India has launched what by any standard is the most draconian and complete nation-wide lockdown of any country affected by novel Coronavirus or Covid-19. This is unusual for several reasons. First, reported infections in India as of March 27, 2020 are 747 out of a total population of 1.3 billion. Of these, 20 have died, and 66 have recovered. What is more, of all the active cases, each and every one of them has shown only mild symptoms. By any metric, India has a far less serious COVID-19 crisis than many advanced and emerging countries.
It has been suggested, including by me, that the low reported infection rate might reflect a very low testing rate in India. So far, India has tested a little more than 27,000 samples or a testing rate of about 20 per million. Compare this to testing rates of more than 6000 in South Korea and more than 600 in Spain.
It is theoretically possible therefore that the infection rate in India is much higher; but if this is true, then it raises another puzzle, why isn’t an already ill prepared public health system overwhelmed with people showing up with COVID-19 like symptoms? Why aren’t thousands of people on ventilators? Given India’s large elderly population and many a kind of respiratory problems due to pollution and other reasons, why aren’t thousands of people on ventilators and in overcrowded hospitals? If any of this has happened, it’s a well-guarded secret. This might just be possible in a totalitarian state like China but it’s impossible in open and democratic India. Despite its well-known flaws, Indian media, to say nothing of foreign media in India, would have jumped on such a story of alleged cover up.
The other noteworthy feature of the Indian situation is a very low fatality rate when computed as a percentage of all reported cases, as low as 2.6 percent. Contrast this with fatality rates as staggeringly high as 10 percent in Italy and in the US of 1.5 percent, just a little lower than India. Apart from outliers such as Italy, most countries have fatality rates around 1-2 percent give or take.
On the face of it, the Indian data suggest no reason for panic, especially given India’s early and aggressive action in closing international borders to travellers from affected regions — much before such action was undertaken in places like the US and Europe. The panic in India and elsewhere emanates from statistical and epidemiological models which predict massive fatality rates if draconian action is not taken to curb the spread of COVID-19. It turns out that these models are very likely wrong and have overstated the probable number of deaths from the virus.
Especially egregious is a recent New York Times op-ed by Ramanan Laxminarayan, an economist and self-described epidemiologist who praised India’s lockdown as necessary to save millions of lives because in his apparently model based judgement 500 million Indians could be infected and the consequences would be dire in the absence of draconian measures. The author claims to have a model but the estimates he provides has a barebones verbal description of what the model purports to do with no presentation of either the mathematical model or statistical methodology and jumping directly to his alarmist conclusions. Yet, this dubious piece of research which cannot be replicated has dominated thinking about the spread of COVID-19 and ways to respond to it.
On March 26, the United Kingdom downgraded COVID-19 and said it is no longer a High Consequence Infectious Disease (HCID). In particular, low mortality rates were cited as a reason for the downgrade. Reacting to the UK announcement, Deborah Birx, the White House Response Coordinator for COVID-19, went so far as to say that the doomsday fatality scenarios predicted by widely publicized models are very likely wrong. Specifically, she argued that either infection rates, especially of those asymptomatic, are much higher than we thought, or the transmission mechanism is not correctly understood. Even more extraordinary, she stated bluntly: “The predictions of the models don’t match the reality on the ground either in China, South Korea or Italy.”
This is the crux of the matter. As argued by Stanford medicine professors Eran Bendavid and Jay Bhattacharya in an op-ed in the Wall Street Journal on March 24, a few days prior to the UK decision, there’s a serious methodological error in the way that fatality rates are being computed. Following conventional practice, as we did above, the fatality rate is computed as the number of deaths out of total confirmed cases, converted into a percentage. They correctly argue that this is an upward biased estimate as it commits the fallacy of selection. In other words, confirmed cases are those who’ve obviously shown symptoms and been admitted for treatment. However, the correct denominator is not confirmed cases but the actual total number of infections, a number which is obviously much larger.
In the event that there are many infected people who’re asymptomatic, the number infected will be much larger than the number of confirmed cases, and the correct mortality rate therefore will be much lower by several orders of magnitude. According to their estimates, the true mortality rate is not what we reported above for the US, 1.5 percent, but more like 0.01 percent. If we assume that their calculation would hold true for India as well, the actual mortality rate in India would be an almost unmeasurably small 0.017 percent.
The public policy implications are very clear.
A disease which has a high infection rate that produces mostly mild symptoms and a mortality rate lower than the common flu is not a non-issue, but it’s not a cause for the levels of panic and paranoia that we’re witnessing around the world, including in India. The obvious issue is that a disease with high infection rates requiring people with even mild symptoms to seek medical help will be a huge burden on the public health system. At some point, as is already happening in some places including the US, hospital beds and ventilators may have to be rationed. There is thus a strong case to ensure testing, screening and enforced self-isolation for those exhibiting COVID-19 like symptoms for the recommended 14 day quarantine period. However, there is no valid argument for a draconian total lockdown of the type that has been imposed in India. For an uncertain and relatively small gain in reduced infections, there is a huge economic, social, and human cost which has already begun to manifest itself.
Good public policy is always based on the best science and most recent and reliable evidence.
Unfortunately, it is not at all clear that science or evidence were the basis of India’s lockdown. The good news is that there is still time to recalibrate in light of new research, as the UK and the US are doing. At a minimum, the government must undertake a cost-benefit analysis to measure the likely gain of a lockdown versus the burgeoning economic costs of basically shutting the economy down, in a country with many poor people already reeling from an economic downturn.
The total lockdown perhaps makes sense in a world in which COVID-19 carries extremely high mortality rates. We now know that this is very likely incorrect. The Narendra Modi government needs to re-think its policy before the lockdown produces an economic and humanitarian disaster.
Rupa Subramanya is an economist and researcher based between India and Canada. Views are personal.
This article was first published by Observer Research Foundation (ORF).