New Delhi: A paper by an Ashoka University assistant professor, which suggests “manipulation” in favour of the Bharatiya Janata Party (BJP) in the 2019 Lok Sabha election, has got the academia, data scientists and psephologists split.
While many have vouched for the robustness of the methods employed and academic rigour, others disagree with the conclusions.
The political class is split predictably, with members of the ruling BJP dismissing the paper, and Opposition leaders seeking an explanation from the government and the Election Commission.
In his paper titled “Democratic Backsliding In The World’s Largest Democracy”, Sabyasachi Das, an assistant professor of economics at the Ashoka University, has suggested that the BJP won “disproportionately” in closely contested seats in 2019, especially in states where it was in power.
The study attributes this to vote manipulation.
“I find evidence consistent with electoral manipulation at the stage of voter registration as well as at the time of voting and counting (turnout manipulation),” Das says in his concluding remarks.
“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 adds.
He, however, says the tests are “not proofs of fraud”, nor does it suggest that manipulation was widespread. He adds that even if manipulation were proved in the seats studied, it would not have affected the outcome of the election.
The paper is yet to be peer-reviewed, and is believed to have come into circulation after Das presented it at an institute.
Ashoka University has been quick to distance itself from the research, saying that the paper had not yet completed a critical review process and had not been published in an academic journal.
ThePrint reached Das through email, but he declined to comment.
Also Read: BJP won 105 LS seats by 3 lakh+ votes in 2019, up from 42 in 2014. What it could mean for 2024
What Das’ paper says
Das says in his paper’s introduction that it seeks to contribute to the discussion on “democratic backsliding” — a retreat from democratic norms towards autocracy — that is a growing concern globally.
“This paper contributes to the discussion by documenting irregular patterns in 2019 general election in India and identifying whether they are due to electoral manipulation or precise control, i.e., incumbent party’s ability to precisely predict and affect win margins through campaigning,” it says. “I compile several new datasets and present evidence that is consistent with electoral manipulation in closely contested constituencies and is less supportive of the precise control hypothesis,” it adds, saying manipulation “appears to take the form of targeted electoral discrimination against India’s largest minority group – Muslims, partly facilitated by weak monitoring by election observers”.
“The results present a worrying development for the future of democracy,” it says.
In the 2019 Lok Sabha election, the BJP-led NDA returned to power with an even greater majority.
To arrive at his conclusion, Das says he used statistical tests such as the McCrary test and Benford’s law.
In simplified terms, the McCrary test is used to detect “discontinuity” in patterns as the result of an intervention.
Benford’s law, employed to detect fraud, suggests that there is an order to how numbers (1-9) occur in a genuine dataset, with 1 likely to be the most frequent digit and the rest following in a decreasing order. Any deviation is considered a suggestion of anomaly (this theory only applies to datasets where numbers 1-9 have an equal chance of occurring as the first digit — it won’t work, for example, on invoices for a product with a fixed price, or a dataset involving the weight of a group of adults).
The data sources used include aggregate election results, two different versions of EVM turnout data released by the Election Commission (the first of which, with partial data, was removed by the EC after discrepancies were pointed out, Das says), observers data, polling-station-level results, and the Muslim voteshare through voter lists.
According to Das, there appear to be significant irregularities in the election data.
In constituencies that were closely contested between a candidate from the BJP and a rival, the former “won disproportionately more of them than lost”.
Das says he did not find “similar discontinuities in the previous general elections for either the BJP or the Congress, the other major national party, or in assembly elections held simultaneously with the 2019 Lok Sabha election and those held subsequently.
Moreover, he says, the BJP’s “disproportionate win of closely contested constituencies is primarily concentrated in states ruled by the party at the time of election”.
The research paper suggests the electoral malpractice is at the local level and not widespread. It says the manipulation could be concentrated in constituencies that have a high share of observers who are state civil service officers from BJP-ruled states.
To examine “whether turnout manipulation was in part facilitated by weak monitoring of counting votes”, Das says he analysed “the assignment of counting observers across PCs (parliamentary constituencies)”.
