Socioeconomic Impacts of COVID-19 in Kenya On Firms Rapid Response Phone Survey Round 1, January 2021 Socioeconomic Impacts of COVID-19 in Kenya On Firms Rapid Response Phone Survey Round 1, January 2021 Table of Contents ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 IMPACT OF COVID-19 ON BUSINESSES IN KENYA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1. Operations of the Business . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Impact on Sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3. Impact on Employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 4. Main Channels of Transmission of the Impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 5. Firm Survival Expectations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 EXPECTATIONS ABOUT THE FUTURE AND UNCERTAINTY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 RESPONSES TO THE SHOCK: DIGITAL ADOPTION AND INNOVATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 THE ROLE OF POLICY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 POLICY RECOMMENDATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 1. General Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2. Recommendations on Targeting Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 APPENDIX 1. DESCRIPTION OF THE SAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 APPENDIX 2. ADDITIONAL RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 APPENDIX 3. RESULTS FROM OLS AND PROBIT REGRESSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 iii LIST OF TABLES Table 1: Estimated number of jobs in businesses affected by the pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Table 2: Change in sales across business characteristics (%) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Table 3: Estimated fraction of workers affected by margin of labor adjustment (% of workers) . . . . . . . . . . . . . . . . . . 11 Table A1.1: Number of firms by sector, size, region, and exporting status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Table A2.1: Expected time to resume operations (fraction of temporarily closed business) . . . . . . . . . . . . . . . . . . . . . 35 Table A2.2: Weeks that business can remain open in current circumstances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Table A2.3: Estimated number of workers affected by margin of adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Table A3.1: Estimated correlation between employment adjustments and business characteristics . . . . . . . . . . . . . 38 Table A3.2: Estimated correlation between shocks and business characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Table A3.3: Estimated correlation between sales reductions and shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Table A3.4: Estimated correlation between firm survival and business characteristics . . . . . . . . . . . . . . . . . . . . . . . . . 41 Table A3.5: Estimated correlation between responses to the pandemic and business characteristics . . . . . . . . . . . 42 Table A3.6: Estimated correlation between type of digital platform used and business characteristics . . . . . . . . . . 43 Table A3.7: Estimated correlation between self-reported most needed policies and business characteristics . . . . 44 Table A3.8: Estimated correlation between policy demand and shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 LIST OF FIGURES Figure 1: Shocks to businesses from the COVID-19 pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Figure 2: Changes in mobility over time (percent change compared to baseline) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Figure 3: Firm operating status by region, size, and female employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 4: Firm operating status by sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Figure 5: Change in sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 6: Distribution of the reduction in sales explained by the observed firm characteristics . . . . . . . . . . . . . . . . . 8 Figure 7: Average adjusted percentage change in sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Figure 8: Margin of adjustment in employment by month . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure 9: Fraction of firms affected by transmission channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 10: Estimated correlation between the change in sales and the shocks of COVID-19 . . . . . . . . . . . . . . . . . . . 12 Figure 11: Number of weeks that businesses can remain open in current circumstances . . . . . . . . . . . . . . . . . . . . . . . 13 Figure 12: Number of days a business can cover costs with available cash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Figure 13: Average change in sales expected for the next six months across scenarios . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 14: Expectations and uncertainty about sales growth for the next six months . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 15: Distribution of expectations about growth in sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 16: Distribution of uncertainty about growth in sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 17: Average change in employment expected for the next six months across scenarios . . . . . . . . . . . . . . . . . 