RECENTLY, as India’s battle against Covid-19 raged on, the New York Times published a damning report regarding the death and disease statistics that India was offering up as ‘official’ figures. According to the NYT article, various statistical models revealed that the Indian government was vastly undercounting the numbers of the dead and diseased. Official Indian figures had listed 26.9 million cases and 307, 231 deaths. Using their models, NYT statisticians said that even based on conservative estimations the actual number of cases was likely to be 404.2m cases and 600,000 deaths.

A more likely scenario (based on the premise that for every reported case there were 20 that were unreported and that the fatality rate was 0.3 per cent) was that there were 539m cases in India and that 1.6m people had died. Finally, the worst-case scenario (based on the premise that every reported case meant 26 unreported cases and a fatality rate of 0.6pc) suggested that a whopping 700.7m cases of Covid-19 infection and 4.2m deaths.

The Indian political class, dominated as it is by supporters of Prime Minister Narendra Modi and his BJP government, were furious at NYT’s audacity in second-guessing the government’s numbers. Sadly, with most opposition voices decimated in India’s emerging Hindu theocracy, it appears that only foreign media are able to raise such questions. The fact that the government numbers are very likely to be inaccurate is widely known given the discrepancy between cremated and buried bodies and the official death counts. But who can dare question a government that has arrested people for things like using Twitter to get blood donations for their family members or putting up posters critical of the Modi government?

Analysts have found that data in various official publications is often not the same.

The whole issue presents a problem that is common in South Asia. In a culture where the value of empiricism is not quite believed, the truth is what the powerful say it is. In Pakistan, this has traditionally been the Pakistani defence establishment and its various sponsored civilian leaders; in India it is the theocratic BJP. The desi powerful know that controlling the narrative of a crisis is more important to their goal of maintaining power than actually solving a crisis; they have misused this equation to such an extent that the fact of their manipulations of data is well known. In fact, few in the countries they control have any trust in the data that is provided on crucial matters.

It follows then that in the past several months when Pakistanis make note of the number of Covid-19 cases they include caveats like ‘according to’. These are apt qualifiers.

Data analysts that have taken a closer look at health issues have found for instance that data in various official publications is often not the same. Writing in this paper, Waqas Younas found that the Pakistan Social and Living Standard Measurement and the Pakistan Demographic and Health Survey did not provide the same data measurements for each indicator. Take for instance the category ‘deliveries with skilled birth attendants’ (in Sindh) in 2012-2013. The value was listed as 53pc in the PSLM and 61pc in the PDHS. Deliveries which have skilled birth attendants have much better outcomes for mothers and their babies, but service provision in this sector is impossible if no one knows how many are actually carried out.

Similarly, the indicator for fully immunised children (12 to 23 months, and again in Sindh) for 2006-2007 was also debatable; it was listed as 65pc in the PSLM database and only 37pc in the PDHS. The discrepancies are huge and it is impossible to tell which (if any) of the numbers are to be believed and made the basis of future calculations.

This lack of respect for statistical accuracy, can like most of our flaws, be pinned on colonialism. The British pioneered the premise of data as the means of controlling the population. Soon after they arrived, they set off a huge data-gathering operation that would become the basis of their administrative state. The numbers of this or that caste in each state, the numbers of Hindus and the numbers of Muslims, the exact size of the acreage belonging to this or that dissolute prince all were known to the white people who ultimately ruled the subcontinent for two centuries. The subcontinent’s introduction to data science (of the sort that enabled the British to shape the nature of ‘reality’) was not a friendly handshake.

That was a long time ago. We are at a point where we inhabit a world ruled by data where data is power and power we know is money. This means that instead of not presenting accurate data, in many cases fudging it, countries in the region must make peace with numbers and the necessity of having the correct figures. Super computers and advanced modelling software that exist today has made it possible to use acquired data to predict future needs and hazards a population could face. For instance, if there were more accurate statistics for health, governments would be better able to predict future surges in disease patterns and prevent some of them with targeted vaccination programmes.

Recognising the value of data and its integral role in our present and future also means we would be better able to evaluate tricky deals presented by foreign powers. Should we allow a certain country to lay fibreglass cables that transmit the data from millions of consumers? Would it be a better national security decision to ensure that a state maintains rights to the data of its citizens? Sadly, the way things are we will never know.

Government functionaries will keep postulating numbers that will please their immediate bosses in a perpetual chain of deception. The consequence is a subcontinent that is suspended in the past, in suspicions and recriminations from the colonial era, looking for easy wins and taking shortcuts to the sort of short-lived stints in power that win battles but lose wars.

The writer is an attorney teaching constitutional law and political philosophy.

rafia.zakaria@gmail.com

Published in Dawn, June 2nd, 2021