New Delhi: A research paper by a faculty member of Ashoka University has triggered a political row by suggesting that there was a possible manipulation of results in some seats by the ruling Bharatiya Janata Party (BJP) in 2019. The paper says that this manipulation resulted in the party winning a disproportionate share of closely contested constituencies in the 2019 Lok Sabha election.
The paper titled Democratic Backsliding in the World’s Largest Democracy was authored by Sabyasachi Das, assistant professor of economics at Ashoka University.
Following the publication of the paper, the author received backlash on social media from pro-BJP handles. In an unusual move, the university publicly disassociated itself from the research paper.
“Social media activity or public activism by Ashoka faculty, students or staff in their individual capacity does not reflect the stand of the University,” said Ashoka University on Twitter on August 1. It also questioned the scholarly credentials of the paper: “To the best of our knowledge, the paper in question has not yet completed a critical review process and has not been published in an academic journal.”
The university has been widely criticised by scholars in India and abroad for throwing their faculty member under the bus and belittling his research paper, even as they applauded Das for his controversial yet ‘bold’ conclusions.
The findings
The 50-page research paper finds that the number of these “excess” BJP wins in close fights in 2019 varies from nine to 18 Lok Sabha seats. The back-of-the-envelope calculation shows that the party won 11 of these seats in which its win margin was less than 5%.
While the figure is much less than the BJP’s required half-way mark of 272 to form a government on its own, the author alarmingly concludes through data analysis that electoral fraud even in a single constituency would imply that such manipulations by incumbent parties are possible. This raises broader questions about the future of electoral democracy in India.
The author makes it clear that he cannot comment on the “overall extent of manipulation” in the 2019 election and focuses on closely contested seats as an empirical strategy to detect the presence of potential manipulation.
The 2019 election gave Prime Minister Narendra Modi a second term, with a large majority as compared to 2014. The BJP secured 303 out of 543 seats nationwide.
In several recent elections, especially in Uttar Pradesh where the BJP has a large number of its MPs and MLAs, opposition parties have accused the incumbent BJP of misusing administrative machinery to influence the elections. They alleged that the saffron party deleted unfavourable names from the voter lists and rigged results.
These allegations run parallel to the claims, though technically unsubstantiated, made by opposition parties that electronic voting machines (EVMs) were being tampered to favour the BJP.
Interestingly, Das in his paper writes that he does not consider the possibility of manipulation of EVMs as a mechanism of electoral manipulation, as given their technology, it is hard to manipulate them at scale. Instead, he argues, electoral manipulation happens at the local level of polling booths and that it was unlikely that the Election Commission of India would engage in direct tampering of turnout data ex-post.
The paper by Das claims that the 2019 general election showed “significant irregularities in the election data”. In technical terms, he is saying that “the density of the BJP’s win margin variable exhibits a discontinuous jump at the threshold level of zero.” Simply put, this implies that the BJP went on to win many more of those ‘closely contested’ seats than it lost.
Das did not find similar “discontinuities” in the previous general elections for either the BJP or the Congress, its biggest national rival, as well as for state assembly elections held simultaneously with the 2019 Lok Sabha election and those held subsequently.
He put together several new datasets in addition to accessing the candidate-level general election results for 1977-2019 and state assembly election results for 2019-2021.
The paper finds that the BJP’s disproportionate win of closely contested constituencies was primarily concentrated in states ruled by the party at the time of election.
Das also indicates that manipulation appears to be concentrated in constituencies that have a high share of observers, who are state civil service (SCS) officers from BJP-ruled states. Unlike the IAS, who are centrally appointed, the SCS officers are more likely to be politically pliable, he argues.
He reaches a conclusion by testing with evidence two different hypotheses: Was there electoral fraud or manipulation? And did the BJP clinch these tight contests simply because it worked harder on them after accurately predicting the possible outcome?
Through a range of arguments, he concludes that he found evidence consistent with electoral manipulation. He put to test the theory of ‘precise control’, the other likely possibility, to investigate if the BJP had managed to win these close fights due to its superior campaigning abilities and resources.
Precise control refers to the possibility that the BJP accurately predicted these tough seats and secured them through better campaigning and mobilisation of party workers to shape voter attitudes at the local level.
However, Das – who studied the National Election Survey of 2019, post-poll voter survey conducted by the Centre for the Study of Developing Societies, that gives micro data on election campaigning by political parties – did not find any evidence to support that. He found that the BJP or any other party did not campaign significantly harder through door-to-door visits in constituencies that the saffron party won by a thin margin.