15 Figure 18: Expectations and uncertainty about employment growth for the next six months . . . . . . . . . . . . . . . . . . . 15 Figure 19: Distribution of expectations about growth in employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Figure 20: Distribution of uncertainty about growth in employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Figure 21: Business responses to the COVID-19 shock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 iv    Table of Contents Figure 22: Predictive effect of firm characteristics on responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Figure 23: Type of digital platform function used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure 24: Average adjusted probability of starting or increasing the use of digital platforms . . . . . . . . . . . . . . . . . . 19 Figure 25: Distribution of expectations about growth in employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 26: Type of assistance received . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 27: Reason for not receiving assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 28: Not being aware of programs by firm characteristic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 29: Self-reported most needed public policies to support businesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure A2.1: Changes in mobility over time (percent change compared to baseline) . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Figure A2.2: Margin of adjustment in employment by month and sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure A2.3: Predictive effect of size and sector on weeks a business can remain open . . . . . . . . . . . . . . . . . . . . . . 33 Figure A2.4: Predictive effect of size and sector on days a business can cover costs with available cash . . . . . . . . 33 Figure A2.5: Average change in sales expected for the next six months by sector and size . . . . . . . . . . . . . . . . . . . 34 Figure A2.6: Average change in employment expected for the next six months across scenarios . . . . . . . . . . . . . . 34 Figure A2.7: Predictive effect of shocks on top three most needed public policies . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Table of Contents   v Acknowledgments This background note was prepared by a team co-led by Marcio Cruz (Senior Economist, ETIFE) and Utz Pape (Senior Economist, EAEPV), and included Alastair Haynes (Consultant, EAEPV) and Henry Stemmler (Consultant, EAEPV), with additional contributions from Jesica Torres Coronado (Consultant, ETIFE) and Kyung Min Lee (Junior Professional Officer, ETIFE). The team also benefitted from the insightful comments provided by Xavier Cirera (Senior Economist, ETIFE) as well as Gabi Afram (Lead Financial Sector Economist, EAEF1), Cecilia Paradi-Guilford (Private Sector Special- ist, EAEF1), and Zenaida Uriz (Senior Private Sector Specialist, EAEF2), and the guidance received by Denis ­ Medvedev (Practice Manager, ETIFE), Niraj Verma (Practice Manager, EAEF1), and Pierella Paci (Practice Manager, EAEPV). The team would like to express its gratitude to REMIT for the outstanding efforts on data collection. In particular, the team would like to thank the manager, Carolyne Nekesa, and the field officer, Sam Balongo, leading the data collection and the enumerators, as well as supervisors collecting data. The team would like to thank the Kenya National Bureau of Statistics (KNBS) and researchers from the University of California, Berkeley, for their vision, commitment to this work, and collaboration in the efforts of data collection. In particular, the team would like to thank Dennis Egger (University of Berkeley) and Bhumi Purohit (University of Berkeley) for their contributions. vi Abstract The COVID-19 pandemic has severe impacts on the Kenyan economy and society as a whole. This report analyzes the impact of COVID-19 on businesses in Kenya based on a nationally representative Business Pulse Survey imple- mented by the World Bank between June and August, 2020. The results indicate that about 93 percent of firms expe- rienced a decline of sales compared to the same period of the previous year. Sales dropped by around 50 percent in the average and median Kenyan firms, and by more than 70 percent for one-quarter of firms. Close to 65 percent of firms are experiencing a decline in demand, cash flow, and available finance. Moreover, firms expect sales to continue declining in the coming months. The pandemic is disproportionally affecting small and female-owned busi- nesses. Firms in Kenya are responding to the crisis through the adoption of digital technologies. About 20 percent of firms have received public support, but lack of awareness of public assistance options is still large among those that did not receive any support. Finally, the COVID-19 Business Pulse Survey (COV-BPS) suggests policy response options divided into four areas: access to finance, firm capabilities, access to new markets, and reducing uncertainty. Additional follow-up surveys are being conducted for monitoring the current circumstances