By Pesach Benson and Omer Novoselsky • September 10, 2026
Jerusalem, 10 September, 2026 (TPS-IL) — The COVID-19 pandemic was not only a test of health systems. It was also a test of whether governments could be trusted to report what was happening — and a New Israeli study suggests that the strength of a country’s institutions May Have mattered more than its ability to count.
Hebrew University of Jerusalem researchers Ariel Karlinsky and Prof. Moses Shayo examined mortality data from 134 countries and territories and found that governments substantially underreported COVID-19 deaths during the first two years of the pandemic.
Published in the peer-reviewed Journal of Economic Growth, the study compares officially reported COVID-19 deaths with estimates based primarily on excess mortality — deaths above the number normally expected. The countries studied officially reported 5.08 million COVID-19 deaths in 2020 and 2021, while the researchers estimate actual COVID mortality at 12.47 million.
The researchers estimate that 45%-55% of governments in their sample misreported COVID mortality, with underreporting far more common than overreporting.
“The two most important findings are: (1) that the toll of the COVID pandemic was much higher than officially reported and (2) we show massive information manipulation by many governments around the world,” Karlinsky, a PhD candidate at Hebrew University, told The Press Service of Israel.
But the study goes beyond establishing that official figures were often wrong. Karlinsky and Shayo examined whether differences in countries’ ability to collect, register and report deaths could explain the gaps.
They found that statistical capacity accounted for only a small share of the variation. Countries with stronger institutional checks and balances generally showed less underreporting, as did countries where citizens were better positioned to evaluate official information, including through higher education and internet access.
The authors also found higher levels of misreporting in countries with a communist legacy and in countries holding elections during the pandemic. The study stressed that these are descriptive associations and do not establish that those factors caused governments to misreport deaths.
Karlinsky pointed to Chile and Russia as a concrete comparison. Both countries scored highly on measures of statistical capacity, he said. Yet Chile, which had stronger institutional constraints, reported COVID mortality relatively accurately, while Russia significantly undercounted it.
Reporting the Reality
The distinction matters because a reporting gap does not automatically prove deliberate deception. Limited testing, incomplete death registration and other weaknesses can cause governments to miss deaths. The researchers attempted to identify misreporting only where official figures fell substantially below the range of plausible excess-mortality estimates.
Testing itself, however, can become part of the problem, Karlinsky said. He cited countries such as Belarus and Egypt, where relatively few COVID tests were conducted despite reports from local hospitals of high infection and death rates.
The largest absolute discrepancy identified by the study was in India, where the government reported roughly 500,000 COVID deaths while the researchers estimate approximately 4.5 million. In Nicaragua, fewer than 250 COVID deaths were officially reported over two years, while excess mortality indicated deaths in the thousands.
The researchers acknowledge that the true death toll may never be known in some countries, particularly poorer nations where death registration is incomplete. Researchers have attempted to fill those gaps using sources including hospital data, obituaries and other indirect measures.
The implications extend beyond COVID. Governments routinely publish statistics that shape decisions about public health, economic policy, foreign aid and other international issues.
“Just like people can lie, so do governments,” Karlinsky told TPS-IL. “I think journalists should either refrain from publishing information from suspected countries or always accompany it by caveating that the information may not or cannot be verified.”
