Page 36 - ES 2020-21_Volume-1-2 [28-01-21]

Saving Lives and Livelihoods Amidst a Once-in-a-Century Crisis 19

Box 4: Assessing the Management of COVID-19 across Countries and
within States in India

To assess the effectiveness of the policy response to COVID-19, we have to estimate the
counter-factual, i.e., what would have been the natural caseload and associated fatalities
purely based on the population, population density, the demographics of the population and
the number of tests conducted. Using a regression model, we estimate the effect on per capita
cases of each of these explanatory variables. Our sample includes the top 30 countries in terms
of total confirmed cases, which represent 86 per cent of the world caseload, from March to
December 2020.

We estimate the following panel regression model:
Log (No of total cases per lakh ) = α + ß * Log (population ) + ß * Log(population
c
2
ct
1
1
density ) + ß * log(Total tests per lakh ) + ß * Log (% of population above 60 years ) +
ct
4
c
3
c
ß * Log (% of population between 0-14 years )+ ß * Log (% of population between 15-59
5
6
c
years ) + ε ,
c
ct
where c denotes country and t denotes month. Note that the inclusion of the log of proportion
of the population above 60 years, 0-14 years and that between 15-59 years does not generate
a problem of multi-collinearity as the log transformation ensures that these variables are not
linearly dependent. In other words, log x , log x and log (1-x -x ) are not linearly dependent.
1
2
2
1
The following panel regression model has been estimated to estimate deaths using the same
group of countries:
Log(new deaths per lakh )= α + ß * Log (No of total cases per lakh ) + ß * Log (% of
1
2
t
ct
ct
population above 60 ) + ß *Log(No of beds per 1000 pop ) + ε ct
c
c
3
where c denotes country and t denotes time period.
As COVID-19 has been more lethal on aged population, taking into account per cent of
population above 60 years helps us to control for demographic heterogeneity across countries.
The number of beds per thousand has been taken as a proxy for health facilities that affects
the number of deaths.
Similar models were estimated across 30 Indian States and Union Territories as well. In this
model, c denotes States instead of countries.
1.31 Although all age groups are at risk of contracting COVID-19, older people face
significant risk of developing severe illness if they contract the disease due to physiological
changes that come with ageing and potential underlying health conditions. Though India has
a young population with only around 10 per cent share of people above 60 years of age, the
population of people above 60 years of age is significantly higher in India than in any of the
30 countries that account for 86 per cent of the cases (Figure 13). If we take the total cases
in India as estimated by the analysis above and apply the CFRs of countries with comparable
proportion of old age people and CFRs of some worse affected countries, it is evident that
India has been able to save a large number of lives (Figure 14).