On the other hand, he found evidence consistent with electoral manipulation at the stage of voter registration as well as at the time of voting and counting (turnout manipulation).
“In both cases, the results point to strategic and targeted electoral discrimination against Muslims, in the form of deletion of names from voter lists and suppression of their votes during election, in part facilitated by weak monitoring by election observers,” he writes.
Electoral discrimination
The paper says that electoral manipulation can take place at the state of voter registration (registration manipulation) or at the time of voting or counting (turnout manipulation). These could be done through strategic deletion of Muslim names from the list of registered voters or electoral rolls, or through strategic suppression of Muslim votes at the time of voting or counting.
He concludes that there was registration manipulation by showing that the growth rate of the electorate (number of registered voters) for each parliamentary constituency between 2014 and 2019 falls discontinuously by five percentage points in constituencies barely won by the BJP. The fall is concentrated in constituencies with a higher share of Muslim electorate.
To test turnout manipulation, Das examined the absolute difference between the two official versions of EVM turnout data for discrepancies. He found that there was a concentration of large discontinuous increase in the magnitude of data revision in the BJP-ruled states. This, he interpreted, was indicative of manipulation done locally at the polling stations rather than resulting from aggregation fraud at the constituency level.
He computed vote share of the BJP at each polling station relative to its vote share in the parliamentary constituency. This is the relative BJP vote share. He found that the spike in the relative BJP vote share was higher in seats with larger discrepancy in turnout data. In those seats barely won by the BJP, the party’s relative vote share exhibits a large spike in polling stations with high turnout, according to him.
The Lok Sabha seats the BJP barely won have a 26% larger likelihood of having a “large” mismatch in the EVM turnout data than the seats that the BJP barely lost. This implies that the sample of closely contested constituencies that were disproportionately won by the BJP also has a disproportionately higher likelihood of “large” turnout revision.
Das also tests the question of manipulation in the form of targeted electoral discrimination against Muslim voters. If the high vote shares of the BJP, which is not viewed favourably by Muslims, in seats it barely won were concentrated in areas with higher presence of the minority group, this would indicate manipulation.
On the other hand, if this was due to better targeted campaigning by the BJP, we would expect the opposite, as the BJP’s increase in vote share from 2014 to 2019 came primarily from Hindus, especially lower caste groups, while its support among Muslims was low in both elections.
In this context, he finds that in parliamentary constituencies that the BJP barely lost, its vote share is less likely to exceed the 95th percentile in polling stations located in assembly constituencies with a high Muslim share, within the Lok Sabha area. This negative relation gets significantly reduced in parliamentary constituencies the BJP barely won.
In those parliamentary constituencies, the likelihood of the event does not fall in assembly constituencies with higher Muslim share. “This again supports the manipulation hypothesis,” he writes.
However, despite forcefully arguing in favour of possible manipulation, he cautions that the tests are not proof of fraud nor do they suggest that manipulation was widespread.
“Proving electoral manipulation in a robust democracy is a significantly harder task that would require detailed investigation of electoral data in each constituency separately,” he says.
Reacting to the controversy over the research paper, Ashoka University said it was “dismayed by the speculation and debate” around Das’s paper and the university’s position on its contents.
While stressing that it values research that is critically peer-reviewed and published in reputed journals, Ashoka University said, “To the best of our knowledge, the paper in question has not yet completed a critical review process and has not been published in an academic journal.”
The University added that it encourages its 160-plus faculty to carry out research, but does not direct or approve specific research projects by individual faculty members.
BJP MP from Jharkhand,Nishikant Dubey, in response to a tweet by Ashoka University, said it was “fine to differ with the BJP on matters of policy but this is taking it too far”.
“…how can someone in the name of half-baked research discredit India’s vibrant poll process? How can any university allow it? Answers needed – this is not good enough a response,” he said.
Congress MP from Kerala Shashi Tharoor, referring to the conclusions in the paper, said that if the Election Commission and the Union government have answers available to refute these arguments, they should provide them in detail.
The paper “offers a hugely troubling analysis for all lovers of Indian democracy,” he said, specifically demanding that the “discrepancy in vote tallies need to be explained, since it can’t be wished away.”
The Congress leader also backed the scholar Das over “political attacks” faced by him over the research.