and updating the policy recommendations. vii Executive Summary The COVID-19 pandemic has severe impacts on the Kenyan economy and society as a whole. Globally, house- holds and firms are struggling to deal with the consequences of the pandemic. The economic and social disruptions are creating multiple challenges for the private sector. Firstly, firms are facing lower demand for goods and services. Secondly, supply chains are disrupted, restricting access to intermediate goods and labor. Thirdly, access to cash and credit is deteriorating. Lastly, uncertainty is dampening investment and innovation prospects. Firms in Kenya are not exempt from these developments. Sales have plummeted for almost all firms. More than 9 out of 10 firms have experienced a decline of sales as a consequence of the pandemic. Sales dropped by around 50 percent in the average and median Kenyan firms, and by more than 70 percent for a quarter of the firms. Firms in the tourism sector, which consists of accommodation and food services, have experienced the largest decreases in sales and are not often fully operating. In the median firms in the tourism sector, sales have fallen by 70 percent. One in three workers is facing high vulnerability. One in five workers has lost their jobs after the start of the pan- demic. Relatively few firms have resorted to other labor adjustment measures, such as reducing the working hours or wages. Firms in tourism and other services have laid off more workers than firms in other sectors. While two-thirds of all businesses are still open, 20 percent of workers are in businesses that are temporarily closed and 16 percent in businesses that are only partially open. Especially workers in small and medium-sized firms in tourism and other services sectors, as well as in older and non-exporting firms, are facing high vulnerability. More than 50 percent of the jobs in the tourism sector and 45 percent of jobs in other service sectors are vulnerable. Close to two-thirds of firms are experiencing a decline in demand, cash flow, and available finance. 62 percent of firms have reduced working hours, and 54 percent face a lower availability of inputs. The pandemic has hit firms in the tourism and other service sectors, as well as firms with a larger female workforce, primarily through fewer working hours, while exporting firms are more affected by a lower availability of inputs. Furthermore, the reduction in cash flow has had a large impact on sales of firms in manufacturing. Lower demand is more severely affecting sales of businesses in retail. Under the current circumstances, the median firm is able to remain open for five months and can cover costs with available cash for about 4 weeks. On average, a firm in Kenya can remain open for close to 19 weeks. The median firm can remain open for 20 weeks, indicating little variation between firms. Larger firms and firms in the agriculture and trade sectors report being able to remain open for a larger number of weeks under the current circumstances. Firms that are only partially open can remain operative for fewer weeks than fully open firms (14 weeks vs. 19 weeks). While Kenyan firms are able to continue to cover costs for 47 days on average, the median firm can cover costs for only 30 days. The large difference suggests a large variability in cash availability. Vulnerable firms can cover costs for less than half as long as fully open firms. Firms in agriculture and manufacturing firms are able to cover costs for longer than firms in the other sectors. Firms expect sales to continue to decline. On average, Kenyan firms expect sales to decrease by 27 percent in the first six months of 2021 compared to the previous year. While the expected decline is large, there is little variation in expectation between firms. Almost all firms expect sales to decline, and only a few firms expect an increase in sales. viii Firms anticipate employment to decrease at a slightly lower rate than sales. Large firms and firms in agriculture or manufacturing are more optimistic about the first 6 months of 2021. Small and micro-sized firms are more severely affected by the pandemic than larger firms. Micro-sized firms are often forced to permanently close or temporarily cease operations as compared to larger firms. Larger firms are most often still open, indicating that they can cope better with the aftermath of the pandemic. A larger drop in sales for micro-sized and small firms further underpins them being primarily affected. While by definition micro-sized firms have laid off a smaller number of workers, controlling for observable firm characteristics, they have reduced more hours of their employees. Large firms are able to cover running costs for longer periods than smaller firms. The pandemic is disproportionally affecting businesses with a large share of female employees. Firms in which more than half of all employees are female are 18 percentage points less often open than firms with fewer female employment. Over 40 percent of businesses with a large female workforce are temporarily closed, leaving women more vulnerable than men. However, firms with more female employees less often lay off workers and more often