“I compute the fraction of counting observers assigned in a PC who are from the State Civil Services (SCS), as opposed to the Indian Administrative Service (IAS). Since SCS officers are appointed by the state government, unlike the IAS officers who are centrally appointed, they are more likely to be politically pliable,” he says.
According to him, his study revealed that for “the fraction of SCS officers from BJP-ruled states, the discontinuity is larger in magnitude in PCs of BJP-ruled states”, while it is smaller and statistically insignificant for non-BJP-ruled states.
Additionally, “in PCs won by BJP, the fraction of counting observers who are SCS and come from BJP-ruled states positively predicts the extent of turnout data discrepancy in the PC; in PCs that BJP lost, no such relationship holds”.
However, Das goes on to say that “the paper is unable to comment on the overall extent of manipulation in the 2019 general election”.
“It focuses on closely contested constituencies as an empirical strategy to detect the presence of potential manipulation,” he says.
“Back of the envelope calculation shows that in PCs with BJP win margin less than 5%, BJP’s ‘excess’ win is in about 11 PCs. Therefore, even if all the disproportionate wins of BJP in closely contested PCs is due to manipulation, it likely would not have changed the government formation,” he adds. Nonetheless, the results “signify a worrying development for the future of democracy in India and consequently, in the world at large”.
Das is an alumnus of Yale University, where he completed his PhD in 2015, Delhi School of Economics and St Xaviers College, Kolkata. He specialises in political economy, public economics, and applied microeconomics. He has previously taught at ISI, Delhi, and has been teaching assistant in various undergraduate level courses at Yale University. His thesis subject in PhD was on the effectiveness of the institution of “Gram Sabha” in Indian village councils.
Das has focused largely on gender and caste in elections at the village level in his studies and has also highlighted the consequences of political alignments in relation to appointments of public servants.
Also Read: Why BJP’s keeping its 37 NDA allies close, even though they notched just 29 Lok Sabha seats in 2019
A host of reactions
In a tweet, BJP MP Nishikant Dubey dubbed the research “half-baked”.
“How can someone in the name of half-baked research discredit India’s vibrant poll process? How can any University allow it?” he said.
Meanwhile, Congress MP Shashi Tharoor tweeted that if “the Election Commission and/or the Government of India have answers available to refute these arguments, they should provide them in detail”.
“The evidence presented does not lend itself to political attacks on a serious scholar. E.g. the discrepancy in vote tallies needs to be explained, since it can’t be wished away,” he added.
Ashoka University teachers have maintained silence, but a former faculty member told ThePrint on the condition of anonymity that the paper “is carefully written and provides a lot of evidence for its claims”.
“It is also clear that evidence, while suggestive, is not ‘smoking-gun’ evidence of fraud and that the results would not change BJP’s victory in 2019.”
The paper, the former faculty member added, “points to genuine concerns about the fairness of national elections in certain contexts and the possible manipulation of electoral rolls in favour of the BJP — whether or not these constitute illegal activities or fraud”.
“In that sense, the paper provides the proper nuance for its claims, but the commentariat (of all political persuasions) are typically commenting without having read the paper. And for this reason, the paper has become controversial — either to oversell its findings or to try to discredit the work altogether.”
The former faculty member said it was “particularly disappointing to see academics make personal attacks without close engagement with the material in the paper, in a manner that hurts India’s academic credibility to score a few cheap political points”.
“Ashoka University’s performance in this has been shameful, as has the radio silence from the author’s colleagues in the economics department,” the academic added. “They should have affirmed the rights of their faculty to publish research, in addition to any critical content that engages with the paper in an academically rigorous manner.”
Mushfiq Mobarak, a professor of economics at Yale University, said the merit of a paper in social science “is not debated on the basis of the conclusions drawn, but on the basis of whether the methods employed were sound, whether the finding is rigorous and robust, and whether the data is high quality”.
