Covid19 Lockdown FirmProfit Jain2023

17593441, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1759-3441.12377 by Indian Institute Of Technology, Wiley Online Library on [25/02/2023].
See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ECONOMIC PAPERS, VOL. , NO. , , 1–18
Effect of COVID-19 Lockdown on the

Profitability of Firms in India*
Ritika Jain1 and Rajnish Kumar2
We examine the effect of COVID-19-induced lockdown on the profitability

of listed firms in India. We use quarterly income statement of 4168 listed
firms for the period between April–June 2020 quarter and April–June 2022
quarter and compare their financial data with previous quarters (2015–
2019). Using a difference-in-difference estimation framework and various
profitability measures, we find that the COVID-19 lockdown has reduced
profits by around 15 per cent for listed firms in India. Our results are robust
to various robustness tests and alternate specifications. We find evidence of
firms losing revenues more than expenses, thus leading to decline in profits.
The main effect is conditioned by firm-specific factors. Specifically, firms that
are smaller, older, unlisted and that do not belong to any group witnessed
larger decline in profitability due to lockdown. Additionally, the effect of
lockdown is more pronounced in areas that had lower mobility and higher
COVID-19 spread. These results underscore the importance of institutional
factors and pre-existing firm characteristics in conditioning the impact of
lockdown on firm profitability.
Keywords: COVID-19, firm performance, lockdown, India.
1. Introduction
On 11 March 2020, after two months since the first case was reported in China, the World Health
Organization (WHO) declared COVID-19 as a global pandemic. To contain the spread of the virus,
various economies implemented measures ranging from stay-at-home orders to extended lockdowns.
This led to a collapse of economic activity reeling with an unprecedented situation of repressed
demand and severe supply shortages. With weak institutional factors, huge informal sector and high
This paper was awarded the second prize for the Best Paper award at Behaviour of Securities Market- Sightings
of the Black Swan organised by SEBI-NISM.
*We thank Kameswar Rao, Upasak Das, Sanchit Arora and Tirtha Chatterjee for useful comments and discus-
sions. We also thank Sudip Chaudhuri, M. Parameswaran, Vinoj Abraham, Beena P.L. and other seminar partici-
pants at the Centre for Development Studies, Trivandrum, Christ University, Bangalore, second SEBI-NISM
Conference, Mumbai, Centre for Technology, Innovation and Economics Research, Pune, SAMVAAD (Seminar in
Applied Microeconomics) by the Society for Economic Research in India, Indian Institute of Management, Kozhi-
kode, Madras School of Economics, Chennai and CTIER, Pune.
1
Centre for Development Studies Trivandrum, Kerala, India.
2
ISS ESG Mumbai, India.
JEL classifications: L25, G38, D22
Correspondence:Ritika Jain, Centre for Development Studies, P.O. Prasanth Nagar, Trivandrum, Kerala 695011,
India. Email: ritika@cds.edu
Accepted date: December 23, 2022
1
ÓÓ 2023 The Economic Society of Australia.
doi: 10.1111/1759-3441.12377
17593441, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1759-3441.12377 by Indian Institute Of Technology, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 ECONOMIC PAPERS
persistent inequality, the cost of lockdowns in developing economies is beyond the revival of
repressed demand and supply.
Moreover, some of these economies had already been exhibiting a slowdown in economic growth.
For instance, the Indian economy grew by 5 per cent in 2019 dipping to a decade low. Negative
growth rates in consumption in the rural economy, unemployment reaching a record high in the past
four decades and the industrial sector stagnating, the pre-COVID-19 India loomed all signs of a weak
economic situation, to begin with (Nagaraj, 2019; Kannan, 2020; Nagaraj, 2020).1 Consequently, a
lockdown may have acted as a brake to industrial output leading to a collapse of economic activity. It
is against this background that this paper examines how the COVID-19-induced lockdown impacted
firms’ profitability in India.
We build on the strand of literature that focuses on how COVID-19 lockdowns affect firm perfor-
mance. Most of the earliest papers have focused on the effect of COVID-19 and lockdowns on the
stock market performance of firms (Alfaro et al., 2020; Bekaert et al., 2020; Bretscher et al., 2020;
Davis et al., 2022; Ding et al., 2020; Zhang et al., 2021). The broad consensus in this strand of literature
is the slump in stock market returns due to COVID-19 and the role of spread of virus, government
restrictions and other policy actions in conditioning the impact. In the context of India, Mathur and
Sengupta (2020) also document a drop in returns for Indian stocks that were more vulnerable to
COVID-19. However, the stock market revived rather quickly and even exceeded the pre-COVID-19
level.
This emphasises the need for examining the effect of lockdown on operational performance indica-
tors. Few studies have documented the negative effect of COVID-19 and lockdowns on operational
indicators (Papanikolaou and Schmidt, 2020; Shen et al., 2020). Similar results have been found by
Hu and Zhang (2021), Brahmana et al. (2022) and Shen et al. (2020). Studies have also documented
the role of managerial experience (Qin et al., 2020), level of R&D investment (Biswas, 2021), country
and sector-specific factors (Golubeva, 2021), firm size and being categorised as essential category
(Stemmler, 2022) on how COVID-19 affects firm performance. In North India, Rathore and
Khanna (2020) use a primary survey of micro, small and medium enterprises to examine the effect of
lockdown find evidence of considerable financial stress.
The current paper aligns closely with the latter strand of literature since we use operational indica-
tors of firm performance. It uses data from the Prowess database that provides financial performance
of listed firms in India. We extract data for all 4168 firms that had disclosed their quarterly financial
statements between April–June 2020 and 2022 quarters and then compare their performance with
the past five years. Since the data pertain to the income statement of these firms, we limit our perfor-
mance indicators to a set of profitability measures and examine the impact of lockdown on it. To cap-
ture the lockdown impact, we discuss the impacts on the supply and demand side of the
manufacturing and services sector separately.
On the supply side of the manufacturing sector production drops drastically due to strict restrictions
on labour mobility and the opening of factories and offices. Distributors, transporters and dealers also
struggle to continue operations and start revising their expected sales accordingly. These events dis-
rupt the supply of manufacturing goods which may also lead to extreme actions such as shutdowns
and layoffs. In contrast, supply in service-oriented sectors will be driven by the extent of information
and communication technology (ICT) compatibility of the business and the proportion of employees
working from home (Papanikolaou and Schmidt, 2020).2 Another important factor while examining
supply disruptions is the list of industries, deemed as ‘essential’ services by the government and that
1
Among other factors, the decision of demonetising its higher denomination currencies (leading to a sharp drop
of 86% of currency in circulation) and replacing the indirect taxes with the Goods and Services Taxes (with ineffi-
cient implementation) contributed to the contraction of output.
2
For instance, information technology, consultancy and financial services are well equipped to continue busi-
ness if employees have the necessary infrastructure to work remotely. However, certain service sectors that
require personal contact with customers and other employees, such as hotels, tourism, restaurants and entertain-
ment will suffer significantly.
Ó 2023 The Economic Society of Australia.

