The lost Covid-19 count: How could the govt only tally a sixth of all who died?

That the number of Covid deaths were likely underreported did not need expertise to suspect — even the general public thought it quite likely, between the lockdown, the state of the public health infrastructure of the country in general, and lack of public awareness.

Many were likely even sympathetic of the job public health officials and administrators had to deal with here, given the shape of the lockdown and the state of (un)preparedness — and we weren't even alone in the world with our struggles.

Why, just last month, even was being bemoaned five years after the first wave.

As for India, the BBC had reported back in 2022 on a WHO study that the truth may be — around 47 lakh, vs the official toll reported for 2020 and 2021 together, being 4.8 lakh. In 2021, CNBC-TV18 had cited a study that suggested by a mere 4.3 lakh, however.

Well, now we know it wasn't that bad (maybe) — or, depending where you started, it is much worse.

New demographic data released by the Union government on 7 May suggests the true extent of the undercount of Covid-19 deaths in India. The Civil Registration System (CRS) — a record of all registered births and deaths in India collected by year — shows there were over 21 lakh more deaths registered in 2021 compared to 2020. This is about six times the official Covid-19 death toll for 2021.

Keep in mind here that while registration of births and deaths is mandatory for a number of legal reasons, both do regularly get put off — or simply, not get done. And in the mayhem that was Covid, the barriers were all the higher.

Then again, registration percentage had been on an upward trend in the pre-Covid years, from an estimated 70 per cent in 2013 to 77 per cent in 2016 and 92 per cent in 2019. So perhaps the disparity isn't that bad? But we don't really know. 

Ideally, deaths and births are meant to be reported year on year. But the chaos of the Covid years likely overwhelmed whatever system we had in place.

Who were they: the counted or the lost?

But also, there was another set of data released on Wednesday — that from the Sample Registration System (SRS), a parallel exercise that estimates the number of births and deaths in the country, not just those actively registered under the CRS.

Also available now is the 2021 report for the medical certifications of cause of death — again, this is not always available in real-world India, though in an ideal world each death would be (1) certified by a medical practitioner and (2) the cause identified as Covid or not at a time when it really mattered.

So between the three, we now have a more credible estimate of the actual dead. However, we still don't have the complete picture.

One reason is that the peak pandemic period was really three years: 2020–22, and we don't have the CRS data for 2022 still.

For 2020, the CRS lists 81.15 lakh registered deaths. Now, if we assume the same percentage of registration as for 2019 (92 per cent), as this does, we might have had 88.21 lakh deaths in actual fact.

That is a disparity of about 4.7 lakh from our typical annual average of 83.5 lakh deaths (the average over the last 15 years) — and is thrice the official Covid-19 toll for that year, reported as 1.48 lakh. And of course, that gulf is hard to account for without ascribing at least a good part of it to the coronavirus.

Now admittedly, this is a rough back-of-the-envelope figure. There may be confounding factors or causes other than Covid-19 that affect the ‘death rate’ that we have not accounted for. For one thing, access to medical care itself was disrupted during the pandemic, so deaths from other causes can be expected to be higher than in prior years as well.

It is also quite possible the registration rate is different from the last year's — and again, that may be upward (following the prior trend) or downward (given Covid chaos)

But now, let's look at 2021.

If we assume that the rising trend of registering deaths held at over 90 per cent even in the Covid years — perhaps hypervigilance where there was healthcare access was offsetting lack of accessible medical care in other places? — then we might, for once, take the official 1.2 crore deaths registered in 2021 at face value.

This is 18.75 lakh deaths over the expected average. Which is approaching six times the official 2021 death toll of 3.3 lakh.

So now, even if we don't try to dream up or adjust for other possible causes, say we attribute this to even mostly Covid-19 — not all, just most — adding up the 'excess' deaths of 2020 (4.7 lakh) and 2021 (18.75 lakh) should take us to at least 20 lakh or so Covid-19 deaths over the both years.

Remember, now, we don't have data for the third year, by which time the mortality rate had also fallen.

The good news, if you can bear the term for such a grim figure, is that a ballpark of say 22-25 lakh deaths is at least less than half the WHO's 47 lakh estimate. The bad news is, this suggests undercounting so massive that the official figure of 4.8 lakh is only a quarter of the actual figure.

There's also one other data set we haven't considered — the medically certified Covid figures.

Typically, only 23 per cent of deaths are in fact medically certified in India.

In 2021, we have about 23.95 lakh medically certified deaths, of which 4.13 lakh were attributed to Covid-19. So, assuming this is less than a quarter of the actual figure, how much of the 1.2 crore might we extrapolate as being caused by Covid-19?

Again, we arrive at a number over 16 lakh — and in truth, this may be closer to or slightly over 20 lakh, if we consider the lack of healthcare access during lockdown and the medical resource crunch we saw in the pandemic.

May 2022: India slams WHO's excess COVID deaths in India claim.
July 2024: India dismisses 'Sciences Advances' report on India's COVID excess deaths.
May 2025: India's official data shows 20 lakh excess deaths, nearly six times the official Covid death tally of 3.3 lakh in that… pic.twitter.com/dLCNoTM6IM

— Karthik 🇮🇳 (@beastoftraal) May 9, 2025

How do we account for such rampant underreporting?

And why are we getting this data from the registrar-general of India only now, after a five-year gap?

Then there's another little confounder. Typically, even the CRS report used to include the SRS' estimated deaths and births for the year. This time, we don't have it. That's right, even the SRS figures reported don't include the baseline estimate of deaths and births for 2020 and 2021, an odd lacuna.

There is also — to add to this sorry state of affairs already — an unevenness of distribution. It's a hard thing to reflect on, but worthy of considering if only to recalibrate public health and healthcare efforts at the state level.

For example, in Gujarat, we have nearly 2 lakh excess fatalities, more than 33 times the state's official count of just over 5,800 reported Covid deaths. Whereas at the other end of the table, Kerala, Uttarakhand, Assam, Maharashtra and Delhi have 'excess death' figures of only 1.5 to 3.1 times their official reported toll.

In between, we have Madhya Pradesh (18 times the official figure attributed to Covid), West Bengal (15 times), and then Bihar, Rajasthan, Jharkhand and Andhra Pradesh, each with a difference in mortality figures over 10 times the official Covid-19 count.

Again, we are looking at 'all deaths we didn't expect' vs Covid-19 deaths, but also again, we can assume a high proportion of the 'excess' does in fact represent the coronavirus' toll.

Meanwhile, there's another question implicit in the report for the medical certification of cause of death: in 2021, Covid-19 was only the second most common cause of death, accounting for 17.3 per cent of deaths that year.

The leading cause (29.8 per cent) was cardiovascular disease — but from what we have since learnt about the coronavirus' impact on blood clotting, comorbidities, and 'long Covid', it may well be that the percentage again represents significant underreporting due to misattribution.

The figures come out at a time when the nation is preoccupied on the front of national security. However, it is to be hoped that the medical community as a whole as well as civic watchdogs and Opposition parties will be considering these figures and asking some hard questions soon.

Also, we wonder what the long-overdue census might uncover. Will it throw up even more discrepancies?

And who will tell us why we have more questions than answers from these reports?