Survey-Based Estimates of Excess Mortality in the Covid Pandemic

by Gayatri Dewan

We conducted the 12th Consumer Pyramids Research Seminar on July 15 2021 at 8.30 PM IST (11 AM ET/ 8 AM PT / 4 PM UK).

Arvind Subramanian from Brown University’s Watson Insitute for International and Public Affairs, Abhishek Anand from the Harvard Kennedy School of Government and Justin Sandefur from the Center for Global Development presented

Survey-Based Estimates of Excess Mortality in the Covid Pandemic

Their presentation was discussed by Murad Banaji from Middlesex University.

You can view a recording of the webinar here.

Event Summary

The researchers presented findings from their paper on all-cause excess mortality during the COVID-19 pandemic. The presentation lasted 45 minutes and was attended by an audience of over 130 individuals. Their study extends the ongoing examination of the massive surge in mortality during the pandemic.

The authors constructed estimates using the Civil Registration System (CRS), the Consumer Pyramids Household Survey (CPHS), and the Seroprevalence surveys. The figures obtained were 3.4-4.4 million, 4.7 million, and 6.1 million excess deaths respectively, across 2 waves. Compared to these, the official estimate was 0.4 million excess deaths.

Prof Subramanian discussed the method of arriving at the figures and the limitations of each estimate. While CRS data was only available for a few states and undercounted deaths, the reliability of seroprevalence surveys, and the validity of applying the international Infection Fatality Rates to India was not assured. For CPHS, the authors were unable to validate the data against other sources and found a mortality spike in 2019.

Murad Banaji, serving as the discussant, spoke about various concerns associated with surveys and registration systems, such as biases with sampling and weighting, choosing a baseline, and drops in registrations. He emphasized the importance of having an accurate age profile in sample data given the age dependence of mortality and discussed the applicability of the IFR to India.

Following a discussion on the comments between Prof. Subramanian and Prof. Murad, Mahesh Vyas, CEO of CMIE, spoke about the age profile of the CPHS sample. Prof Sandefur responded to Q/As and suggestions from the audience regarding areas for future work.


India lacks reliable nationwide estimates of the death toll from the Covid pandemic. We report excess mortality estimates from the Consumer Pyramid Household Survey (CPHS), a longitudinal panel of over 800,000 individuals across all states interviewed up to 20 times from September 2014 to April 2021. CPHS mortality estimates for the pre-Covid period exhibit several deficiencies: they often do not track mortality estimates from other sources in the aggregate, or at spatial, state, and demographic levels; and they show an inexplicable pandemic-like mortality spike in 2019 well before Covid. The reliability of the post-Covid estimates is therefore unclear. With all these caveats, and recognizing the dearth of reliable alternatives, all-India mortality data sources, we report all-cause excess death estimates for the first wave of the pandemic between May 2020 and April 2021.

About the CPHS Research Seminar Series

The CPHS Research Seminar Series features work based on the Consumer Pyramids Household Survey. It is a platform for researchers to receive critical and technical feedback from accomplished peers. It is also meant to engage with the larger research community who may gain from technical discussions. You can find all previously presented work on the Events Section of our website. You can also view a recording of every seminar so far at the CPHS Research Seminar Playlist on our YouTube channel.

The Consumer Pyramids Household Survey is India’s largest regular household survey and the world’s largest household panel survey. CPHS has collected data on over 232,000 households and 1.19 million individuals surveyed since 2014. The survey collects information on household demographics, individual identities, employment, health status, financial inclusion, individual and household incomes, consumption expenditures, ownership of assets and intentions to buy them, household amenities and consumer sentiments. Income and expenses data are a monthly time-series since January 2014.

Record-level data from all 22 Waves of CPHS is available through a subscription service. Data from the Jan-Apr 2021 Wave of CPHS was released to subscribers on May 1, 2021. Please visit our website to learn more about CPHS and how to gain access to the record-level data.

If you have written a paper using CPHS and would like to present in the Seminar Series, please write to Kaushik Krishnan at

Use of CMIE Product implies acceptance of the usage agreement & privacy policy   ♦   FAQs   ♦  Diagnosis