COVID 19 LOCKDOWN AND FIRM PROFITS 3
were exempted from the lockdown as these were considered to be critical for the basic sustenance of
the economy.3 The Government of India (GoI) had also released an official order before the com-
mencement of lockdown that exempted a set of industries and services from it. Since this was an
order released by the central government, the list of essential services was uniform at the subnational
level.
On the demand side, Bekaert et al. (2020) describe that a massive lockdown of the economy
directly reflects a large demand slump. Commencing a sudden lockdown leads to a rise in income
uncertainty followed by drastic income cuts. Guerrieri et al. (2020) describe that with the lockdown
disrupting supply in some sectors more than others, the demand impact will be driven by comple-
mentarity between goods and services with sectors that face a strong supply shock. However, in the
presence of incomplete markets and weak institutional setting, these non-uniform supply shocks will
unravel an adverse demand shock across most sectors. To summarise, lockdowns lead to an inter-
twined shock to both revenue and expenses of firms. However, revenue shortfalls are rapid and
instantaneous with sudden lockdowns and cost reduction is more structural and delayed. Against,
this background we examine the impact of lockdown on firm profitability.
Further, the economic consequences of announcing a lockdown and its impact on firm perfor-
mance also needs to be contextualised using the role of institutions. Adverse economic shocks that
are limited to a particular region or industry can have substantial impacts on the forward and
backward-linked industries and regions (Inoue and Todo, 2019), consequently leading to an aggre-
gate shock (Gabaix, 2011; Acemoglu et al., 2016). In the current context, institutional quality influ-
ences how lockdowns affect firm performance (Ferraresi et al., 2020). Lockdown stringency will
depend on the degree of law enforcement by the police force. Consequently, we examine the role of
conditional factors on the impact of lockdown on profitability. We specifically examine the role of
weaker law enforcement as measured by lower mobility and the rate of infection spread on the effect
of lockdown.
Additionally, firms’ resilience and ability to recover from unexpected shocks will moderate the
effect of lockdown on firm profitability. We use age, size, group affiliation and stock market listing for
measuring firm resilience. Larger firms have better safety nets such as economies of scale, capacity
building and managerial capabilities that may enable them to withstand external shocks better than
smaller firms. Older firms have better experience and have existed through self-selection, ageing of
firms may lead to inflexibility and inertia in adapting to sudden changes in the external environment.
Khanna and Palepu (2000) argue that group affiliation may be beneficial in developing economies
since the scope of operations of group firms spreads beyond single industry groups and regions, they
are better placed in diversifying the external risk and mitigate market failures using these internal
institutions.
We use a difference-in-difference estimation framework to ensure causal inference of the lock-
down restriction on firm performance. We use industry and state-specific location identifiers for firms
to design our empirical strategy. We identify our treatment and control group of firms for the estima-
tion model using an exogenous approach of identifying industries that were exempted from the lock-
down. The GoI had released an official order before the commencement of lockdown in the country
that deemed certain industries as essential and hence were exempted from the lockdown. This is simi-
lar to the approach used by Shen et al. (2020).
To integrate the spatial variation in lockdown stringency at the subnational level, we use daily data
on workplace mobility, from the Google Community Reports and take account of mobility patterns at
the state level in the post-event period. We use interactions within the difference-in-difference model
to examine whether lockdown stringency conditions affects firm profitability. While using the mobil-
ity measures for the state where the firm operates, we account for multiple locations and compute a
3
There was significant variation across economies on the identification and implementation of ‘essential’ service
operations. For instance, in the US various state governments were responsible for deeming certain industries as
‘essential’ in their respective states. In contrast, Sweden had a softer approach by allowing work from home and
encouraging people to practice social distancing when stepping out (Fischer, 2020).

4 ECONOMIC PAPERS
measure of mobility at the firm level based on reported locations. We follow a similar methodology
for examining the effect of COVID-19 spread across various states. We find that lockdowns reduced
profitability of firms in India validating our first hypothesis. All measures of profitability- real profit
after tax, return on sales, real earnings per share and profit growth exhibited a decline of around 15
per cent. Our results are robust to different model specifications and alternate tests. As expected, we
find that while both expenses and revenue suffered due to the lockdown, the decline in revenues is
more than that of the expense, thus leading to profit loss.
We also find that lockdown has a higher effect on firm profitability if the external environment of
a firm has a lesser workplace mobility vis-à-vis a firm that operates in an environment with a relaxed
lockdown. Similarly, the effect of the lockdown is higher in firms that are located in areas that wit-
nessed a rapid rise in COVID-19 infection. Finally, we find that lockdown effects are more pro-
nounced for firms that are smaller, older and not part of any business group.
The current paper aligns well with the set of studies that have examined the impact of lockdown
on firm performance such as Bachas et al. (2020), Carletti et al. (2020), Liu et al. (2021), Hoehn-
Velasco et al. (2021), Yang and Yang (2021). We also highlight the non-uniform nature of the impact
that may be conditioned by various factors such as pace of infection spread, governance quality, the
efficacy of lockdown implementation and pre-existing firm-specific factors. This line of thought has
been explored in the past by Apedo-Amah et al. (2020), Golubeva (2021) and Zou et al. (2020).
2. The Indian Context

India recorded its first COVID-19 infection on 30 January 2020. By March 24, India had 618 con-
firmed cases of COVID-19. As a response, the GoI announced a twenty-one-day nationwide lock-
down from the following day as a steep measure to control the spread of the virus. According to the
COVID Government Stringency Index, India scored a full score of 100 in late March indicating a high
level of stringency. The timeline of events related to the outburst of COVID-19 in India and major
policy actions taken by the government has been compiled in Figure 1.
This lockdown was extended twice until May 31, when the Ministry of Human Affairs issued an
order announcing a phased reopening that lifted most prohibitory orders barring containment zones.
There are two striking and unique features of the Indian lockdown – the lockdown was a sudden sur-
prise event with a lot of future uncertainty about opening up and extension. Second, its commence-
ment coincided with the beginning of the June quarter lasting for almost two of the three-month
quarter. This provides us with an interesting opportunity to examine the effect of this nationwide
Pre- lockdown Central govt State-level

period Lockdown lockdown*
First Case of Lockdown Phased unlocking

COVID-19 in India announcement commences
Jan 27, 2020 Mar 25, 2020 Jun 1, 2020
Figure 1. Timeline of Important Events Related to COVID-19 in India. Notes: Lockdown in India was lifted in a phased
manner

event on the Indian industries. Lockdown in India was under the Central government only until 31
May 2020. With the lifting of the lockdown since June 1, except in the containment zones,4 the state
governments of India were given authority to devise lockdown restrictions and relaxations.
What followed was huge spatial and temporal variation in the type of policies being announced not
just at the state level but also within each state. Furthermore, there was considerable variation in the
type of lockdown and the duration of lockdown.5 Hence, a comprehensive way to capture lockdown
changes will involve tracking newspapers and other official documents to account for changes in con-
tainment zones, nature of lockdown and type of activities within states, across states and overtime.
While studying each of these restrictions in a detailed manner will be an excellent attempt but it is
likely that despite all our efforts, the data that we will collect will not be exhaustive since many of
these restrictions were not even documented.
To circumvent this problem, we employ two approaches – we consider the April to June 2020 quar-
ter, wherein two of the three-month quarter was under the Centre lockdown. Second, we use our
post-event period to start from the day the lockdown was announced. Since no district in India was
uninfected, we assume the post-lockdown window to be the period since April 1st, 2020. According
to the Google Community Reports, workplace mobility declined sharply in Maharashtra (which har-
bours two economic hubs Mumbai and Pune) and Delhi. In contrast, states such as Telangana (that
harbours Hyderabad, another commercial city), Andhra Pradesh and Karnataka (that harbours Ban-
galore) went lax. Similarly, the rise of COVID-19 cases was fast-paced in states like Maharashtra and
New Delhi whereas it was slow in Rajasthan and Kerala. The inter-state coefficient of variation of
cases between April and June 2020 has remained high and fairly stable in India.6 Unusual spikes in
case spread may also affect mobility across states, consequently inhibiting smooth business
operations.
3. Data and Variables
3.1. Data Source