grant leaves of absence. Almost 50 percent of all firms with five or more employees are starting to use, or increase the use of digital plat- forms. The type of response to the shock varies by firm characteristic. Larger firms more often increase the use of digital platforms. Firms in the tourism sector are less able to reap the benefits of digital platforms. Exporting firms and firms in agriculture are less likely to repackage their product mix. Of the firms that use digital platforms, 75 percent use them for business administration purposes and close to 50 percent for marketing. Larger firms more often use digital platforms in other ways, for instance, for supply chain management, payment methods, or sales. Also, younger firms are more versa- tile in their use of digital tools. Firms in food services could potentially take more advantage of service delivery options. About 20 percent of firms have received public support, but a large share of those firms that did not receive any support have reported lack of awareness of government assistance programs. 20 percent of firms in Kenya have received public support during the COVID-19 pandemic. Smaller firms less often receive assistance. 80 percent of firms report not having received assistance because they were not aware of any government measures. Awareness of existing government programs is similar over firm characteristics. Of the firms that received any assistance, 40 per- cent received cash transfers and 33 percent received tax deferrals. Firms most often refer to loans with subsidized interest rates as one of the three most needed policy responses. Just over 50 percent of all firms call for loans with subsidized interest, 42 percent for monetary transfers and 25 per- cent for tax deferrals. The type of assistance reported as “most needed” by the firms depends on their characteristics. For instance, exporters disproportionately demand tax deferrals, and agricultural firms and firms in retail are most likely to call for monetary transfers. Micro-sized firms more often call for deferral of rent, mortgage, and utilities, and less often for fiscal exemptions or deferral of loan payments than larger firms. Similarly, firms report different policies to be most needed depending on the type of shock they experienced. Firms hit by a decrease in demand are around 24 percentage points more likely to call for cash transfers, while firms that report a decrease in hours worked are more likely to report tax deferrals as the most needed policy response. To help mitigate the adverse impacts of COVID-19 on firms, the COV-BPS suggests policy response options divided into four areas: access to finance, firm capabilities, access to new markets, and reducing uncertainty. As the crisis continues to evolve, policies must find a balance between short-term interventions to help businesses “keep the lights on” and a sustainable recovery plan that facilitates the selection of the most productive firms. In the recovery phase, policies should be geared toward supporting growth-oriented enterprises, promoting the reallocation of resources to more efficient companies, and avoiding measures that prop up zombie firms (i.e., inefficient firms that survive). Executive Summary   ix Introduction 1.  The COVID-19 Business Pulse Survey (COV-BPS) aims at providing critical information to help policy makers monitor the effects of the pandemic on businesses. The COVID-19 shocks affect businesses through five distinct channels (Figure 1). Firstly, lockdown measures and regulations to control the spread of the pandemic directly affect businesses’ ability to operate and consumers’ ability to procure goods. Secondly, firms are hit by a reduction in demand due to lower consumption, lower export demand, and lower demand of intermediates from other busi- nesses. Thirdly, a decline in the availability of labor and intermediate goods as value chains are disrupted, which limits firms’ capacity to produce and supply goods. Fourth, firms are becoming more financially constrained due to a deterioration in the availability of cash and credit conditions. Lastly, a surge in uncertainty leads to lower incentives to invest and increases risks associated with innovation and entrepreneurship.  FIGURE 1: Shocks to businesses from the COVID-19 pandemic Lockdown effects Supply shocks Public health measures require Global and local value chains are nonessential businesses to close disrupted Decreased supply of labor and intermediate Temporary shock, targeting nonessential inputs, affecting, e.g., firms that rely on imports businesses, mostly in retail, hotels/restaurants (tourism), and personal services Financial shocks Opportunities for finance becoming further constrained Demand shocks Deterioration in availability of credit while Economic downturn drives down demand increases affects access to finance demand domestically and abroad Broad-based shocks will especially hit firms producing durables, apparel/textiles, and those Uncertainty reliant on exports (manufacturing and services, Uncertainty drives down e.g., tourism) investment and innovation Source: World Bank 2020b. 2.  