![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
Abstract
Drawing on a representative survey of firms in 38 countries, eight of which are in Sub-Saharan Africa (SSA), this paper documents the impact of COVID-19 and firms’ coping strategies in SSA, benchmarking with other regions. The paper shows that the impact of the pandemic is more pronounced in Sub-Saharan Africa compared with other regions. This disproportionate impact is not explained by differences in sectoral composition and other firm characteristics, but likely by the level of development. This underscores the important economic and structural contexts that predate the pandemic in understanding the differential impact. Contrary to expectations, the findings show that businesses in Sub-Saharan Africa are more likely to adjust their operations or products and services to adapt to the shock than those in other regions. However, firms in the region lag in leveraging digital technologies, remote working, and e-commerce, compared with those in other regions.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
The data that support the findings of this study are freely available for download from the Enterprise Surveys website (https://login.enterprisesurveys.org/content/sites/financeandprivatesector/en/signin.html). The STATA do file to generate all results presented in the paper are available from authors upon request.
Notes
1 According to the World Bank’s Enterprise Surveys, SSA firms are 35 percent smaller than those in other locations.
2 In fact, according to a recent COVID-19 focused business survey conducted by the Enterprise Analysis Unit of the World Bank, only 3 per cent of firms in SSA have received some form of government support for CVOID-19 relief compared with 60 per cent for ECA, 15 per cent for LAC and 30 per cent MNA. Firms in SSA are in effect left to either sink or swim.
3 The list of countries with this survey can be found here: https://www.enterprisesurveys.org/en/covid-19 Anonymized firm-level data and survey documentation including questionnaires can be found from the ES website: https://bit.ly/3bFHlVB.
4 Similar surveys have been conducted for additional countries in SSA (https://www.worldbank.org/en/data/interactive/2021/01/19/covid-19-business-pulse-survey-dashboard). However, as of the write-up of this paper, the firm-level data for these surveys are not available. In addition, these additional surveys don’t have pre-COVID baseline information. Consequently, our sample of countries is restricted to those covered by the Enterprise Analysis Unit’s follow-up survey, for which anonymized firm-level data has been published following the tradition of the World Bank Group Enterprise Surveys data.
5 We use index of stringency of containment measures compiled and produced by Oxford University’s COVID-19 policy response tracker. Data and documentation can be found at: https://www.bsg.ox.ac.uk/research/research-projects/covid-19-government-response-tracker.
6 In all estimations, we use sampling weights computed for the ES COVID-19 follow-up surveys, an adjusted version of the standard ES sampling weights to account for firms that may have existed since the ES and those that have become ineligible. For the regression where dependent variable is permanent closure, we use the standard ES weights. The model is also estimated by using re-scaled sampling weights in order to minimize the results being driven by countries with large sample size. The results are broadly unchanged (see in the Online Appendix).
7 We control for survey period in all specifications since data collection period is staggered across countries, coinciding with different stages of the pandemic.
8 For binary dependent variables, the specification is estimated using a linear probability model and a logit model. The results using both specifications are nearly identical. For ease of interpretation, we present the results for linear specifications in the main text, and provide estimates of the logit model in the Online Appendix (see ).
9 The lower income category also includes economies that are classified as lower-middle income economies; hence two of the countries in our sample, Zambia and Zimbabwe, that are lower-middle income countries also fall in to the lower income classification.
10 We re-run the same regression with all controls by comparing SSA and low-income countries vs. high-income economies. The results confirms that the effects of the shock is significantly higher for firms in lower income economies compared with those in higher-income economies regardless of whether they are in SSA or elsewhere; see Online Appendix .
11 For instance, redundancy costs are higher in SSA and South Asia regions compared to high income economies. In Zambia, one of the countries in our sample, redundancy costs is about 20 months of salary for an employee being retrenched after 10 years in the business (Word-Bank, Citation2020). Adoption of employment protection rules are also on the rise in SSA, particularly in the Francophone countries (Kudzordzi, Citation2019).
12 While this result is robust to controlling for host of firm-level control variables (including sectoral composition, age, access external finance etc.), full explaining this interesting result is left for future research.
13 Within SSA, Somalia and Togo are among the most significantly affected. Interestingly, while about 90 per cent of firms in Chad and Somalia report facing liquidity and cash flow shortages, only about 15 per cent of firms in Chad delayed payment to service providers compared to about 90 per cent of firms in Somalia.