The data have been collated from multiple sources. We extract all firm-level information from the
Prowess database owned by the Center for Monitoring India Economy. Spanning across all listed and
few unlisted firms, Prowess provides audited financial data, data on the location of firm offices, own-
ership groups and industry categories of firms. The study uses quarterly income statement of all firms
that have disclosed these reports between the quarters of April–June 2020 and April–June 2022 per-
iod. We track these firms for five years before the announcement of the lockdown making the final
time period of the study ranging from April–June 2015 quarter to April–June 2022 quarter. We also
use the GoI order dated 24 March 2020 to classify all the selected firms into essential (exempted
industries) and non-essential categories by mapping the industry information. To account for labour
mobility during the lockdown, we use Google Community Mobility (GCM) data set at the state level.
Based on GPS location, GCM records changes in mobility trends across various categories including
the workplace. Using the five weeks of 3 January–6 February 2020, the GCM dataset uses the median
value of mobility for each day of the week as a reference point. Based on this reference point, each
data point compares the percentage change in mobility on a particular date. Finally, we also use daily
data on the spread of COVID-19 in India at the state level from http://www.COVID19india.org, a
4
Containment zones were districts that were identified to have the highest share of COVID-19 cases in India.
5
For instance in the first week of June, states such as Bihar and Sikkim had complete lockdown, West Bengal
announced a two-day lockdown every week and Uttar Pradesh, Madhya Pradesh and Punjab opted for a weekend
lockdown. The central government order explicitly mentioned that states were not allowed to impose any lock-
down rules outside the containment zones. For instance, urban areas of certain districts in Chhattisgrah were in
lockdown whereas the rest of the area was not a containment zone. Similarly, Kerala had containment zones
according to ward numbers within a district. Furthermore, the identification of containment zones changed
according to the spread of COVID-19 infection in the local zones.
6
Based on authors’ calculations using COVID-19 data. Source of the data www.Covid19india.org

6 ECONOMIC PAPERS
crowdsourced initiative that curates and verifies data related to COVID-19 cases in India from state
bulletins and official handles.
3.2. Variables
To examine the effect of COVID-19 on firm performance, we use four measures of profitability – (i)
natural log of real profits after tax (PAT), excess of income over expenses, (ii) the log of return on
sales (ROS), ratio of profit after tax to operating income, (iii) the log of real earnings per share (EPS),
ratio of profit after tax to the total number of outstanding shares and (iv) the growth in profit after
tax (GP), quarter-on-quarter proportionate change in real profit after-tax. We convert the nominal
value of PAT and EPS into real value by using quarterly WPI from the Reserve Bank of India website.
Following the previous studies on factors affecting firm performance, we include a set of control vari-
ables such as age, leverage, growth of revenue and research and development expenses. The defini-
tion of these variables has been presented in Appendix S1.
Besides these financial variables, we also control for firm ownership (group, state-owned, foreign
and private), industry group and location using dummy variables. To account for the change in con-
centration at the industry level overtime, we compute the concentration ratio for the top four firms
at the three-digit NIC code level for every period. To map our firm-level data with state-level external
factors, we identify the states in which the firm has headquarters, registrar’s office, corporate office
and registered office.7 Thus, our final dataset provides information on firm-specific factors, industry-
level factors and state-specific factors. Descriptive statistics of the important variables along with their
definition is presented in Appendix S1 (Table S1).
4. Identification Strategy
According to the Securities and Exchange Board of India (SEBI) guidelines, all listed firms are
required to disclose their quarterly financial statement within forty-five days of the end of each quar-
ter.8 However, as a response to the pandemic, SEBI had relaxed the deadline for filing the financial
report for the quarter ending June 2020 to September 15. Our analysis covers all firms that have dis-
closed their income statement in any quarter since April–June 2020. We construct a panel data using
these 4168 firms and tracing them for the past few years.9 We also winsorize profitability measures at
1 per cent to avoid estimation bias due to the presence of outliers. The final data set is a panel of 4168
firms across all quarters between 2015 April–June quarter and 2022 July–September quarter.
Our main empirical strategy is a difference-in-difference (DD) regression framework to examine
the effects. We define the nationwide lockdown announcement of March 25 as the event. Since the
countrywide lockdown started from March 25 and continued till May 31, as presented in Figure 1, it
provides us a natural advantage of considering quarterly data. We define the quarter ending June
2020 as the post-event period and all the prior quarters as the potential pre-event period. However, it
may not be viable to include all pre-event quarters in the analysis. Differences in demand conditions,
business decisions and fund allocation across quarters within the same year may cause seasonality
across quarters in the same year. Hence, to provide a reasonable comparison of the pre-event period
with the post-event period (April–June 2020), we examine the same quarter (April–June) of all years
between 2015 and 2019 as the pre-event period.10
The GoI order mentioned that public utilities such as petroleum, CNG, LPG, PNG, power generation
and transmission units, post offices, electricity, water and sanitation corporations would not remain
closed. Additionally, hospitals and related medical establishments ranging from manufacturing,
7
All firms may not have four offices. Depending on the number of offices reported in Prowess database, we use
the relevant states for the particular firm.
8
According to clause 41 of the guidelines, a company may submit audited or unaudited quarterly financial
report. If the financial results are unaudited, it will be subjected to a limited review by the statutory auditors of the
issuer. Source: https://www.bseindia.com/downloads1/claause41.pdf
9
While Prowess database covers a large number of firms as far as annual report information is concerned, only a
handful of them report quarterly earnings statements.
10
As a comparison to the baseline model, we present the results for the pre-event period with all quarters.