The Kenya COV-BPS is based on a nationally representative sample. The sample consists of 2,070 firms ran- domly selected from the universe of 138,186 firms available in the 2017 Census of Establishments from the Kenyan National Bureau of Statistics (KNBS). The sample was stratified by firm size and sector of activities. The analysis regrouped the size and sectors of activities for the sake of comparability across over 50 countries. Table A1.1 in Appendix 1 describes the distribution of the sample. The response rate was 37 percent, including all firms that the 1 survey team attempted to reach. The survey took place in Kenya between June and August, 2020. Phone interviews were conducted between June 10 and August 30, 2020. The team used Computer Assisted Telephone Interviewing (CATI) through a Survey CTO platform. The results presented in this report use sampling weights. 3.  This report analyzes the results from the Kenya-BPS and provides in-depth information about how the pan- demic is affecting private sector firms and which policies are most needed. The main reaction of the Kenyan gov- ernment has been centered around tax reliefs and reducing fees and obstacles for monetary transfers and access to credit. This report aims to provide a better understanding of the effects of the COVID-19 pandemic on firm operations, labor demand, and expectations of future operations. Thereby, the results can assist in identifying the most needed policies to support the private sector throughout the COVID-19 pandemic. 4.  The Government of Kenya also implemented a series of economic stimulus measures.1 A Central Bank order for banks to waive fees for individuals who move money between their bank account and a mobile wallet came into effect on March 17. The upper limit for mobile money transfers was increased. Authorities reached a deal with com- mercial banks to restructure nonperforming loans caused by COVID-19 layoffs. Additionally, loans and grants are available through government and private funds, including the National Business Compact on CoVid19 (NBCC), as well as special loans through Stanbic Bank and Standard Chartered Bank, among others. The Government of Kenya also disbursed KSh 1 billion for the health care sector and US$5 million for the tourism sector. In addition, the Inter- national Finance Corporation disbursed a US$50 million loan to Equity Bank Kenya to support small and medium enterprises (SMEs). 5.  Mobility decreased substantially in the early months of the pandemic. In the first month after the start of the pandemic and the government initialized restrictions, mobility decreased by up to 50 percent in key areas (Figure 2).2 Activities related to retail and recreation had the sharpest decline. Individuals in turn spent more time at home. As COVID-19 infection rates fell over the course of the year and the government eased restrictions, mobility patterns  FIGURE 2: Changes in mobility over time (percent change compared to baseline) 40 20 0 Percent –20 –40 –60 –80 Feb Feb Mar Mar Mar Apr Apr Apr May May May Jun Jun Jun Jul Jul Jul Aug Aug Aug Sep Sep Sep Oct 15 25 06 16 26 05 15 25 05 15 25 04 14 24 04 14 24 03 13 23 02 12 22 02 Retail/recreation Grocery/pharmacy Workplace Residential Note: Data comes from Google’s “COVID-19 Community Mobility Reports.” The baseline values are each day’s five-week median value between January 3 and February 6, 2020. 1  KPMG. “Kenya: Government and institution measures in response to COVID-19.” https://home.kpmg/xx/en/home/insights/2020/04/kenya- government-and-institution-measures-in-response-to-covid.html 2  Mobility data come from Google’s “COVID-19 Community Mobility Reports” and are based on location histories of individual’s Google accounts. The data are therefore not representative of the country as a whole, but give an indication of how the pandemic has changed mobility patterns. 2    Socioeconomic Impacts of COVID-19 in Kenya slowly converged back to the baseline rates. However, at the beginning of October, people still less often engaged in retail and recreational activities or went to the workplace. Mobility hasn’t recovered equally across all locations in Kenya, however. In Nairobi, retail and recreational and workplace activities are still 13–20 percent below the baseline value, while in Mombasa the gap is only between 1 and 10 percent (Figure A2.1). 6.  The remainder of this report is organized as follows. Section 2 describes the impact of COVID-19 on operations, sales, and employment, and analyzes the main transmission mechanisms. Section 3 analyzes the current status of firms in terms of expectation and uncertainty. Section 4 analyzes how firms are reacting regarding the adoption of digital technologies and changes in the product mix in response to the COVID-19 shock. Section 5 analyzes the main policy responses. Finally, Section 6 provides a set of policy recommendations based on these findings and comple- mentary analyses across countries. Introduction   3 Impact of COVID-19 on Businesses in Kenya 1.  Operations of the Business 7.  