Ever since the Covid-19 pandemic had hit the world last year, foreign liberal media has been showing a strange attitude while reporting on India. When India’s infection numbers were lower compared to developed nations like the USA, Italy and the UK, the international left-liberal media had wondered how it can be possible. For them, the low death rate in India due to Covid-19 was a mystery.
One year has passed since then, India has faced the second wave of the pandemic, with more deadly variants of the virus causing increased infection and more deaths. But the western media is not satisfied, they don’t believe the official numbers, believing the actual infection and fatality numbers are much more.
On May 25, New York Times published an article claiming that India’s actual Covid-19 death numbers are exponentially more than the officially published numbers. They predicted that India’s death numbers range between 6 lakh to 42 lakh, against the official number of 3 lakh. The article by Lazaro Gamio and James Glanz also estimated that the total Covid-19 cases in the country should be 40 crore to 70 crore, while the official number is 2.7 crore. Thus, NYT is claiming that India’s Covid-19 positive and death numbers are highly under-reported.
But the entire analysis done in the article is based on ‘estimates’ based on randomly selected numbers, and it does not explain why and how those estimates were made.
The article talks about three likely scenarios, conservative, more likely, and worse scenario. Based on these scenarios, they have estimated 40 crore, 54 crore and 70 crore infections and 6 lakh, 16 lakh and 42 lakh death respectively for each scenario.
NYT uses arbitrarily chosen random numbers
Now, New York Times has not used any statistically relevant theory to arrive at these numbers. They just randomly assumed by what degree the actually infections are more, and calculated the fatality numbers from a similarly arbitrarily assumed fatality rate.
For the ‘conservative’ estimate, they assumed actual infections are 15 times more, and multiplying current number of 2.69 crore on May 24, they arrived at the estimated infections at 40401 million (40.42 crore). After that, they assumed a fatality rate of 0.15%, and decided that estimated deaths should be 6 lakh.
What was the basis for increasing infections by 15 times and assuming 0.15% death rate? That is not answered in the report.
Similarly, for ‘more likely’ and ‘worst’ scenario estimates, they increased infections by 20 and 26 times, and assumed fatality rates of 0.30% and 0.60%, again giving any explanations. They reason that fatality rates go up with higher infections as the healthcare system gets overwhelmed and may people don’t get treatment. While that is true, no reason is given for choosing the rates they have used.
For each subsequent scenario, they increased infections by higher magnitude, and then used a higher fatality rate to arrive at highly inflated death numbers.
The entire News York Times article is nothing but an exercise of ‘what if’ analysis using simple arithmetic, it does not have any statistical analysis needed to arrive the kind of conclusion they are claiming. It just multiplies the official infection numbers with random numbers, and then further multiplies them with randomly chosen fatality rates to calculate the estimated deaths.
As proof that the entire article is nothing but an arithmetic exercise of multiplying India’s official Covid-19 number with random numbers, the article also included the estimated numbers if the actual cases are 10 times more, and the fatality rate is 0.10%.
There is no accompanying text for this calculation, it does not say what category of scenario it is. If the conservative estimate is 15 times, then what kind of estimate is 10 times, the article does not explain. This is purely a random calculation.
Cites serosurveys but don’t use their numbers
The article says that they relied on the three serosurveys done in India last year. In serosurveys, a small number of people are tested for Covid-19 antibodies, and the from the percentage of people showing the antigen in their bodies, the total infection in the entire population is estimated. The survey is based on the fact that people development antibodies for a pathogen after it is exposed to that pathogen, therefore, people who carried the Covid-19 antibodies was infected with the virus, and their bodies fought the virus without showing any symptoms.
Although the scientific basis for the serosurvey is sound, it suffers from many issues just like any other sample survey, like issues with sample selection, geographical spread of the samples etc. This means, the NYT attempted to debunk official Covid-19 numbers of India using data estimated from surveys.