distribution, transportation, shops, laboratories and hospital support services are to be permitted.
Even within commercial and private establishments, ration shops, all food and food products, dairy
products, meat, groceries and so on were to remain functional. Importantly, the order also specifies
that the manufacturing of all essential units and production units that require continuous process
was to remain open. This implies basic industrial sectors such as chemicals, coal and minerals would
not be disrupted. The non-essential category of firms is our treatment group, and the essential cate-
gory is our control unit for lockdown. For firms producing multiple products, we use the industry
group that generates the highest share of revenues as identified by the main product. We present the
spatial and sectoral distribution of the firms in our analysis in Figure S1A,B of the Appendix S2.
Goodman-Bacon and Marcus (2020) discuss the issues related to examine the impact of policies
related to COVID-19 using a difference in difference research design. Using two different cities located
in different countries, the difference-in-difference estimation design in Goodman-Bacon and Mar-
cus (2020) uses the country imposing the early lockdown as the treatment category and limits the
period of analysis to the date before the commencement of lockdown in the other country. The chal-
lenges in this approach are on grounds of different countries having customised approach to policy
actions with a package of policies being implemented by each country, reverse causality between
COVID-19 spread and lockdown policies, voluntary precautions and conscious behavioural changes
by individuals in both the countries and spillover between the treatment and control groups.
Most of these challenges arise due to categorising treatment and control groups by using two sepa-
rate countries. First, the approach to policy packages and COVID-19 spread is not uniform across the
two cities since they are in different countries weakening the possibility of parallel trends assumption
to be satisfied. Further spillover effects may also bias the estimates – since COVID-19 cases and poli-
cies in nearby regions will affect outcomes and behaviour of residents in the study group. Goodman-
Bacon and Marcus (2020) recommend prudence while choosing the treatment, control groups and
suggest using a setting where the policy is implemented at the national level.
In the current research design, we deviate from location-based categorisation of treatment and con-
trol groups, thereby reducing the possibility of the difference-in-difference estimates to be biased.
Furthermore, we use the government circular for identifying our essential and non-essential group-
since this circular was released by the Central government of India, it is exogenous. Furthermore, the
circular does not vary across states and hence is uniform across all the firms in India eliminating dif-
ferences in policies arising due to different regulatory governments. However, there are some issues
that need to be discussed. First of all, lockdown in India was under the Central government only until
31 May 2020. With the lifting of the lockdown since June 1, except in the containment zones,11 the
state governments of India were given authority to devise lockdown restrictions and relaxations. We
employ two approaches – we consider the April to June 2020 quarter, wherein two of the three-
month quarter was under the Centre lockdown. Second, we use our post-event period to start from
the day the lockdown was announced. This approach has been employed by other India-based studies
as well. Alsamhi et al. (2022) use descriptive tools for examining the effect of COVID-19 on few
selected firms in India and use the April–June quarter of 2020 as the post-lockdown period.
Soni (2021) examine the impact of lockdown on anthropogenic air pollution in various cities in India
by considering the April–June 2020 quarter again. Tripathi et al. (2021) analyse the impact of lock-
down on agricultural sector and rural economy using the same quarter as the post-lockdown period.
Mitra et al. (2022) also use the same quarter-based approach to examine the effect of lockdown on
the urban labour market in India.
Additionally, since no district in India was uninfected, in a second version of our model, we assume
the post-lockdown window to be the period since 1 April 2020 to the latest available time period. We
do take into account the lockdown stringency variable by capturing the mobility using the Google
Community Mobility database. Furthermore, the approach of dividing industries into essential and
non-essential group has been followed by Shen et al. (2020). We estimate the effect of lockdown in
the following DD regression framework
11
Containment zones were districts that were identified to have the highest share of COVID-19 cases in India.

8 ECONOMIC PAPERS
yit ¼ α0 þ α1 D2020,t þ α2 NEtreat,j þ α3 D2020,t NEtreat,j

þα4 x it þ α5 z jt þ α6 I j þ α7 Og þ α8 Ss þ α9 T t þ ϵit (1)
where yit denotes the profitability measure of firm i, operating in industry j, located in state s, owned
by entity g in quarter t.12 D2020,t is a dummy variable that takes a unit value for the year 2020 and zero
otherwise. α1 measures the difference in average profitability between the pre-event and post-event
period. NEtreat,j takes a value one if the industry is categorised as non-essential and zero otherwise. α2
measures the time-invariant difference between the treatment and control group. The effect of lock-
down is denoted by the product D2020,t NEtreat,j and we, therefore, focus on the coefficient α3 . If α3 is
negative, it implies that the lockdown hurts the profitability of the enterprises in India. x it and z jt are
vectors of confounding firm-specific and industry-specific characteristics respectively. Finally, we use
dummy variables at the industry, state, year and ownership level.
5. Results
5.1. Main Model

The DD model gives precise causal estimates only if the parallel trends assumption is satisfied. The
parallel trends assumption requires the treatment and control groups to follow a similar trend before
the announcement. This would ensure that differences across control and treatment groups are
Figure 2. Profitability Trend by the Treatment and the Control Group
12
In our analysis, t refers to the April–June quarter of each year and not the entire year.

time-invariant. Figure 2 presents profitability trends followed by our treatment and control groups
during the pre-treatment period for the April–June quarters of 2015–2019.13
The treatment and control groups follow a parallel trend before 2020 and then diverge (Figure 2).14
We test the parallel trends assumption formally by re-estimating Equation (1) for the period before
the lockdown announcement. Table 1 compiles the results of the parallel trend assumption. We
report only the coefficient of treatment with a year trend for the pre-event period.
We find that the interaction of our time trend with treatment for both the samples is insignificant,
thus validating the parallel trend assumption. After validating the parallel trends assumption, we
examine the impact on all four measures of profitability by estimation Equation (1). Table 2 compiles
the results.
Table 2 presents the impact of lockdown on profitability by considering four versions of the main
model – with and without controls and by considering the post-lockdown period to be 2020 and by
extending it to 2022. Lockdown has a negative and significant (at 1 per cent level) for all the models.
According to the last panel, PAT declined by 15 per cent due to lockdown. Similarly, ROS and EPS
plunged 3 and 7 per cent, respectively.15 PG rate plummeted 4 per cent due to the lockdown. The
other panels also indicate a similar pattern. The detailed regression result with firm-specific variables
and other controls have been discussed in Appendix S3, Table S2. As a robustness check, we re-
estimate this model for all quarters (Appendix S4; Table S3). The effect sizes are similar to those in
Table 2 affirming that the lockdown is a strong blow to the Indian corporate sector.
It is important to evaluate the effect size in detail. According to official statistics, India’s GDP faced
a worst-ever contraction, falling by 23.9 per cent in the June 2020 quarter. Our sample consists of
Table 1. Results for the Parallel Trends Assumption
(1) PAT (2) ROS (3) EPS (4) PG
April–June quarters (2015–2019)

Non-essential industries ×pre lockdown −0.003 −0.000 0.001 −0.006*
trend (0.007) (0.001) (0.003) (0.004)
Other controls Yes Yes Yes Yes
No. of observations 26,003 25,996 24,830 19,644
R2 0.11 0.32 0.14 0.05
All quarters (April–June 2015 to January–March 2020)
Non-essential industries ×pre-lockdown 0.002 (0.005) −0.001 0.002 −0.001 (0.002)
trend (0.001) (0.002)
R2 0.15 0.31 0.14 0.05
Notes: Estimates from regressions are presented with robust standard errors in parentheses. The dependent variables used in Mod-
els (1)–(4) are measures of profit after tax (PAT), return on sales (ROS), earnings per share (EPS) and profit growth (PG) respec-
tively. All regressions use clustered standard errors at the industry-state-year level. Non-essential industries is a dummy variable
that takes value 1 if the industry is non-essential and 0 otherwise. All regression models include firm-specific variables such as
age, leverage and size. Industry-level concentration ratios are also used. The models also control for industry, state, year and own-
ership fixed effects.
*, ** and *** indicate significance at 10%, 5% and 1% respectively.
13
The trend followed by the two categories for all quarters before the lockdown has been presented in Figure IIC
of the Online Annexure.
14
The red vertical line does not indicate the date of announcement of the lockdown. Instead it separates the pre-
event period from the post event period.
15
Since our dependent variable is in log terms, the coefficient is converted into effect size by using [1-
exponential (coefficient size)].