More than one-third of all firms are temporarily closed or only partially open. More than half of the firms were fully open and one-tenth of firms was partially open at the time of this survey. Firms based in Nairobi are more often fully open as compared to firms in other regions and are less often mandated to close temporarily.3 The pan- demic is disproportionally affecting businesses with a large female employment share. Firms in which more than half of employees are female are 18 percentage points more often closed than firms with fewer female employees (Figure 3).4 8.  Larger firms were more likely to remain fully operative than smaller firms. Firms’ operating status largely dif- fers by size. Eighty percent of large firms (100+ employees) are fully open, compared to only 50 percent of small (5–19 employees) and medium-sized (20–99 employees) firms. Micro-sized firms (0–4 employees) are more often forced to permanently close or temporarily cease operations by their own choice, indicating that they are less able to deal with the aftermath of the pandemic on their own.  FIGURE 3: Firm operating status by region, size, and female employment 100 80 Percent of firms 60 40 20 0 All Other Nairobi Micro Small Medium Large Below 50% Above 50% regions (0–4) (5–19) (20–99) (100+) Region Size Female employment Open Partially open (mandated) Temporarily closed (mandated) Temporarily closed (own choice) Permanently closed 3  Being closed by mandate refers to government regulations which ordered firms to close temporarily. These estimates need to be interpreted with caution, given that among nonresponses there might be a disproportionally larger number of temporary or permanently closed businesses. 4  The median share of female employment is 35 percent. In roughly 20 percent of firms more than 50 percent of employees are female. 4 9.  The pandemic is affecting some sectors of the economy more strongly than others. The majority of agricultural and manufacturing firms have been able to remain open. Within the service sector there are large differences in the operating status of firms. As the government mandated schools to close, almost all enterprises in education were closed temporarily. Furthermore, the pandemic is particularly affecting firms in the accommodation and food service sectors. They are less likely to be fully open and more likely closed by mandate than firms from other sectors. More- over, the transportation and storage sectors, as well as accommodation and food service sectors, have a relatively large share of firms only partially open by mandate. This reflects curfews and lockdown restrictions primarily affecting firms of these sectors (Figure 4).  FIGURE 4: Firm operating status by sector 100 80 Percent of firms 60 40 20 0 Agriculture/mining Manufacturing Construction/utilities Retail/wholesale Transportation/storage Accommodation Food services Information/communication Financial/ real estate Education Health/social work Other services Open Partially open (mandated) Temporarily closed (mandated) Temporarily closed (own choice) Permanently closed 10.  Thirty-three percent of workers are in firms facing high levels of vulnerability. Firms are defined as vulnerable if they are partially open or temporarily closed, as these firms could potentially run into liquidity problems and are more likely to permanently close. By this definition, 48 percent of workers in small firms and 50 percent of workers in medium-sized firms are employed in vulnerable firms, compared with only 26 percent of those working for large firms. There is also a large variation of vulnerable firms between sectors. More than 50 percent of jobs are vulnerable in the tourism sector, compared to 8 percent in manufacturing firms.5 Moreover, workers in older and non-­ exporting firms, as well as firms with more than 50 percent of employees being female are more vulnerable. Despite only 1 percent of workers being in permanently closed firms, the large proportion working in vulnerable businesses is concerning (Table 1). 5  The tourism sector consists of accommodation and food services. Impact of COVID-19 on Businesses in Kenya   5  TABLE 1: Estimated number of jobs in businesses affected by the pandemic Vulnerable (partially Partially Temporarily Temporarily open + open closed closed temporarily Permanently Number Open (mandated) (mandated) (by choice) closed) closed of   (%) (%) (%) (%) (%) (%) workers* Total 65 16 15 4 34 1 3,112,467 Micro (0–4) 68 10 8 11 29 3 82,958 Small (5–19) 51 11 30 7 48 1 341,923 Medium (20–99) 49 11 35 4 50 1 725,240 Large (100+) 74 19 5 3 26 0 1,962,347 Agriculture/mining 75 8 1 14 23 2 339,061 Manufacturing 90 6 1 1 8 2 258,874 Retail/wholesale 83 11 4 2 17 1 605,929 Tourism 45 23 22 7 52 3 72,444 Other services 55 20 22 2 45 0 1,836,159 Nairobi 70 26 2 2 30 0 1,231,716 Other regions 63 9 23 5 36 1 1,880,751 Young (0–4) 67 10 9 15 33 0 111,626 Maturing (5–14) 66 12 15 6 33 1 1,166,703 Established (15+) 65 19 15 1 35 0 1,834,138 Exporter 76 16 4 3 24 0 132,762 Non-exporter 65 16 15 4 34 1 2,979,705 Below 50% 67 17 12 4 32 0 2,659,370 female employment Above 50% 56 11 30 2 43 1 453,097 female employment *Note: The total number of paid jobs adds the number of full-time paid jobs and half the number of part-time jobs. 2.  Impact on Sales 11.  