E-Paper
Terms of Trade | Nerds, trolls, statistics and politics
India’s public discourse, that way, is supremely un-educating. Opinions are sacrosanct, and facts, unnecessary
Prefer HTon Google
As someone whose primary job is to run a data journalism team in a newsroom, statistics are my constant companion. Some of them are straightforward, some not so much. When they are not, you need to look at the fine print. Doing this requires a lot of time, patience and most importantly, domain knowledge. None of these constraints, however, prevents shrill polemics around statistics when they are released.
This week has seen two such debates: one over India’s April-June quarter gross domestic product (GDP) numbers and another over the Special Intensive Revision (SIR) results for a number of states, including the national capital. The numbers have surprised many; some positively, some negatively. India’s GDP growth was 7.8% in a quarter when it was badly affected by the ongoing war in West Asia. As far as SIR is concerned, India has now deleted more people from its electoral rolls than the number of valid votes polled in Brazil's 2022 national elections. Delhi deleted more than a third of its voters.
Both of these numbers have their preconceived constituency in India and outside. Some believe the economy has been sinking, some think it’s a flying unicorn. Some think the SIR has been purging illegal voters who infiltrated the country from Bangladesh; others think all deletions are politically motivated to rig things in favour of the ruling party in the country.
Not even a small fraction of people peddling these narratives bother about the fine print of the data. India’s public discourse, that way, is supremely un-educating. Opinions are sacrosanct, and facts, unnecessary. This is not to say that the GDP or SIR numbers are necessarily absolutely kosher or set in stone.
Take the GDP data, for example. If one did a mechanical comparison with the GDP numbers for the quarter ending March 2026 released in June, headline growth is unchanged at 7.8%. Many newswires, which perhaps had prewrites ready, ran with this comparison. But the past numbers changed due to the adoption of a new deflator, and growth slowed by 0.8 percentage points between the March and June quarters. The entire GDP series itself was changed earlier this year. As with previous base-year revisions, the exercise has generated very different estimates from the previous series. Those interested in the nuances can read through thousands of pages of expert committee notes to understand what happened. Those waxing eloquent on social media will, of course, not read them. Some of India’s most credible private sector economists actually got the number right in their estimates. They used available data to arrive at their estimates. Indicators such as corporate profits and credit growth did well in the period. To be sure, the data does raise larger questions about the drivers of India’s growth performance in the recent past. We are yet to get a longer back-series data for the latest GDP series, which will likely be available by the end of the year. We will hopefully know better.
The SIR numbers are even more perplexing. 118 million deletions, so far, is an extremely large number. There is no way all of them have been wrongly deleted by the government. It also cannot be the case that more Bangladeshi adults migrated illegally to vote in India than there are adults in Bangladesh (estimated at about 117 million). These are the two extreme and very successful narratives we have been hearing about SIR. What is more likely to have happened is a mix of wrong exclusions, purging of the dead, etc., and de-duplication of voters who might have been registered in more than one place in India after migration. Our data journalism team has done multiple data stories since the beginning of the SIR process to underline these points.
The sad part is that, unlike GDP data, where we can at least hope to be wise in hindsight, we will perhaps never have better quality data around the SIR exercise. The Election Commission of India could perhaps learn a lesson or two from the Ministry of Statistics in this case. Right now, the only possible resolution could come once the 2027 census exercise is complete, and perhaps shows that India’s actual population is lower than what official estimates suggest at the moment.
This issue is not just about the GDP data or SIR numbers. It is about the larger poverty of our public discourse. Statistics are used in the most cavalier manner in India. Examples are galore.
Important government functionaries compare nominal numbers -- they do not account for inflation -- over time to claim that they are allocating more budgetary resources than their predecessors. They have enough people working for them who understand these numbers to tell them that they are wrong. Similarly, opposition leaders allege large-scale irregularities in electoral rolls when they lose elections. Such alibis go along with glaring organisational lapses in local-level politics and absolutely no political introspection about these deficiencies. Reality often gets buried in such extreme polemics.
More than a million people were removed from electoral rolls, pending adjudication, in West Bengal and are now at the mercy of India’s notoriously delayed judicial process. This is a shocking subversion of the spirit of India’s universal franchise. There is a very high possibility that many poor people have been wrongly excluded from electoral rolls elsewhere due to the heavy documentary burden that the exercise entails. In an ideal world, the SIR debate should have focused on the asymmetry between lack of state capacity to do this exercise effectively and the disproportionate burden it places on people on the margins. On this front, the judiciary deserves the biggest indictment.
The economy is no different. Debates around GDP and per capita GDP notwithstanding, palliatives of similar denominations are used to woo voters in elections in India’s richest and poorest states. India’s headline economic numbers mean very little, even if they are absolutely statistically robust, for most of its people. But these are structural issues. They are too complicated to engage trolls. The trolls move promiscuously from one number dominating the news cycle to another. Unlike the sensationalism around numbers on a particular day, India’s bigger challenge is solving structural problems. This will also require educating the public at large about their seriousness.
The trolls are not bothered about all this. The nerds can continue to be sad about the lack of data or keep writing about the nuances involved without any larger traction. Our public discourse today works on what is best described by paraphrasing Keynes: in the short run, we are all trolls, even inadvertently, and in the long run, we are all dead.
(Roshan Kishore, HT’s Data and Political Economy Editor, writes a weekly column on the state of the country’s economy and its political fall out, and vice-versa)