10 ECONOMIC PAPERS
Table 2. . Effect of Lockdown on Firm Profitability in India
(1) PAT (2) ROS (3) EPS (4) PG

Non-essential industries × post-lockdown −0.594*** −0.173*** −0.315*** −0.073***
(D2020,t NEtreat,j Þ (0.085) (0.036) (0.039) (0.023)
Other controls No No No No
Non-essential industries × post-lockdown −0.422*** −0.046*** −0.202*** −0.082***
(D2020,t NEtreat,j Þ (0.082) (0.018) (0.039) (0.026)
R2 0.13 0.34 0.14 0.06
Non-essential industries × post-lockdown −0.225*** −0.072*** −0.123*** −0.031**
(D > 2020,t NEtreat,j Þ (0.043) (0.018) (0.020) (0.014)
Other controls No No No No
Non-essential industries × post-lockdown −0.149*** −0.018** −0.066*** −0.029*
(D > 2020,t NEtreat,j Þ (0.043) (0.009) (0.020) (0.015)
R2 0.19 0.35 0.15 0.08
tively. All regressions use clustered standard errors at the industry-state-year level. Non-essential industries are a dummy variable
that takes value 1 if the industry is non-essential and 0 otherwise. Post-lockdown is a dummy variable that takes a unit value for
quarters after April–June 2020. All regression models include firm-specific variables such as age, leverage and size. Industry-level
concentration ratios are also used. The models also control for industry, state, year and ownership fixed effects.
firms that are listed on the stock market and that have managed to disclose their financial informa-
tion at the end of the quarter even under lockdown. It is reasonable to assume that these firms are
the least vulnerable group of firms and hence, are less affected as compared with firms that are facing
a complete collapse of production activity. Thus, a decline of 15 per cent in real profits is a reasonable
estimate against the 23.9 per cent decline in overall economic activity in the country for the same
quarter.
5.2. Robustness Tests
5.2.1. Falsification tests

To re-instate our causal effect of lockdown on profitability, we implement two sets of falsification
tests (similar to Aggarwal and Narayanan, 2021; Jain, 2021). We generate a new treatment indicator
taking values zero and one, as a random assignment to all firms and we re-estimate our models. If
our true models indeed capture the effect of the lockdown, the placebo models with this new treat-
ment indicator should not capture any effects. For the second falsification test, we assume that our
event (lockdown) was announced in 2019 instead of 2020. We present the 2015–2022 results in
Table 3.

Table 3. Robustness Tests for the Effect of Lockdown on Firm Profitability
(1) PAT (2) ROS (3) EPS (4) PG
Falsification test 1 – randomly generated treatment indicator

Random treatment ×post-lockdown 0.024 (0.037) −0.003 (0.009) −0.005 (0.015) 0.007 (0.015)
R2 0.12 0.38 0.15 0.08
Falsification test 2 – 2019 lockdown
Non-essential industries × 2019 lockdown 0.088 (0.063) 0.023 (0.016) 0.017 (0.025) 0.043* (0.028)
R2 0.12 0.30 0.13 0.10
tively. All regressions use clustered standard errors at the industry-state-year level. Non-essential industries is a dummy variable
that takes value 1 if the industry is non-essential and 0 otherwise. Post-lockdown is a dummy variable that takes a unit value for
quarters after April–June 2020. All regression models include firm-specific variables such as age, leverage and size. Industry-level
concentration ratios are also used. The models also control for industry, state, year and ownership fixed effects.
The results indicate that the DD variable is insignificant in all the specifications. Thus, our models
capture the causal impact of lockdown in India.
5.2.2. An alternate difference in difference estimation

Since our lockdown variable is based on an aggregate approach, it may be limiting in capturing the
state-wise ground reality. To counter this issue we use an alternate difference-in-difference technique
Table 4. Effect of lockdown using an alternate difference in difference estimation
(1) PAT (2) ROS (3) EPS (4) PG
Period of analysis: 2015–2020

2020 dummy ×April–June quarter −0.103** −0.088*** −0.108*** −0.047**
dummy (0.047) (0.023) (0.030) (0.019)
R2 0.14 0.48 0.15 0.09
Period of analysis: 2015–2022
Post 2020 dummy ×April–June quarter −0.063*** −0.041*** −0.029*** −0.027**
dummy (0.023) (0.010) (0.013) (0.010)
R2 0.13 0.48 0.16 0.07
Note: Estimates from regressions are presented with robust standard errors in parentheses. The dependent variables used in Models
(1)–(4) are measures of profit after tax (PAT), return on sales (ROS), earnings per share (EPS) and profit growth (PG) respectively.
All regressions use clustered standard errors at the industry-state-year level. We define treatment as all the years 2020 onwards.
April–June quarter in each year is considered as the post-event period and January–March quarter is the pre-event period. All
regression models include firm-specific variables such as age, leverage and size. Industry-level concentration ratios are also used.
The models also control for industry, state, year and ownership fixed effects.

12 ECONOMIC PAPERS
where in we use time–time space rather than state-time space (similar to Aggarwal and Naraya-
nan, 2021). We define the event of lockdown as the treatment, all years since 2020 as the treatment
unit and the preceding years, 2015–2019 as the control units. The three quarters before (April–June)
is the pre-treatment period and the April–June quarter each year as the post-treatment period. The
results have been compiled in Table 4.
We find that the coefficient of lockdown, as measured by this new DD variable, continues to be
negative and significant, strengthening our main findings.
5.2.3. Multiple hypothesis problem

Since we analyse a large set of profitability outcomes, we address concerns with multiple hypothesis
testing by using the Romano–Wolf correction (Anderson, 2008). The Romano–Wolf correction con-
trols for the family-wise error rate, which is the probability of rejecting at least one true null hypothe-
sis among a family of hypotheses under a test. We report the Romano–Wolf p value associated with
our key coefficient of interest for each outcomes in Table 5.
All the p-values continuously be significant indicating that our main findings are consistent with
multiple hypothesis testing issue.
5.3. Potential Channels

Total expenses of a firm encompass the supply side of the firm, that is costs associated with the pro-
duction inputs including labour, capital, raw material, energy and so on. On the other hand, total
revenue is the value of sales made by the firm which reflects the demand pattern of the firm. In line
with Section 1, we focus on the effect of the lockdown on both expenses and revenue. We use the log
of real expenses and the growth rate of real expenses as compared to the previous quarter to capture
the supply side. Similarly, we use the log of real operating income and the growth rate of real operat-
ing income to capture demand. Table 6 compiles the results.
Expenses of firms have plunged by a smaller amount due to the lockdown as compared to the rev-
enue of firms. Growth rates also exhibited a similar trend. With sharper declines on the revenue side,
firms struggle to continue operations and lose money since these slumps are instantaneous. In con-
trast, reducing business expenses is more structural in nature.
Table 5. Multiple hypothesis testing- impact of lockdown (DD estimate)
(1) PAT (2) ROS (3) EPS (4) PG

Non-essential industry × post-lockdown −0.422*** −0.046*** −0.202*** −0.082***
estimate (D2020,t NEtreat,j Þ (0.082) (0.018) (0.039) (0.026)
Model p-value 0.000 0.000 0.000 0.003
Romano-Wolf p-value 0.009 0.009 0.009 0.009
Non-essential industry × post-lockdown −0.225*** −0.072*** −0.123*** −0.031**
estimate (D > 2020,t NEtreat,j Þ (0.043) (0.018) (0.020) (0.014)
Model p-value 0.000 0.000 0.000 0.095
Romano-Wolf p-value 0.009 0.009 0.009 0.069
Notes: p-Value sensitivity to multiple hypothesis testing. For reference, we include point estimates from Table 2. Adjusted p-values
obtained following the procedure of Romano and Wolf (2016), as implemented in the RWOLF command in Stata (Clarke, 2020).

