When the people of Uttar Pradesh cast their ballots in April, one characteristic promises to figure prominently in their votes: ethnicity. Before 1980, the Congress was the dominant party in the state and the nation, with very limited evidence of voter polarisation along either ethnic or class lines. Over the past 25 years, however, ethnic polarisation has played an increasingly dominant role in UP elections. Voters have exhibited a tendency to support parties that share their caste and ethnic affiliations.
Exit poll data from UP suggests that by 1999, the propensity for a voter to cast his vote for a party representing his ethnic group was close to 70 per cent. This polarisation of voting along ethnic lines has reshaped the state’s political landscape. It has ushered in an era of political dominance by low caste parties such as the BSP and the Samajwadi Party, in competition against parties such as the BJP and the Congress that receive support from upper caste voters.
What voters may not realise, however, is that this ethnicity-based voting has contributed to a second, more disturbing trend in UP politics: a sharp rise in the election of corrupt and criminal politicians. The criminal backdrop of UP politics is no secret. An analysis of affidavits filed before the 2004 national elections shows that a startling 27 per cent of the legislators elected from UP had a criminal record. Survey evidence collected by us shows that the fraction of UP politicians who either won or were placed second in the assembly polls and had a criminal record, such as convictions for rioting, assault or attempted murder, has more than doubled from nearly 8 per cent in 1980 to 16 per cent in 1996.
These simultaneous trends of criminalisation and ethnic parochialism are intricately related. The data suggests that the increase in corruption is attributable to the rise of ethnic parochialism and the resulting domination of majority ethnic parties. Importantly, the additional corruption is not associated with any particular caste. Instead, the study finds that while the identity of the dominant caste groups varies from one district to another, politicians elected from the party that is generally considered to represent the dominant groups in that district have become much more corrupt over this period. But this is not true of the politicians elected from the same areas from the other parties.
So, in an area where the share of the lower castes has historically been low, the BJP and Congress winners have become much more corrupt over the 1980-1996 period, unlike in the case of the SP and BSP winners. The opposite is true in low caste dominated areas. To worsen matters, the trend since 1980 has been that the winner in every area is more likely to come from the party representing the dominant groups.
The study suggests a simple explanation for this: candidates are judged on the extent to which they favour a voter’s particular ethnic group, as well as on the candidates’ perceived quality. Quality is measured by characteristics such as incorruptibility, respect for the law, competence, commitment and charisma. Since the policy pledges made by candidates during campaigns can always be broken once elected, voters are forced to rely on the ‘signals’ of perceived quality and ethnic party membership to predict how these candidates will perform — and, importantly, which groups they will favour — once they assume office.
What would happen if voters adjust the importance that they assign to these two ‘signals’ and start to place a stronger emphasis on candidates’ ethnic affiliation, as has happened in UP over the past 25 years? As more voters choose candidates along ethnic lines, the party affiliated with the dominant ethnic group will win a greater share of overall votes. As a result, the number of winning candidates from the dominant party will grow, while the number of competitive candidates from the minority party simultaneously shrinks — even if the quality of the candidates from each party stays exactly the same. This has the effect of lowering the quality threshold that candidates from the majority party must cross in order to win the election. At the same time, it ensures that minority party candidates must be of even higher quality in order to win.
The more that voters care about ethnic affiliation, the greater is their willingness to compromise on candidate quality. Consequently, the quality of the winners in the election will fall relative to the losers, and create opportunities for criminal and corrupt candidates from the majority party to enter office. But this effect is less likely to show up in constituencies reserved for scheduled castes. There, the contest is always between candidates with similar ethnic affiliation. So polarisation along ethnic lines matters less.
The study shows that dominant group party winners were significantly more likely to have a criminal record compared to winners from the same party in areas where it did not represent the dominant social group. The study also surveyed journalists and politicians in 102 representative constituencies to get data on a range of ‘politician-quality’ indicators (whether the politician had a criminal record on entering office, was associated with criminals, used his office for political or personal gain, was perceived to be corrupt, and whether he or his family made substantial financial gains after he entered office, etc). This is the data that was used to come up with the finding about the divergent trends in corruption among dominant group party winners and other winners. The study also found that the quality of winners has fallen relative to losers in all constituencies. As anticipated, these trends are absent in reserved constituencies.
For voters who will cast their ballot in April, selecting candidates who reflect ethnic identity may come at the cost of electing politicians of dubious quality. However, while ethnic-based parties continue to dominate UP’s political landscape, one hopes that the maturing of the SP and the BSP will create more inner-party competition for seats and thus throw up better candidates. And maybe, just maybe, the novelty of being represented by one’s own ethnic party is beginning to wear off.
Abhijit V Banerjee is Professor, Department of Economics, Massachusetts Institute of Technology
Rohini Pande is Professor, Kennedy School of Government, Harvard University