages compared to either cross-country studies or studies that, even within a country, compare scandals which occurred in different periods, concerning different sorts of political actors and different types of wrongdoing. Moreover, the scandal was salient in public debate for several months and was followed by a general election only one year after it began.
Following the scandal, an investigation chaired by Sir Thomas Legg was held that led to an accurate reconstruction of the amount misappropriated by each MP in the February 2010 ‘Review of past ACA payments’ (hereafter ‘the Legg report’). This provides another characteristic of the scandal that makes it particularly suitable for empirical study: the availability of an objective, accurately defined measure of monetary wrongdoing.
Although most theories tend to study accountability mechanisms by focusing on a simplified voter-politician relationship, democratic processes rely on a number of actors who often play a crucial role in the process of “throwing out the rascals”. Our empirical analysis takes the complexity of the accountability process into account and studies the scandal from a variety of angles. We analyse three key links in the accountability process: 1) a link from malfeasance to news, with respect to which we ask questions about possible media bias and the role performed by media outlets as watchdogs of power; 2) a link from news to perception, with respect to which we ask how partisanship and other individual characteristics affect the way news are processed and incorporated into perceptions about MPs; 3) a link from perceptions to action, whereby voters punish corrupt politicians in the ballot box, or expected punishment induces politicians to stand down (or political parties to de-select corrupt MPs).
We collected data on the coverage of the scandal in seven UK newspapers (including the Sunday editions): the Daily Telegraph, The Guardian, The Times, The Independent, The Sun, Daily Mail, and the Scotsman. The sample of newspapers was selected to include widely read national broadsheets and widely read national tabloids, along with an important regional newspaper (the Scotsman), as well as in order to have sufficient ideological variety. Two indicators were used to gauge the media salience of each individual MP’s involvement in the expenses scandal. First, we use the number of articles in which an MP’s name appears alongside the word ‘expenses’ in the period from 8th May 2009 to 7th August 2009 (i.e. for three months after the Telegraph revelations). However, since some MPs naturally had a higher profile, and therefore attracted more coverage, whether related to scandal or not, we also take into account in the statistical analysis the number of articles in which the MP’s name appears during the three months preceding the scandal.
Our first finding concerns the British mass media: our results provide a rather benign view of the British press, whose coverage of the scandal appears to have been positively linked to actual monetary wrongdoing (from the Legg report), and mostly focused on government members. We find little evidence of a partisan bias in the sense that newspapers were not giving comparatively less coverage to ideologically close MPs.
Then we find that MPs who received more attention in the press in relation to their expenses were also, ceteris paribus, more likely to stand down and not run in the 2010 election. Those MPs that, in spite of having been involved in the scandal, decided nevertheless to run in 2010, were punished in the ballot box: scandal-related press coverage affected negatively the 2010 performance of standing MPs. Punishment in the ballot box, however, was relatively small and generally not sufficient to remove MPs involved in the scandal from their seat. Hence, the main mechanism to “throw out the rascals” appears to have been pre-election politics. We can assume this was a consequence of pressure from party leaders and party members as well as of the fear to face the voters again.
This is in line with findings from the US where primaries have been found to be the main mechanism to remove corrupt politicians. In primaries ideological considerations are less important than in general elections and, therefore, being involved in a scandal has maximal negative effects on candidates. In the context of British politics this means that greater attention should be given to the processes that lead to candidates’ selection. This often obscure process can be more important than the actual election in determining the quality of the personnel who gets eventually elected to represent the British voters.
One of the reasons why in general elections the punishment of corrupt politicians can be modest is because ideology matters and many voters may still prefer a corrupt politician to an ideologically more distant alternative. By using data from the British Election Study we corroborate this conclusion by providing evidence of a strong effect of partisanship on the voters’ perception of their MPs’ honesty. Ceteris paribus, voters perceive co-partisan MPs to be less involved in the scandal. This is an instance of cognitive dissonance: not only voters trade off ideology with valence but they also tend to underestimate bad information concerning co-partisan MPs and possibly overestimate bad information on MPs from the opposite side of the political spectrum. Our evidence points to the complexity of the role played by partisanship in voters’ mind and shows that biased perceptions and sticky beliefs can represent formidable obstacles to accountability.
Last but not least, we uncover a significant gender bias: we provide robust evidence that female MPs were subject to higher scandal-related coverage in the press, had a higher probability to stand down as a reaction to press coverage, and suffered higher loss of votes in 2010 compared to 2005. All these results are obtained ceteris paribus, i.e. for the same level of money actually misappropriated, same party affiliation etc. Our evidence shows that female MPs were more vulnerable in many ways during the scandal and suggests that the path to a balanced gender representation can be more complex than what is often assumed, involving both external pressure (higher for women, as evidenced by higher press coverage) and internal constraints (women MPs were overall more likely to voluntarily leave Westminster if involved in the scandal).
To conclude, our evidence provides support for theories that stress the importance of information availability for a well-functioning democracy. The disclosure of information on MPs’ detailed expenses items led to a wave of resignations and eventually to voters’ punishment of the most involved MPs. Crucially, we find that, while information available on the press matters for resignations and electoral returns, an objective monetary measure of wrongdoing does not: the amount of money misappropriated (from the Legg report) does not independently explain patterns of resignation and does not affect voting returns. It is only through press coverage that misbehaviour led to electoral punishment: MPs who were involved in the scandal but not covered in the press managed to scrape through the 2010 election keeping their seat. Independent and plural mass media are an essential ingredient of democracy: politicians’ accountability crucially rely on them acting as watchdogs of power.
Note: This article gives the views of the author, and not the position of the British Politics and Policy blog, nor of the London School of Economics. Please read our comments policy before posting.
Valentino Larcinese is Associate Professor (Reader) in Public Policy in the Department of Government, LSE.