Table 6. Main Channels that Drive the Result
Category Cost Revenue
(1) Ln (real (2) Expense (3) Ln (real (4) Income

expenses) growth rate income) growth rate
2015–2020 April–June
Non-essential industries × post- −0.295*** −0.341*** (0.040) −0.380*** −0.401*** (0.045)
lockdown (0.034) (0.045)
R2 0.67 0.10 0.80 0.11
2015–2022 April–June
Non-essential industries × post- −0.138*** −0.198*** (0.021) −0.192*** −0.209*** (0.022)
lockdown (0.018) (0.022)
R2 0.80 0.11 0.84 0.10
els (1)–(4) are measures of expenses, growth of expenses, income and growth in income, respectively. All regressions use clustered
standard errors at the industry-state-year level. Non-essential industries are a dummy variable that takes value 1 if the industry is
non-essential and 0 otherwise. Post-lockdown is a dummy variable that takes a unit value for quarters after April–June 2020. All
regression models include firm-specific variables such as age, leverage and size. Industry-level concentration ratios are also used.
The models also control for industry, state, year and ownership fixed effects.
Table 7. Characteristics of Firms That are Dormant Since the Onset of the Pandemic
Firm-specific factors during the period of (1) Active status since (2) Inactive status since (3)
2015–2019 April–June 2020 April–June 2020 Difference
Number of firms 5307 1232 4075

Proportion of larger firms 0.609 0.430 0.178***
Proportion of older firms 0.389 0.474 −0.084***
Proportion of group firms 0.280 0.250 0.030**
Proportion of listed firms 0.821 0.273 0.547***
Profit after tax (in Rs. million) 208.8 −75.94 284.74***
Proportion of firms belonging to the 0.439 0.472 −0.032**
non-essential category
Note: The table presents the difference in means for the period of 2015–2019 by dividing firms into two categories – firms that were
active in the period after the lockdown (since April–June 2020) and firms that were inactive. Larger firms are firms that have
assets above the median assets in the industry group for 2015–2019. Older firms are firms that have an age above the median age.
Listed firms are firms that are listed on the stock market.
*, ** and *** denote statistical significance at 10%, 5% and 1% levels.
5.4. Some Extensions
5.4.1. Missing firms during the lockdown

It is imperative to examine the other side of Indian industry during the lockdown – firms that became
dormant during that period. We look closely into the number, type and characteristics of firms that
were active until April 2020 and became dormant sometime after that. We found 1232 such firms. To
better understand the characteristics of these firms, we group firms into two segments – firms active

14 ECONOMIC PAPERS
during 2015–2022 and those that were not. As a next step, we contrast the characteristics of these
two groups on the basis of their profile in the five years prior to lockdown (2015–2019). Results are
compiled in Table 7.
We find that the ‘dormant’ group had a lower proportion of larger, newer, grouped and listed firms.
Furthermore, the average profitability of firms in this group was significantly lower (to the extent of
making losses), than firms that continued to be active during COVID-19 lockdown. These results
strongly suggest that the relatively vulnerable firms ended up discontinuing operations during the
pandemic.
5.4.2. Firm resilience and the main effect

We also explore the effect of certain firm-level characteristics (size, age, indicators for group owner-
ship and being listed on the stock market) in driving our main result. We define small (old) as a
dummy that takes a unit value if the total asset in the last financial year (incorporation year) was
below (earlier than) the industry mean asset size (year of incorporation) and zero otherwise. Simi-
larly, we use no-group (not a NIFTY-500) dummy variable if the firm is part of a business group (part
of NIFTY-500) and zero otherwise. We discuss our results in Table 8. We find that the effect of lock-
down is more severe for firms that are more vulnerable – smaller, older, not part of a business group
Table 8. Effect of Lockdown on Firm Profitability Conditioned on Firm-Specific Factors
(1) PAT (2) ROS (3) EPS (4) PG
Firm size
Non-essential industries ×post- −0.206*** −0.015 (0.013) −0.077*** −0.019
lockdown ×small (0.056) (0.022) (0.017)
All controls Yes Yes Yes Yes
R2 0.12 0.33 0.15 0.10
Firm age
Non-essential industries ×post- −0.124*** −0.072*** −0.092*** −0.024
lockdown ×old (0.050) (0.015) (0.023) (0.017)
R2 0.10 0.34 0.14 0.09
Group ownership
Non-essential industries ×post- −0.274*** −0.014 (0.015) −0.077*** −0.041**
lockdown ×no-group (0.066) (0.025) (0.019)
R2 0.11 0.45 0.10 0.11
Listed
Non-essential industries ×post- −0.314*** −0.001 (0.015) −0.075*** −0.017
lockdown ×unlisted (0.049) (0.022) (0.018)
R2 0.15 0.40 0.10 0.08
Note: Estimates from regressions are presented with robust standard errors in parentheses. The dependent variables used in Models
(1)–(4) are measures of profit after tax (PAT), return on sales (ROS), earnings per share (EPS) and profit growth (PG), respectively.
All regressions use clustered standard errors at the industry-state-year level. Non-essential industries are a dummy variable that
takes value 1 if the industry is non-essential and 0 otherwise. Post-lockdown is a dummy variable that takes a unit value for quar-
ters after April–June 2020. All regression models include firm-specific variables such as age, leverage and size. Industry-level con-
centration ratios are also used. The models also control for industry, state, year and ownership fixed effects.
*, ** and *** denote statistical significance at 10%, 5% and 1% levels.

or NIFTY-500 index. These results emphasise the role of pre-existing firm-specific conditions in influ-
encing the effect of lockdown on profitability. We also find that these results are valid after control-
ling for multiple hypothesis testing (Table S4A in Appendix S5).
5.4.3. Lockdown stringency and COVID-19 spread

We postulate that the effect of lockdown on profitability is moderated by a set of external variables.
We explore these extensions by employing an interaction of these moderating variables with the dif-
ference in difference variable in the regression framework from Equation (1). To examine heteroge-
neous effect of COVID-19 of firm profitability, we mainly consider two types of heterogeneous effect
– lockdown stringency and COVID-19 spread. For lockdown stringency, we use the state-level daily
data from GCM that provides the percentage change in mobility on that day as compared to the same
day of the week in the reference period (median value for five weeks between January and February
2020).
We consider workplace mobility on the first day of the post-event quarter (April 1, 2020) and on
the last Wednesday of the quarter (24 June 2020, since 1 April 2020 was also a Wednesday). Using
the national workplace mobility on the respective dates, we group states under more stringent lock-
down and less stringent lockdown on both the dates. As a next step, we construct our ‘strict lock-
down’ indicator variable if the state is under a strict lockdown on both dates.
The top panel of Table 9 indicates that firms that were under a stricter lockdown, experienced
lower PAT as compared to firms operating in a less strict lockdown environment in our sample. Simi-
larly, these firms also recorded lower ROS, EPS and PG as compared to their counterparts. These
results highlight the significance of institutions and external factors in shaping the profitability of
firms.
Our COVID-19 spread is based on the variation on the pace of COVID-19 at the state-level in India
to test the relevance of external factors in influencing the impact of lockdown. We use data on
growth in COVID-19 cases between April 1 and June 30 for each state and merge it with our main
dataset using location identifier(s). We estimate model (2) and examine the effect of the COVID-19
interaction term with our DD variable. The lower panel of Table 9 compiles the results. We find that
firms that have seen rapid growth in infections have experienced lower PAT, EPS, ROS and PG as
compared to firms that are operating in states with a lower spread of COVID-19. This validates the
role of external factors in affecting how the lockdown affects profitability. These results are valid even
after controlling for multiple hypothesis testing (Table S4B of Appendix S5).
Table 9. Effect of Lockdown on Profitability Conditioned on Lockdown Stringency and COVID-19 Spread
PAT ROS EPS PG
Lockdown stringency
Non-essential industries × post- −0.283*** −0.112*** −0.108*** −0.023
lockdown × strict lockdown (0.085) (0.021) (0.031) (0.024)
R2 0.12 0.10 0.15 0.10
COVID-19 spread rate
Non-essential industries × post- −0.133*** −0.085*** −0.069*** −0.028*
lockdown × high spread (0.045) (0.014) (0.021) (0.016)
R2 0.12 0.34 0.15 0.09
Note: *, ** and *** denote statistical significance at 10%, 5% and 1% levels.

