The national lock down must continue till we have reliable data on the incidence of Covid-19.
Since the onset of Covid-19, the world has generally fumbled and faltered. And though there are some much-admired examples like China and some other South East Asian countries, the jury is still out on this issue. We also do not know what will happen during the second surge in East Asia nor is the European record enviable, except perhaps that of the Germans.
It looks as if the world has been collectively taken aback; there is almost a mass unnerving of sorts that has limited our ability to act rationally. We have laid off our wits, that ancient ally of sapiens in times of adversity.
This mass unnerving largely comes from the absence of information: not enough is yet known about the virus, for example, how deadly it is, its various modes of transmission and its pathogenic pathways. Moreover, the exponential speed of its spread has pressurised and unnerved most public leaders — politicians, bureaucrats, and public health professionals alike.
The key is to understand how to deal with this lack of information.
A lockdown can be completely effective only in a single-stage game. However, Covid-19 is a multi-stage game: there will be multiple peaks and troughs. A strict lockdown will show a downward trend in infection rates. People will relax, infections will rise, and the cycle will start again. The time lag between actions and payoffs further disincentivises people in terms of their choices for appropriate actions. Perceived severity is directly related to appropriate behaviour modification and behaviour modification affects the perceived severity of an epidemic. The time lag between behaviour modification and resultant changes in perceived severity produces peaks and troughs.
In game-theoretic terms, various forms of social distancing, including lockdowns, constitute a cooperation game whose positive network effects only materialise if there is no significant deviation in its early phase; otherwise there are massive negative network externalities. Even if a couple of dozen people from the neighbourhood continue to meet for Friday prayers or for a weekend cricket match, the purpose of social distancing is defeated, and the spread of the virus continues unabated.
A nationwide lockdown of the entire country assumes that the nature and level of disease is uniform across the country. This is an incorrect assumption: the disease spreads in clusters and mass contamination happens only when clusters spill over. While the spread of the disease is exponential, it is uneven too. Therefore, a complete lockdown might be effective for a limited time to control the deepening of the disease (such as a fortnight) but it cannot stay forever.
Managing Covid-19 without vaccine or medicine requires social distancing and a lockdown is an excellent strategy for social distancing in the short run. Unfortunately, lockdown capacity is a finite source. Every country has a maximum reserve of resources, in terms of individual or household savings, and the social credit and stamina to survive without working. This stamina, when aggregated across households sets the limits for national stamina for lockdown: let us call it the national lockdown reserve (NLR). Broad-spectrum, long term deployment of NLR exhausts the society, depletes the resources, drains administrative capacities and compels significant sections of society to deviate. Moreover, using our limited national stamina for lockdown only makes sense when the payoff in terms of averted spread of the disease is maximised. Given that it is a multi-stage game, it makes little sense to exhaust a finite resource prematurely, especially when that resource gives us our only fighting chance.
Given the uneven spread of the disease in spatial and temporal terms, only a differentiated approach can work. Different levels of lockdowns will have to be defined, with varying severities. On the softer end of the spectrum of a lockdown, soft measures (personal hygiene, ensuring six feet distance from each other, closing of places of daily communal interactions such as schools and colleges) might be good enough. On the stricter end of the spectrum, a kind of curfew may be required.
But a differentiated approach towards lockdowns can only be applied if different areas can be categorised on the basis of disease incidence.
In addition to differentiated lockdowns, a signaling system is needed. This signaling system will allow us to identify pockets where disease incidence or spread is stronger or weaker. The immediate need of the hour, then, is to create micro viral maps of the entire country, with distinct boundaries. Without such a mapping, a decision support system cannot be enabled to maximise payoffs vis-à-vis our finite stamina for lockdowns.
The assumption of uneven transmission of Covid-19 is at the heart of this argument. Given the transmission dynamics of Covid-19, an even spread of disease can only occur if two conditions are met: (a) space is homogeneous and (b) the motion of individuals is random and isotropic. Real world environments are heterogeneous and heterogeneity affects human motion as well as human behaviour.
If we look at the spatial distribution of Covid-19 across countries, the key assumption of uneven transmission has so far been confirmed. Even within cities, Covid-19 spreads in clusters of socioeconomic interaction. And given an uneven spread of the disease in spatial and temporal terms, only a differentiated approach can work. Different levels of lockdowns will thus have to be defined. Decentralised and differentiated containment strategies become even more important when we factor in the finite resource constraint of neighbourhood lockdown reserve, a national aggregate whereof is the national lockdown reserve (NLR). When households are grouped in geographic zones, the severity of lockdown measures (total curfew vs. ban on public gatherings) can vary according to disease levels or level of infecting potential.
Social distancing as a strategy hinges on the cooperation of entire neighbourhoods; therefore, moving full zones or full neighbourhoods instead of households along the social distancing spectrum creates incentives for congruent households to exert social pressure on delinquent households for compliance. A robust testing regime that rewards or punishes the neighbourhoods on degree of cooperation wards off the moral hazard of delinquent behaviour.