Almost all firms have experienced a decline in sales. Ninety-three percent of firms report a reduction of sales in the last 30 days compared to the same period in 2019, while only 2 percent report an increase (Figure 5, panel a). Sales have declined across all firm types, but the magnitude differs between firms depending on their size and sec- tor. For instance, sales have decreased more for firms in the accommodation and food sectors as compared to most other sectors. The pandemic has larger effects on firms in these sectors due to lockdowns, capacity restrictions, and people avoiding larger crowds. Firms in the agriculture, information and communication, financial, and real estate and the social services sectors have experienced the smallest declines in sales (Figure 5, panel b). Furthermore, micro-sized and small firms have experienced a larger decline in sales than medium-sized and large firms (Table 2). 6    Socioeconomic Impacts of COVID-19 in Kenya  FIGURE 5: Change in sales a. Share of firms with an increase, decrease, or same b. Estimated changes in sales across sectors level of sales Other services Health/social work Education Finance/real estate Information/communication 2% 93% Food services Accommodation 5% Transportation/ storage Retail/wholesale Construction/utilities Manufacturing Agriculture/mining Decrease Remain the same Increase –80 –60 –40 –20 0  TABLE 2: Change in sales across business characteristics (%) Average Median firm 10th 25th firm 75th 90th   decline percentile percentile decline percentile percentile Total –51 –90 –70 –50 –30 –10 Micro (0–4) –54 –90 –80 –50 –30 –10 Small (5–19) –53 –90 –80 –50 –40 –10 Medium (20–99) –46 –80 –60 –50 –30 –10 Large (100+) –39 –80 –60 –50 –20 0 Agriculture/mining –38 –75 –50 –40 –20 0 Manufacturing –50 –90 –75 –50 –30 –10 Retail/wholesale –53 –90 –75 –50 –40 –15 Tourism* –63 –95 –80 –70 –50 –25 Other services –50 –90 –70 –50 –30 –10 Nairobi –50 –90 –75 –50 –30 0 Other regions –52 –90 –70 –50 –30 –15 Young (0–4) –52 –85 –80 –50 –40 –10 Maturing (5–14) –53 –90 –80 –50 –40 –10 Established (15+) –47 –80 –70 –50 –25 –10 Exporter –41 –90 –60 –35 –25 0 Nonexporter –52 –90 –75 –50 –30 –10 Below 50% female –51 –90 –70 –50 –30 –10 employment Above 50% female –51 –90 –75 –50 –30 –10 employment *Note: Tourism refers to accommodation and food services activities. Impact of COVID-19 on Businesses in Kenya   7 There is no statistically significant difference in sales declines between other observable characteristics (region, age, exporting status, and female employment). 12.  Sales dropped by 51 percent, on average, compared to the same period in the previous year. For the median firm in Kenya, the drop was 50 percent (Table 2). For a quarter of firms, sales dropped by over 70 percent. Within each firm type, the decline in sales was highly heterogeneous. For the bottom 10 percent of firms, sales declined by 90 percent, while for the top 10 percent of firms, sales dropped by 10 percent. For the 90th percentile of large, agri- cultural firms in Nairobi, the sales remained the same. As the median and mean decline are similar, outliers are not likely driving the estimated reductions in sales. 13.  Observable business characteristics explain only a small part of the heterogeneity in the reduction in sales, implying that the shock is affecting similar firms differently. The BPS suggests heterogeneity in the shock to the sales of the firm. The interaction of size, sector, region, age, exporting status, and female employment only accounts for 20 percent of the deviations of sales reductions from the overall mean and cannot explain the remaining 8 percent.6 Observable characteristics overpredict reductions in sales closer to the mean, but can not explain large declines nor increases in sales (Figure 6).  FIGURE 6: Distribution of the reduction in sales explained by the observed firm characteristics .020 .015 Probability .010 .005 0 –100 –50 0 50 100 150 Change in sales (%) Change in sales observed in the data Change in sales unexplained by observed characteristics Change in sales explained by observables Note: The figure presents kernel density estimates of the distribution of the change in sales, the change in sales predicted by a linear regression, and the residual of the regression. The estimations of the kernel density use the default specifications in Stata. The linear regression uses dummies for the interaction of size, formality status, sector, region, age, and exporting status. 14.  Firms in Kenya experience a larger average reduction in sales than the majority of African countries. When accounting for country, size, sector, and timing of the survey, firms in Kenya experienced a 65 percent reduction in sales. The magnitude of the change in sales is large when compared with other African countries. For instance, only South Africa reported a larger average decline (76 percent), while in countries such as Côte d’Ivoire, Senegal, and Tanzania the average decline was below 40 percent (Figure 7). 