It has been more than a year since the Covid pandemic first struck India. While uncertainties may exist as to whether the virus emerged naturally or was man-made, one thing is certain — Covid has been weaponised by some researchers, if not by sovereigns. Engaging in leaps of inference, interpretation and imagination and relying on unsuitable data, they have sought to estimate Covid deaths. The most recent production in this genre is a paper by Abhishek Anand, Justin Sandefur and Arvind Subramanian (Three New Estimates of India’s All-Cause Excess Mortality during the COVID-19 Pandemic, Working Paper 589, Centre for Global Development, July 2021).
First, underreporting of deaths in India is a reality. The gap between estimated and registered deaths is declining, but is still high. As per the report, Vital statistics of India as per the Civil Registration System 2019, the level of registration of deaths was in the range of 50 to 80 per cent for vast swathes of India in 2019. Further, of the 31 states and Union territories combined, only 11 of them had deaths registered within the mandated 21 days in 2019. In 2020 and 2021, with lockdown and movement restrictions, even longer delays are both inevitable and unsurprising. So, the level of excess deaths estimated will vary depending on when one chooses to estimate them. The difference between estimated deaths and registered deaths was 6.6 lakh in 2019, with an average difference of around 15.4 lakh per year for the five-year period ending 2019. There is really nothing to see here.
In the light of this, let us examine the three methods the authors use to estimate all-cause excess mortality in India during the pandemic.
The first of their methods relies on CRS which is maintained by states. They use data from seven states without disaggregating them. However, data for 14 states are available for 2020. Based on an analysis done by a reliable source (to be published soon) on the data on registered deaths from these 14 states, it can be stated that the growth in all-cause mortality in 2020 is neither below nor above the historical trend growth rate of 4 per cent.
The second method their paper uses is the infection fatality ratio (IFR) of America to assess the proportion of infected individuals with deaths in India. There is a minor tweak of the American IFR for the demography of India. Yet, this may not be sufficient. That can be understood through a simple extrapolation. MedRxiv has published preliminary scientific reports that put the combined infection fatality rate of 0.25 per cent in India for the first and second wave. If we take that as the base and combine this with the latest nationwide serological survey conducted by the Indian Council of Medical Research (ICMR), which estimated that 67.6 per cent in the age group of six and above have antibodies, the excess mortality count drops to 16 lakh.
Next, the paper estimates excess mortality using the CMIE consumer pyramid household survey (CPHS). Although the focus of the CPHS has been employment, income and consumption, a collateral benefit has been its data on mortality. Using the survey yields an excess death count of 49 lakh!
Estimating excess mortality using a survey is contingent on the availability of information on the reasons why a respondent dropped out. Attrition is a general problem for any study that draws on a panel survey. Marriage, cohabitation, separation, divorce, or childbirth may lead to a residential move, and agencies conducting the survey are often unable to keep track of people as they move. Some panel studies (like the PSID and the US Health and Retirement Survey) can verify that a respondent dropped out due to death by linking his or her information to official death registers. In other countries, it is never possible to confirm deaths in this manner as the paper has done.
The world over, estimating Covid deaths has been a challenge. Confounding the confusion further, in the initial months, there was suspicion that governments were overstating deaths attributed to Covid-19. The Institute for Health Metrics and Evaluation (IHME) in Washington State reports that globally the underestimation of Covid-19 deaths is of the order of 2.1 times (Estimation of excess mortality due to COVID-19, May 13, 2021). That is, the ratio of excess Covid-19 deaths (they estimate this from excess all-cause mortality) to reported Covid deaths is 2.1. In developed countries, it is 1.5. In India, they find the ratio to be 2.97. If we multiply the reported deaths in India of around 4 lakh by 2.97, we get close to 12 lakh, an excess of 8 lakh, not too far from the 6.6 lakh deaths not reported in CRS for 2019.
In sum, we believe that there is no case for haste in estimating Covid deaths, except to improve the health infrastructure and prepare the country for future eventualities. On that score, India has made considerable progress. Between March 2020 and June 2021, Covid dedicated hospitals rose to 4,389 from 163 (27x) and dedicated Covid health and care centres are up 8,340 and 10,015 respectively from zero. Oxygen supported beds are 4.17 lakh, up from 50,583 and total isolation beds (excluding ICU beds) are 18.1 lakh, up from a mere 41,000 before.
M/s Anand, Sandefur and Subramanian state that engaging with data-based estimates is necessary. We agree. Better data should become available in 2023, if not sooner. It is sensible to wait for them than to conclude in haste. There is time. In the interim, we suggest that the authors shed light on how the country of origin of the virus has escaped waves of infections and deaths that continue to stalk nations around the globe. That would be consistent with the authors’ pursuit of accountability for now and in the future.
This column first appeared in the print edition on July 27, 2021 under the title ‘It doesn’t add up’. Ghosh is group economic advisor, State Bank of India, and Nageswaran is member, Economic Advisory Council to the Prime Minister of India. Views are personal