“But most importantly for a paper like the one Dr Das presents, we examine whether the author is offering reasonable interpretations of the data, or if there are alternative explanations that would fit the same pattern observed in the data,” he added. “It is strange that the hordes of people complaining about this paper are not engaging with the data or methods in any serious or rigorous way. Instead they are using difficult to evaluate statements like ‘this does not pass the smell test’ to attack the findings.”
Anant Sudarshan, professor in the Department of Economics at University of Warwick, found the paper useful and important because, he said, it is necessary that academics examine election data regularly. That is the best way to ensure trust, he added.
“However, in my opinion, this paper does not show what the title claims, ie., it does not prove election fraud or ‘democratic backsliding’. The paper collects a lot of data and makes one key point, namely very close elections in 2019 disproportionately went in favour of one party, namely the BJP,” he said. “This is the only documented time this has happened and it is very odd because normally close elections are toss-ups and on average you win as many as you lose. But this does not necessarily imply manipulation.”
Practically “all the other evidence in the paper, which the author interprets as suggestive of manipulation, is also consistent with a party that is able to get last-minute voters to the polling booth when the election is very tight”, he added.
“In addition, others have pointed out that some of the discrepancies such as anomalies in registration numbers are in seats the BJP narrowly loses rather than those they won. Bottom-line, the paper is worth paying attention to, but it does not make a watertight case for fraud or manipulation. In fact, the fact that these disproportionate wins by BJP in close elections are seen only in one election (2019 Lok Sabha) could also be interpreted as evidence that this is not manipulation and instead due to a better ground game.”
If it were manipulation, Sudarshan said, “you’d expect to see it happen in state elections too, where presumably oversight is weaker and fraud would be even easier”.
Political activist and psephologist Yogendra Yadav hailed the study as the “first sophisticated and systematic documentation backed with empirical evidence” in the matter. “In my mind, it is in sync with what I get to hear on the ground,” he said. “Instead of making wild allegations, he is giving systematic reasoning for limited things. That is the core strength of the paper. The strength of the paper is that it is specific and not general. It looks at specific hypotheses with careful empirical evidence.”
Yadav criticised the university for distancing itself from the paper. “It is a comment on the nature of our times that a university has to disassociate itself from the findings of a research paper. Does it mean that the university owns all the other research papers? A scholar’s paper is a scholar’s paper and a university doesn’t own it.”
Sanjay Kumar of the Centre for the Study of Developing Societies (CSDS) stressed the need for more evidence.
“Some more evidence and data was required to make such a statement,” he said.
Yashwant Deshmukh, founder editor of C-Voter, said he had gone through the entire paper “as a data scientist”, adding that “very serious data crunching has been done but there are some critical issues”.
“Even if the entire hypothesis is concerned we are looking at BJP winning 11 Lok Sabha seats out of 543 through potential manipulation. He himself says that this would not have made any impact on the outcome of the elections. The entire paper is written in a way that kind of tries to prove that the 2019 mandate was manipulated which is not the case. He himself says there is no fraud.”
“Mamata Banerjee in West Bengal. Arvind Kejriwal in Delhi, Punjab, Stalin in Tamil Nadu assembly, Left Front in Kerala, Congress in Karnataka — all these elections are in everyone’s plain sight — even in these states, the majority of the marginal seats have been won by these parties and leaders. Are we trying to say all these people have manipulated elections?
“Regardless of whether you are in power or not, the fundamental truth is that the leading party always picks up the marginal seats,” he added. “On the issue of minority seats, it is a historical fact that the BJP tends to win more seats in Muslim areas. Their strike rate is extremely high because of the communal polarisation as turnout is higher. There’s nothing new that this has proven,” he added.
Aaditya Dar, a professor of economics at the Indian School of Business, said it was very common “in our discipline to circulate the paper for discussion before it’s submitted to a journal”.
“Looks like the paper was presented at NBER [National Bureau of Economic Research] Summer Institute just a few days ago, and that would explain the timing of this working paper,” he added. “Working papers go through revisions when they undergo peer review and that’s standard practice; I am looking forward to reading the final version in print when it is ready.”
(Edited by Sunanda Ranjan)