Does crime pay for politicians in India?

A 2003 Supreme Court ruling in India made it obligatory for all political candidates to declare criminal record to the Election Commission. This information was then collated by NGO Education Watch and released to the media. The Education Watch data reveal that a substantial fraction of Members of Parliament (MPs) have a criminal record.

An interesting question to pose is whether the revelation of criminality affected the electoral chances of politicians in the subsequent national election in 2004. We have been assembling electoral data for the 1999 and 2004 elections which will allow us to look at two substantive issues. The first is whether for those who won in 1999 there was a penalty for being revealed to having had a criminal record. The second is to look at candidates within constituencies to see whether votes are flowing disproportionately to those who have been revealed to have clean records.  

Quite a number of countries are moving to trying to promote accountability by making characteristics of politicians more transparent. Whether this works in terms of punishing criminal politicians (and hence encouraging clean politicians) is unknown. As we have data on the nature of crimes it will also be interesting to look at whether politicians are electorally punished for committing particular types of crimes. This work will begin to get into these important issues.  

We have been working with an Indian NGO Election Watch to assemble a comprehensive data base on the criminal records of all candidates who contested the national election in 2004. From this data base we are able to ascertain which types of crimes were committed and also the number of crimes. We will develop a typology of crimes committed to see whether different types of crimes attract different types of electoral responses. Data on electoral results are being assembled from the Indian Election Commission. There is information on who won the elections (as well as runner ups) , party affiliation, and share of votes won. Splicing together the Election Watch and Election Commission data will allow us to examine whether the revelation of criminality (and of the type of criminality) affects the propensity to be re-elected and also the share of votes won. From data sources within the National Indian Parliament we will also assemble data on the backgrounds of different Members of Parliament so that we can also examine which politician characteristics are associated with the propensity to commit crime.