We must use a data-driven approach to decide what and where to ease lockdowns
In a remarkable feat, India has been rather successfully observing an almost 100 percent lockdown over more than two weeks now, and there is an increasing speculation that the current lockdown may get extended further.
While such a lockdown - and extensions of it - are imperative to arrest the spread of the virus, given the heterogeneous nature of the Indian economy, the economic costs of a blanket lockdown can pile up sooner than we think and get more acute than we can imagine. And this begets the question: Is 100 percent lockdown too excessive for India?
To address this, the key insight I draw upon is the following: At the heart of both the mechanism of virus transmission and economic activity is the human-to-human contact. Now, every economic activity is different in terms of the frequency with which one has to interact with other economic agents: A barber's occupation involves more contact relative to say, a mathematician.
Could such differences in contact intensity across occupations be usefully exploited for a lockdown policy? Yes, because, intuitively, an occupation which involves lower contact and contributes more to output should not be subjected to the same degree of lockdown as that occupation which has higher contact intensity and contributes less to output. It is these combinations which I attempt to quantify -- only at a much broader level: the entire Indian economy.
To begin with, I define and create a measure of ‘Economic Contact Intensity’ (ECI Index) at the district and economic sector level (To the best of my knowledge, this is the first work which attempts to do this exercise for India; please refer to the research paper for details https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3572564.)
I find substantial variations of ECI across both districts and economic sectors, which highlights its important role as a useful instrument in rationalising the lockdown policy. I then identify districts (see Figure 1) and economic sectors (see Figure 2), with a mutual combination of the ECI Index and Output (Sales) share. (Please read the caption notes of the respective figures.)
Bottom line: Policy wise, pockets (whether districts or sectors) with low contact intensity environment and high output share could be preferred to have a lenient lockdown. Overall, incorporating the dimension of ECI in policymaking could help optimally navigate the dual challenge: keeping the transmission rate of the virus low while keeping the economy engine running, just about enough. The figures given below tell us which areas and sectors should be opened up first.
Figure 1: ECI dispersion across geography (district-wise)
Source: Author’s calculations
Please see Figure 1. Both panels are at the district level and the colour coding emphasise the degree of ECI. Panel A plots the raw ECI Index for the districts where the district colours lean towards an increasing red, imply a greater ECI. Other things remaining the same, a district with a higher ECI should perhaps stay in the lockdown for a higher period relative to a district with a lower figure.
Panel B plots the weighted ECI where weights are the respective aggregate district level output (sales) share. While Panel A plots the raw ECI, Panel B tries to juxtapose this raw ECI with the district-level output share. A greener district is an indication of lower contact-intensity and higher economic output share. Therefore, other things remaining the same, the lockdown could be made relatively lenient for the areas shaded green.
Figure 2: ECI across economic sectorsSource: Author’s calculations
Figure 2 plots the sector-wise dispersion of ECI and opens up another dimension to conducting the lockdown policy. The size of the bubble stands for the asset size: the bigger the bubble, the greater the sector size. The colour of the bubble stands for a geometric average of the sales (output) and the inverse of ECI, averaged at the sector level and thereby, highlighting the economic size impact of the sectors. The lighter colour bubbles (sectors) have, in general, a combination of lower ECI and higher output share and the darker colour bubbles (sectors) have a combination of higher ECI and in general lower output. The lockdown, therefore, could potentially be made relatively lenient for the lighter coloured sectors.
In short, the data and methodology given above can be used by policymakers to plan how to open up the economy gradually after the lockdown.(Vipul Mathur is Faculty, Economics, IIM Calcutta. Views are personal.)