Moreover, NYT cited three serosurveys, while two of them estimated that actual numbers are 26 and 28 times of actual numbers, one estimated it was 13 times. This is a very wide fluctuation in the numbers, raising doubt whether they can be used for any meaningful estimation.
Even though they have cited the serosurveys, they didn’t use the exact numbers to arrive at the estimates. They only used the 26 times figure for their ‘worst’ analysis as mentioned above. For the rest two, the authors just randomly allotted numbers of 15 and 20 to project the estimated infections without giving any basis for using those numbers.
No basis for death rate estimates
The NYT article has a section titled ‘How we estimated death rates’. While it raised the hope that there is an explanation behind the assumed death rates of 0.15%, 0.30% and 0.60% they have used, actually there was no explanation. They used a misleading title for the section, as it does not say how exactly they estimated the death rates. The section talks about how it is difficult to estimate the death rate as it depends on various factors like the age of the population. It does not contain a single sentence to inform how they estimated the death rates.
Quotes people proven wrong earlier
The article quotes several persons in support of its claim, and one of them is Dr. Ramanan Laxminarayan, director of the Center for Disease Dynamics, Economics & Policy. He is the same person who had predicted very high infection numbers for India last year, and after his predictions proved to be inaccurate, he is now peddling the theory that India must be underreporting cases and deaths.
Last year in March, several Indian media houses had interviewed him, where he had claimed that India will face 30 to 50 crore 300 million to 500 million cases by July 2020. By the end of July 2020, India had less than 16 lakh positive cases.
After the claims by is Dr. Ramanan Laxminarayan, an Indian, proved to be false, the same claims about Covid-19 in India are now being made by foreign authors in a foreign publication. Somehow, it looks like a part of the ongoing efforts by the left liberals to target the Modi govt using international left-wing publications. Suddenly media houses like The New York Times, Washington Post, Time magazine etc are publishing lots of anti-Modi govt articles written by Indian left liberal columnist, and the NYT article on the estimated Covid-19 deaths also looks like the part of the same agenda.
Central govt not responsible for reported deaths
India’s top opposition leader Rahul Gandhi shared the NYT article to attack the Modi govt at the centre. However, what Rahul Gandhi forgets is that the central govt is not responsible for any underreporting of Covid-19 deaths, if any.
Healthcare is a state subject, which means the entire Covid-19 management, along with recording the infection and death numbers, are done by the state govts. And BJP is not in power in all the states, several of them are ruled by non-NDA parties. If Rahul Gandhi thinks that India’s Covid-19 numbers are not real, he is also saying that Congress-ruled states like Punjab, Rajasthan, Maharashtra, Chhattisgarh etc are also not reporting correct Covid-19 infection and death numbers.
Therefore, if Rahul Gandhi believes the NYT numbers, he should tweet, ‘Numbers don’t lie, state governments in India do’.
‘Shaky’ record keeping system?
Showing the racist attitude of the white authors, the article also claims that ‘the system for keeping vital records in India is shaky at best’, while arguing that India’s Covid-19 numbers are not correct. Like always, they forget that India is a major IT powerhouse, and the country had developed a robust IT infrastructure to track the pandemic at its earlier stages. According to ICMR protocols, all testing laboratories are required to report Covid test results at the ICMR database. The agency has even taken steps to record the results of home testing after the self-testing kit ‘Coviself’ was approved. Users taking the test using the kit are advised to submit the result at the ICMR Covid1-19 Testing portal.
Therefore, the claim that India’s system of keeping Covid-19 test records is shaky is also completely false.
Similarly, India also has an established system of recording births and deaths, and it is not possible to underreport deaths to a large extent. While some Covid-19 deaths may be recorded as normal deaths, the variation can’t be as large as the NYT report is claiming. Today, families need to register the death of their family members due to various reasons like inheritance, financial, insurance etc. Therefore, most deaths are recorded, even if they don’t take place in hospitals.