16 ECONOMIC PAPERS
6. Discussion and Conclusion

While lockdowns are essential for inhibiting the spread of COVID-19 infections and hence human
lives, it costs the economy both on the social and economic front. We document the economic cost
by characterising the effect of lockdown on the profitability of firms in India. We use the earliest
available data on firm operations after the lockdown and employ a DD regression framework to eval-
uate the impact. We observe a dramatic drop in all profit measures which are driven by both demand
and supply-side shocks to the firm.
The decline in profitability is higher in firms which are operating in an environment with higher
stringency of lockdowns and higher COVID-19 exposure. These results strongly highlight the role of
institutions in shaping firm performance. Moreover, it also links to supply chain management and
diversification in location while setting up plants. Even though a concentrated production line might
lead to better coordination among different branches, but it also suffers from high risks associated
with labour strikes, political changes in the given state and disruptions due to adverse climate
changes.
This analysis pertains to firms that were positioned favourable since they were able to publish the
quarterly financial report even during the lockdown. Based on the effect size of lockdown for these
firms, one can assert that the adverse effects of the more vulnerable set of firms will be even higher.
These adverse effects on profitability will deepen in the future quarters as the demand and supply
shocks due to the pandemic continue to expand. These results point towards a crisis in the Indian
industrial sector that requires the immediate attention of the policy-makers.
In this paper, we do not focus on the state relief measures as most of these were implemented after
period of our analysis. The paper pushes for government stimulus packages to be more focused on
the vulnerable group of firms such as the MSMEs and firms in stressed sectors such as aviation and
construction. The government also needs to provide income support to boost demand to avoid the
deepening of these effects. With the historically highest contraction in economic activity in the Indian
economy, the government must ensure a quick, easy and effective implementation method to avoid
a spiralling crisis to widen.
Data availability statement

The data is not a publicly available free dataset and hence cannot be shared due to copyright issues.
REFERENCES
Acemoglu, D., Akcigit, U. and Kerr, W.R. (2016), ‘Innovation Network’, Proceedings of the National Academy of
Sciences, 113 (41), 11483–8.
Aggarwal, N. and Narayanan, S. (2021), ‘Impact of India’s Demonetization on Domestic Agricultural Markets’,
Available at: SSRN 3066042.
Alfaro, L., Chari, A., Greenland, A.N. and Schott, P.K. (2020), ‘Aggregate and Firm-Level Stock Returns During
Pandemics, in Real Time (No. w26950)’, National Bureau of Economic Research.
Alsamhi, M.H., Al-Ofairi, F.A., Farhan, N.H., Al-ahdal, W.M. and Siddiqui, A. (2022), ‘Impact of Covid-19 on
Firms’ Performance: Empirical Evidence from India’, Cogent Business & Management, 9 (1), 2044593.
Anderson, M.L. (2008), ‘Multiple Inference and Gender Differences in the Effects of Early Intervention: A Reeval-
uation of the Abecedarian, Perry Preschool, and Early Training Projects’, Journal of the American Statistical Associa-
tion, 103 (484), 1481–95.
Apedo-Amah, M.C., Avdiu, B., Cirera, X., Cruz, M., Davies, E. and Grover, A. (2020), ‘Unmasking the Impact of
COVID-19 on Businesses’.
Bachas, P., Brockmeyer, A. and Semelet, C. (2020), The Impact of COVID-19 on Formal Firms’.
Bekaert, G., Engstrom, E. and Ermolov, A. (2020), Aggregate Demand and Aggregate Supply Effects of COVID-19:
A Real-time Analysis. Available at: SSRN 3611399.
Biswas, S. (2021), ‘Can R&D Investment Reduce the Impact of COVID-19 on Firm Performance?—Evidence from
India’, Journal of Public Affairs, 22 (S1), e2773.
Brahmana, R.K., Setiawan, D. and Trinugroho, I. (2022), ‘The Impact of Government Nonmarket Policy on a
Firm’s Financial Performance: A Lesson from COVID-19 Pandemic Lockdown’s Policy’, Asia-Pacific Journal of
Business Administration. doi: 10.1108/apjba-05-2021-0204
Bretscher, L., Hsu, A., Simasek, P. and Tamoni, A. (2020), ‘COVID-19 and the Cross-Section of Equity Returns:
Impact and Transmission’, The Review of Asset Pricing Studies, 10 (4), 705–41.