Countries like Pakistan do not have the option of offering extraordinary payouts for a national lockdown. Consequently, we have no option but to allow congruent neighbourhoods to function normally with minimalist precautions like a ban on public gatherings and lockdown only those neighbourhoods that have dangerous levels of viral loads.
The success of zoning strategy hinges on strict enforcement and strict enforcement requires sensible zoning of neighbourhoods. The proposition is thus to divide the country into corona management zones and corona management circles.
The perimeter of the zone or circle should be a clearly demarcated boundary, one defined either by natural or man-made markers such as a road or a canal. Economies are not organised on the basis of some arbitrary number of households but there is nonetheless the fundamental concept of a market town and its hinterland. If the relationship of the hinterland with the primary towns is respected while demarcating the perimeters, spillover effects (negative externalities) of quarantining certain zones or aggregate of zones can be neatly minimised.
So how do we divide a country like Pakistan on these lines? Pakistan has roughly about 33,000,000 households. If each zone has 16,000 households, we will have 2,000 zones. Commercial or industrial areas can either be included as part of such zones or can be separated into independent zones with limited access.
Each of the zones can further be divided into 10 circles each, of roughly 1,600 houses. What is important will be geographic contiguity, easily identifiable natural boundaries and socio-economic cohesiveness, in determining a zone or a circle.
Testing remains extremely crucial for combating the spread of the disease. By and large, tests are being conducted on people who either report symptomatic or have come into contact with symptomatic people.
Approximately 30% of the population infected with Covid-19 remains asymptomatic and acts as extremely potent viral vectors. To make things worse still, the current testing regime has roughly a false negative percentage of 15%. This adds up to 40% of the unidentified infected persons, the deadliest weapon in Covid-19's arsenal.
With 40% unidentified, total lockdown remains the only option because then one must assume that every individual is infected. Without isolating this 30% asymptomatic mass and another 10% of false negatives, any relaxation of the lockdown would defeat the entire social distancing strategy. However, testing kits are scarce and any diversion to asymptomatic testing is currently viewed as wasteful.
To overcome testing limitations, Germany started testing pooled samples; 10 individual swabs/samples with different viral transportation mediums are collected, pooled into one, and tested. If the pooled sample result is negative, then all 10 people are cleared. If none are positive, you save the cost of nine additional tests.
To further make pooled testing an effective tool in drawing up data regarding disease incidence and spread, randomised sampling can serve as an effective strategy. In a detailed strategy paper, this author has calculated annual cost of a robust testing regime at PKR 35 billion for the entire country. Rapid testing and serological testing may reduce this cost significantly.
Robust testing of the kind described above will throw up micro viral maps, enabling the authorities to apply a decision support system to place zones and circles in appropriate level of social distancing severity.
The question remains what if the data shows an infection in majority of circles. In this situation, a complete lockdown of 14 days will be observed in the entire country as shown in the stylised map here. Prior to the curfew-lockdown, ample time (a week or so) will be given for the population to prepare for a 14-day lockdown. People may travel back to their homes, with sufficient stocks for a long vacation to be spent with family alone. This lockdown may take place every quarter. Such a lockdown will result in across-the-board suppression of disease, enabling us to have the luxury of targeted lockdowns for the remaining 10 months.
The case of opening the economy while more strictly implementing rules of quarantine, isolation and social distancing where needed does not need to be justified. It has its own merits in terms of lives and livelihoods saved. Nonetheless, there is a strong economic and fiscal case for liberal spending on testing and on stricter enforcement around disease zones.
The existing government approach is primarily defined by a complete lock-down period followed by opening of selected sectors of the economy with a possible full lockdown in the future (with uncertain timing). Let’s call the government’s existing approach 'Approach A' and the proposed approach in this piece 'Approach B'. Both Engro (supported by Mackenzie) and PIDE have estimated significant impact on economic and fiscal state of the country under Approach A.
Implementation of Approach B will lead to the opening of more than 90% of the areas with minimal lockdown. This will result in increased economic activity. Based on the assumption of 90%+ opening of areas and economy, it is expected that implementation of Approach B will lead to reduction in unemployment rate from 16% to 7%, saving 5.6 million people from being unemployed. Similarly, the reduction in GDP as a result of implementing Approach B will be 8% as compared to 18% (Engro estimate) under Approach A.
Under Approach A, estimates are that the fiscal deficit could increase by PKR 2.16 trillion. However, if Approach B is implemented, it is estimated that fiscal deficit could be reduced to PKR 0.984 trillion; a net fiscal advantage of PKR 1.176 trillion.
The national lock down must continue till we have reliable data on the incidence of Covid-19, collected through robust testing of households selected via representative sampling. Zoning can be done in three days and if the state really moves into action, the first round of testing can be conducted through pooled method within 10 days. Any other approach including incremental opening of the country may lead to disaster.
This article is based on a strategy paper written by the author. You can read the complete paper here.
Rashid Mahmood Langrial is civil servant and a practitioner of design thinking and implementation in the public sector.
The views expressed by this writer and commenters below do not necessarily reflect the views and policies of the Dawn Media Group.