Chennai: World Health Organization estimates released on Thursday suggest that India experienced 4.7 million excess deaths in 2020 and 2021, as compared to the country's official Covid-19 toll of 481,000 for the same period. The Indian government has rejected the findings, making it the only one of the WHO's 194 member states to push back against these estimates.

The WHO's estimates suggest that while India experienced roughly 830,000 excess deaths in 2020 (by the end of its first wave), the country saw over 3.9 million excess deaths in 2021. Of these, 2.3 million took place over just two months--April and May 2021--at the height of India's brutal second wave driven by the Delta variant.

The WHO's estimates place "the full death toll associated directly or indirectly with the Covid-19 pandemic (described as "excess mortality")" between January 1, 2020 and December 31, 2021 at 14.9 million deaths (within a range of estimates from 13.3 million to 16.6 million) for the world.

By the WHO's estimates, India experienced the largest absolute excess mortality numbers of any country in the world, over four times the excess mortality seen by the next most affected countries, the Russian Federation and Indonesia.

However, proportionate to population, Peru had the highest excess mortality for the 24-month period, the WHO estimates show, with Russia in fifth place. Relative to population, India was 33rd among countries in terms of excess mortality, with 171 excess deaths for every 100,000 people during the two-year period.

Globally, as well as in India, men accounted for 57% of estimated excess deaths. In India, those over the age of 60 accounted for 78% of excess deaths, while globally 82% of excess deaths were among those aged 60 and above.

Excess mortality includes deaths associated with Covid-19 directly, as well as indirectly from the pandemic's impact on access to health and support. It may also be affected by deaths averted by the lowered likelihood of road accidents and lowered emissions as a result of interruptions in economic activity.

It is, thus, the difference between the number of deaths that have actually occurred and the number that would be expected in the absence of the pandemic, based on data from earlier years.

Both sets of estimates involve some demographic and statistical assumptions that are at the heart of the Indian government's objections to the WHO's estimates. Despite the WHO beginning this process in February 2021, the Indian government did not share all-cause mortality data for either 2020 or 2021 with them.

As a result, the WHO used monthly all-cause mortality from 18 states and Union Territories collected by Indian journalists from state-level Civil Registration System portals, and extrapolated them for the country, to prepare their estimates. Then, on the evening of Tuesday (May 3), less than 48 hours before the WHO's estimates were expected, the Indian government released Civil Registration System data on deaths in the country in 2020.

The trends in the 2020 data collected by Indian journalists match those in the government's data for 2020, historian Chinmay Tumbe, who has written a book on pandemics and co-authored a paper on excess mortality from Covid-19 using the numbers, said.

"We have seen that the Indian government has released their mortality data for 2020 on Tuesday," Samira Asma, Assistant Director-General for Data, Analytics and Delivery at the WHO, said in response to a question from IndiaSpend. "We haven't yet been able to look at and compare the data. Our estimates are for 24 months, while the Indian data is for 2020 only. It will take time for us to have a careful look at the data and run it through our models.

"[Our] estimates use the best available data and have been produced using a robust methodology and a completely transparent approach."

Asked by IndiaSpend for a comment, a health ministry spokesperson directed us to the government's remarks from last week pushing back against the expected numbers, as also the newly released 2020 CRS data.

Following the release of estimates on May 5, the ministry issued a statement where it reiterated its objections (see IndiaSpend's report here for more details), calling the methodology "statistically unsound and scientifically questionable". "India has been consistently objecting to the methodology adopted by WHO to project excess mortality estimates based on mathematical models," the ministry said. "Despite India's objection to the process, methodology and outcome of this modelling exercise, WHO has released the excess mortality estimates without adequately addressing India's concerns."

Further, referring to the CRS data, the ministry said, "India firmly believes that such robust and accurate data generated through Legal Framework of a Member State must be respected, accepted and used by WHO rather than relying on less than accurate mathematical projection based on non-official sources of data."

The CRS data released by the Indian government on Tuesday show that India registered 8.1 million deaths in all in 2020, which is an increase of 470,000 deaths over 2019. However, this increase is smaller than the increase between 2018 and 2019, for instance, leading NITI Ayog member V.K. Paul to claim on May 4 that there had been no significant increase in deaths in India during the pandemic.

The CRS data--which are for 2020 only; 2021 data will be made available next year--however do not estimate how many deaths took place in India in 2020. They only look at how many were registered. India's Sample Registration System (SRS, which studies a representative sample of households) produces estimates of how many deaths actually took place in a country in a given year. However, the SRS for 2020 was not conducted.

India's officially recorded deaths were problematic even before the Covid-19 pandemic: nine out of ten deaths in India were registered in 2019, but rural areas and less developed states reported lower levels of registration. Only a fifth of deaths in 2019 were medically certified, and more often than not with the underlying cause of death recorded incorrectly, as IndiaSpend reported in June 2020.

Data from past SRS reports show that prior to the pandemic, mortality in India was flattening, with the death rate in decline and population growth slowing. Absent a substantial increase in registrations, India should not have expected a large uptick in deaths in 2020. The WHO's estimates of expected (in the absence of the pandemic) mortality too show only a very slight increase in how many deaths India should have seen.

Moreover, while Paul claimed that registrations improve every year, it isn't apparent that the registration of deaths would have increased during the pandemic. Some states like Telangana have reported to the CRS that citizens faced severe difficulties in reporting deaths on account of the pandemic. Additionally, registered births actually declined which, given that fertility is not yet on the decline in India, could be an outcome of depressed registrations, demographers pointed out. "Until we have the SRS data, it is difficult to say how much of the increase in mortality is on account of increased registration, and how much as a result of the pandemic," Usha Ram, Professor in the Department of Biostatistics and Epidemiology at the International Institute of Population Sciences in Mumbai, and an expert at mortality estimation, told IndiaSpend.

(Nileena Suresh, data journalist at IndiaSpend, contributed to this story.)

Update: At 6.45 p.m. on May 5, the story was updated to include the health ministry's rebuttal to the WHO's estimates.