Carletti, E., Oliviero, T., Pagano, M., Pelizzon, L. and Subrahmanyam, M.G. (2020), ‘The COVID-19 Shock and
Equity Shortfall: Firm-Level Evidence from Italy’, The Review of Corporate Finance Studies, 9 (3), 534–68.
Clarke, D., Romano, J.P. and Wolf, M. (2020), ‘The Romano–Wolf Multiple-Hypothesis Correction in Stata’, The
Stata Journal, 20 (4), 812–43.
Davis, S.J., Liu, D. and Sheng, X.S. (2022), ‘Stock Prices and Economic Activity in the Time of Coronavirus’, IMF
Economic Review, 70 (1), 32–67.
Ding, W., Levine, R., Lin, C. and Xie, W. (2020), Corporate Immunity to the COVID-19 Pandemic (No. w27055)’
National Bureau of Economic Research.
Ferraresi, M., Kotsogiannēs, C., Rizzo, L. and Secomandi, R. (2020), COVID-19: "lockdown" and Institutions. Soci-
età italiana di economia pubblica.
Fischer, C. (2020), ‘Six Degrees of Separation for COVID-19: The External Costs and Benefits of Extended Distanc-
ing for Different Social Groups. COVID’, Economics, 1, 1–22.
Gabaix, X. (2011), ‘The Granular Origins of Aggregate Fluctuations’, Econometrica, 79 (3), 733–72.
Golubeva, O. (2021), ‘Firms’ Performance during the COVID-19 Outbreak: International Evidence from 13 Coun-
tries. Corporate Governance: The International Journal of Business’, Society, 21 (6), 1011–27.
Goodman-Bacon, A. and Marcus, J. (2020), ‘Using Difference-in-Differences to Identify Causal Effects of COVID-
19 Policies’, Survey Research Methods, 14, 153–8.
Guerrieri, V., Lorenzoni, G., Straub, L. and Werning, I. (2020), ‘Macroeconomic Implications of COVID-19: Can
Negative Supply Shocks Cause Demand Shortages? (No. w26918)’, National Bureau of Economic Research.
Hoehn-Velasco, L., Silverio-Murillo, A. and de la Miyar, J.R.B. (2021), ‘The Long Downturn: The Impact of the
Great Lockdown on Formal Employment’, Journal of Economics and Business, 115, 105983.
Hu, S. and Zhang, Y. (2021), ‘COVID-19 Pandemic and Firm Performance: Cross-Country Evidence’, International
Review of Economics & Finance, 74, 365–72.
Inoue, H. and Todo, Y. (2019), ‘Firm-Level Propagation of Shocks Through Supply-Chain Networks’, Nature Sus-
tainability, 2 (9), 841–7.
Jain, R. (2021), ‘Information and Communication Technology Adoption and the Demand for Female Labor: The
Case of Indian Industry’, The BE Journal of Economic Analysis & Policy, 21 (2), 695–722.
Kannan, K.P. (2020), ‘A Low Growth, No Employment and No Hope Budget for ‘Aspirational India’’, Economic &
Political Weekly, 55 (9), 27.
Khanna, T. and Palepu, K. (2000), ‘Is Group Affiliation Profitable in Emerging Markets? An Analysis of Diversified
Indian Business Groups’, The Journal of Finance, 55 (2), 867–91.
Liu, H., Yi, X. and Yin, L. (2021), ‘The Impact of Operating Flexibility on firms’ Performance during the COVID-19
Outbreak: Evidence from China’, Finance Research Letters, 38, 101808.
Mathur, A.K. and Sengupta, R. (2020), ‘Covid-19 and Corporate India’, The Leap Blog. Available at: https://blog.
theleapjournal.org/2020/06/covid-19-and-corporate-india.html
Mitra, A., Shrivastav, P.K. and Singh, G.P. (2022), ‘Livelihood Volatility in the Urban Labour Market’, Economic &
Political Weekly, 57 (35), 63.
Nagaraj, R. (2019), ‘Make in India: Why didn’t the Lion Roar?’, The India Forum, 3. Available at: https://blog.
theleapjournal.org/2020/06/covid-19-and-corporate-india.html
Nagaraj, R. (2020), ‘Understanding India’s Economic Slowdown?’, The India Forum. Available at: https://www.
theindiaforum.in/article/understanding-india-s-economic-slowdown
Papanikolaou, D. and Schmidt, L.D. (2020), ‘Working Remotely and the Supply-Side Impact of COVID-19 (No.
w27330)’, National Bureau of Economic Research.
Qin, X., Huang, G., Shen, H. and Fu, M. (2020), ‘COVID-19 Pandemic and Firm-Level Cash Holding—Moderating
Effect of Goodwill and Goodwill Impairment’, Emerging Markets Finance and Trade, 56 (10), 2243–58.
Rathore, U. and Khanna, S. (2020), ‘From Slowdown to Lockdown: Effects of the COVID-19 Crisis on Small Firms
in India’, Available at: SSRN 3615339.
Romano, J.P. and Wolf, M. (2016), ‘Efficient Computation of Adjusted P-Values for Resampling-Based Stepdown
Multiple Testing’, Statistics & Probability Letters, 113, 38–40.
Shen, H., Fu, M., Pan, H., Yu, Z. and Chen, Y. (2020), ‘The Impact of the COVID-19 Pandemic on Firm Perfor-
mance’, Emerging Markets Finance and Trade, 56 (10), 2213–30.
Soni, P. (2021), ‘Effects of COVID-19 Lockdown Phases in India: An Atmospheric Perspective’, Environment, Devel-
opment and Sustainability, 23 (8), 12044–55.
Stemmler, H. (2022), The Effects of COVID-19 on Businesses: Key Versus Non-Key Firms, ILO Working Papers
doi: 10.54394/bzkj3115
Tripathi, G., Pathak, H. and Kumar, A. (2021), ‘Impacts of COVID-19-Induced National Lockdown on Agriculture
and Rural Livelihoods in India: A Macro Perspective’, Indian Journal of Agricultural Economics, 76 (3), 333–351.
Yang, J. and Yang, C. (2021), ‘Economic Policy Uncertainty, COVID-19 Lockdown, and Firm-Level Volatility: Evi-
dence from China’, Pacific-Basin Finance Journal, 68, 101597.

18 ECONOMIC PAPERS
Zhang, W. and Hamori, S. (2021), ‘Crude Oil Market and Stock Markets During the COVID-19 Pandemic: Evi-
dence from the US, Japan, and Germany’, International Review of Financial Analysis, 74, 101702.
Zou, P., Huo, D. and Li, M. (2020), ‘The Impact of the COVID-19 Pandemic on Firms: A Survey in Guangdong Pro-
vince, China’, Global Health Research and Policy, 5 (1), 1–10.

Covid19 Lockdown FirmProfit Jain2023

Uploaded by

Copyright:

Available Formats

Covid19 Lockdown FirmProfit Jain2023

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Covid19 Lockdown FirmProfit Jain2023

Uploaded by

Copyright:

Available Formats

17593441, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1759-3441.12377 by Indian Institute Of Technology, Wiley Online Library on [25/02/2023].

Effect of COVID-19 Lockdown on the

We examine the effect of COVID-19-induced lockdown on the profitability

Ó 2023 The Economic Society of Australia.

Ó 2023 The Economic Society of Australia.

2. The Indian Context

Pre- lockdown Central govt State-level

First Case of Lockdown Phased unlocking

Jan 27, 2020 Mar 25, 2020 Jun 1, 2020

Ó 2023 The Economic Society of Australia.

3. Data and Variables

3.1. Data Source

Ó 2023 The Economic Society of Australia.

Ó 2023 The Economic Society of Australia.

Ó 2023 The Economic Society of Australia.

yit ¼ α0 þ α1 D2020,t þ α2 NEtreat,j þ α3 D2020,t NEtreat,j

5.1. Main Model

Figure 2. Profitability Trend by the Treatment and the Control Group

Ó 2023 The Economic Society of Australia.

Table 1. Results for the Parallel Trends Assumption

(1) PAT (2) ROS (3) EPS (4) PG

April–June quarters (2015–2019)

Ó 2023 The Economic Society of Australia.

Table 2. . Effect of Lockdown on Firm Profitability in India

(1) PAT (2) ROS (3) EPS (4) PG

April–June quarters (2015–2020)

5.2. Robustness Tests

5.2.1. Falsification tests

Ó 2023 The Economic Society of Australia.

Table 3. Robustness Tests for the Effect of Lockdown on Firm Profitability

(1) PAT (2) ROS (3) EPS (4) PG

Falsification test 1 – randomly generated treatment indicator

5.2.2. An alternate difference in difference estimation

Table 4. Effect of lockdown using an alternate difference in difference estimation

(1) PAT (2) ROS (3) EPS (4) PG

Period of analysis: 2015–2020

Ó 2023 The Economic Society of Australia.

5.2.3. Multiple hypothesis problem

5.3. Potential Channels

Table 5. Multiple hypothesis testing- impact of lockdown (DD estimate)

(1) PAT (2) ROS (3) EPS (4) PG

April–June quarters (2015–2020)

Ó 2023 The Economic Society of Australia.

Table 6. Main Channels that Drive the Result

Category Cost Revenue

(1) Ln (real (2) Expense (3) Ln (real (4) Income

Number of firms 5307 1232 4075

5.4. Some Extensions

5.4.1. Missing firms during the lockdown

Ó 2023 The Economic Society of Australia.

5.4.2. Firm resilience and the main effect

Table 8. Effect of Lockdown on Firm Profitability Conditioned on Firm-Specific Factors

(1) PAT (2) ROS (3) EPS (4) PG

Ó 2023 The Economic Society of Australia.

5.4.3. Lockdown stringency and COVID-19 spread

PAT ROS EPS PG

Note: *, ** and *** denote statistical significance at 10%, 5% and 1% levels.

Ó 2023 The Economic Society of Australia.

6. Discussion and Conclusion

Note: , and denote statistical significance at 10%, 5% and 1% levels.