On April 15, as the country’s first lockdown ended and the second began, India’s health ministry published its first district categorisation list, and identified 170 as ‘hotspots’. Then, on April 30, four days before India went into its third lockdown, the ministry issued a new district-classification list, which differed in two important ways: (i) the categorisation criteria had changed, and were no longer explicitly stated for every category; and (ii) individual states had been given some power to reset its districts’ categories.
The health ministry’s April 30 letter to the chief secretaries of the states and the union territories stated that 736 districts will be classified as ‘red’, ‘orange’ or ‘green’ zones according to some rules. A district will be “considered under green zone, if there are no confirmed cases so far or there [has been] no reported case [for] the last 21 days in the district”.
However, the letter didn’t spell out when a district would be considered a ‘red zone’. Instead, it states that the “classification is multifactorial and takes into consideration incidence of cases, doubling rate, extent of testing and surveillance feedback”. The orange zone ultimately includes all districts that are neither red nor the green zones.
The letter also gives the states limited powers to “reclassify districts”. Specifically, a state can intensify the lockdown status of a district – from green to orange or from orange to red – if the local situation demands it, but cannot relax the lockdown status – from red to orange or from orange to green. In addition, the states have the right to declare certain “high disease-load areas” within a red or orange district as ‘containment zones’, and there additional restrictions will come into force.
Over the last few weeks, a debate has raged about the extent to which government decision-making regarding the coronavirus epidemic should be decentralised. Some have argued that the Centre needs to devolve more authority to the states. The new district categorisation exercise seems like a step in this direction. But others have called this and other recent proclamations by the Centre an attempt to avoid making tough decisions.
One way or another, India’s state governments now face a dilemma. Given that most of India’s citizens are desperate for the lockdown to be relaxed so they can get back to work, any state government will indeed be hard-pressed to modify the Centre’s classification of a district to restrict economic activities further.
Consequences of the new categorisation
In the Centre’s first attempt, some districts had been categorised in a puzzling manner. The Centre’s second attempt, in similar vein, includes some surprising choices.
Recently, there was a sudden, unexpected surge of COVID-19 case numbers in some parts of India. These spikes could be the result of a large number of pilgrims, students and labourers moving from one region to another, as well as of increased testing. Such surges have the potential to quickly invalidate a district’s assigned category. This in turn makes the case for the Centre to be more flexible with categorising districts.
A classification scheme does have important implications for a district’s long-term ability to fight the epidemic. It’s possible that if a stressed district is identified early enough and is given adequate support, it will over time turn into a less-vulnerable district, and the opposite might happen if such a district is denied timely support.
For our analysis, we restrict our attention to district-level infection data (from covid19india.org) to nine states: Andhra Pradesh, Bihar, Delhi, Gujarat, Kerala, Maharashtra, Punjab, Rajasthan and Tamil Nadu. (This is only due to our limited time and energy, and, for some states, limited data). This said, these nine states had 36,000 people with COVID-19 on May 4 – 78.2% of all of India’s cases.
Quickly increasing infection numbers
In late March and early April, 3,500 pilgrims from the Sikh shrine in Nanded, Maharashtra, returned to different districts in Punjab. In the same period, 7,500 workers and traders went back to their home districts in Tamil Nadu from the Koyambedu market in Chennai. A large number of these people subsequently tested positive for COVID-19. Later, from April 30 to May 4, the case load in the corresponding districts of Punjab and Tamil Nadu surged: by 204 in Amritsar; by 77 in Hoshiarpur; by 63 in Bhagat Singh Nagar; by 59 in Sangrur; by 135 in Cuddalore; and by 85 in Villupuram.
If these six districts had experienced their respective surges before April 30, they would likely have been included in the red zone. But today, they remain in the orange zone. The only change we could identify (from news reports) was that the Tamil Nadu government has set up some ‘containment zones’ in Cuddalore.
Of course, no classification exercise can prepare for such unanticipated shocks. The health ministry letter clearly states that the announced list “is a dynamic list” and that it “will be revised on a weekly basis or earlier and communicated to states for further follow-up action.”
We hope that this is done as quickly as possible in response to spikes in individual districts. Similar rapid responses might also be called for in districts like Hingoli in Maharashtra, where the case load shot up after many soldiers garrisoned there tested positive in a short span of time.
There is also a distinct scenario in which small increases in the case load have the potential to confound the Centre’s district classification scheme. In explicitly stating the criteria for green zones – if and only if there have been no cases in the last 21 days – the Centre has run a risk. A district pronounced ‘green’ on April 30 may fail to remain so on May 4, which is when the third lockdown began.
More surprisingly, many districts – such as Araria, Katihar, Sitamarhi and West Champaran in Bihar; Baran and Pratapgarh in Rajasthan; and Fatehgarh Sahib and Rupnagar in Punjab – have reported new cases over the last week of April, but still sit in the green zone. Any attempt by state governments to alter the status of such districts to orange, especially when there have been a few new infections, will now require significant political will (notwithstanding the desirability of such a change).
On the other hand, Beed in Maharashtra is orange on the Centre’s list, while we found no evidence of a new infection there after April 8. In this case, the state government is not empowered to elevate Beed’s status from orange to green.
Other puzzling choices
And then there are districts that haven’t suffered any recent spikes but whose stated categories are just strange.
From April 4, the number of new cases in Raigarh has been consistently greater than the number of cases in Dhule. Yet Dhule is in the red zone while Raigarh is in the orange zone. The following district-pairs have the same anomaly: Rajkot > Aravalli in Gujarat from April 4; Nalanda > Gaya in Bihar from April 11; Coimbatore > Thiruvarur in Tamil Nadu from March 28; and Tonk > Banswara in Rajasthan from March 28.
Since the Centre has not specified the criteria for red zone districts, we can’t say that the above classification is incorrect. But they’re bewildering.
Appropriate categorisation at the right time
Of course, we should be so concerned about the correctness of the government’s classification exercise only if the classification helps. And while we don’t yet have sufficiently rich outcomes data, a preliminary answer is still available.
If a district between Yavatmal (Maharashtra) and Ranipet (Tamil Nadu) were to be tagged ‘red’, it should have been the latter, but the Centre tagged the former in its first classification exercise. Similarly, the Centre classified Pathanamthitta (Kerala), Thiruvananthapuram (Kerala) and East Godavari (Andhra Pradesh) ‘red’. However, Anand (Gujarat), Raigarh (Maharashtra) and Thanjavur (Tamil Nadu) should have been tagged ‘red’ but weren’t.
In the Centre’s exercise on April 30, it (correctly) classified Pathanamthitta (Kerala), Thiruvananthapuram (Kerala) and East Godavari (Andhra) as ‘orange’, and Anand (Gujarat) and Thanjavur (Tamil Nadu) as ‘red’. The Centre also tagged Raigarh (Maharashtra) as ‘orange’ – but incorrectly.
Thus, it’s tempting to claim that if the Centre had correctly classified the districts in its April 15 exercise, the inherently ‘red’ districts of Anand, Thanjavur and Raigarh could have turned ‘orange’ by now. Again, this is a preliminary finding, but it also suggests that appropriate district classification at the right time can lead to better outcomes.
Jyotsna Jalan is a professor of economics at the Centre for Studies in Social Sciences, Kolkata, and co-director of CTRPFP. Arijit Sen is a professor of economics at IIM Calcutta. The views here are the authors’ own.