6  Toobtain this estimate, the percentage change in sales is regressed on dummies for the interaction of size, sector, region, age, and exporting status, and both the prediction from the regression (the explained component) and the residual (the unexplained component) are computed. For each observation, the deviation of the sales drop from the overall mean is decomposed into deviations unaccounted for (deviations from – [ yi – y the linear prediction) and deviations accounted for (deviations of the linear prediction from the overall mean): yi – y ^i] + [ y – ], where y ^i – y i – is the overall mean. The ratio of each ^i is the prediction from the linear regression, and y corresponds to the change in sales for business i, y component to the deviation from the overall mean is computed and averaged across all observations (using sampling weights). 8    Socioeconomic Impacts of COVID-19 in Kenya  FIGURE 7: Average adjusted percentage change in sales 0 –10 –20 –30 –40 –39 –32 –50 –35 –47 –45 –60 –54 –51 –59 –54 –70 –61 –58 –65 –64 –80 –76 –90 ZAF KEN MDG GIN ZWE NER GAB ZMB TGO LBR TCD TZA CIV SEN Source: World Bank—Business Pulse Survey Sub-Saharan Africa Results. Note: Average adjusted mean from a linear regression that controls for country, size, sector, and timing of the survey. Computations use weights equal to the inverse of the number of observations per country and exclude countries where the fraction of missing values in the dependent variable excludes 60 percent. 3.  Impact on Employment 15. Firms have adjusted labor on the extensive (e.g., lay-offs) rather than the intensive (e.g., leave, reduced wages and hours) margin. More than 20 percent of businesses in Kenya have fired workers (Figure 8, panel a). Labor adjustments on the intensive margin have been smaller. Relatively few firms have reduced working hours of at least one employee (12 percent), reduced wages (8 percent), or granted leave of absence with or without pay (5 and 11 per- cent, respectively). A small number of firms even increased employment after the start of the COVID-19 pandemic. Given the large decreases in sales, labor adjustments have been relatively modest. Yet, the larger adjustment in the extensive margin puts Kenya in a pattern that is different than observed in other countries (see Apedo-Amah et al. 2020 for comparison). 16.  The response through labor adjustment changed as the crisis developed. When splitting the labor adjustments through different months, it is observed that firms in Kenya were more likely to adjust through the intensive margins in June, but this pattern changed over July and August (Figure 8, panel b), with a clear trend toward reducing the likelihood of adjusting through granting leave of absence, reducing hours, or reducing wages, in comparison to an increasing likelihood to adjust through laying off. 17.  Labor adjustments have taken place at both the extensive (e.g., layoffs) and intensive (e.g., reduced wages and hours) margins, but so far have been relatively modest given the large decreases in revenue for firms. More than 20 percent of businesses in Kenya laid off workers. Labor adjustments on the intensive margin were smaller on average; relatively few firms reduced the working hours of at least one employee (12 percent), reduced wages (8 percent), or granted a leave of absence with or without pay (5 and 11 percent, respectively) in all sectors over time (Figure A2.2). Impact of COVID-19 on Businesses in Kenya   9  FIGURE 8: Margin of adjustment in employment by month a. Average for the full sample 25 20 Percent of firms 15 10 5 0 Hired Fired Granted leave Granted leave Reduced Reduced workers workers of absence of absence wages hours with pay worked b. Average per month 35 30 25 Percent of firms 20 15 10 5 0 Hired Fired Granted leave Granted leave Reduced Reduced workers workers of absence of absence wages hours with pay worked June July August Note: Fraction of businesses reporting at least one employee in each category; it excludes businesses that are permanently closed. 18.  Firms have resorted to different labor adjustment measures depending on their characteristics. Firms in the other service sectors have laid off 25 percent of workers and firms in tourism 20 percent of workers. Firms in manu- facturing and retail trade have laid off less than 10 percent of workers (Table 3). Controlling for observable character- istics, there is, however, no statistically significant difference between firms of different sectors with respect to laying off of workers. Exporters and established firms less often fire workers than non-exporting and younger firms. While firms overall have more often resorted to labor adjustments on the extensive margin, firms in the tourism sector and manufacturing firms have more often adjusted on the intensive margin. Compared to other sector firms, they have more often reduced hours worked, reduced wages, or, together with retail firms, granted leave of absence (without pay). Medium and large firms in turn are more often able to grant leave of absence with pay than smaller firms. Also, exporting firms are more likely to reduce hours than non-exporting firms. Firms with a larger female workforce less often lay off workers and instead have larger labor adjustments on the intensive margin (Table A3.1). 10    Socioeconomic Impacts of COVID-19 in Kenya  TABLE 3: Estimated fraction of workers affected by margin of labor adjustment (% of workers) Businesses open or temporarily closed Workers granted Workers Workers in Workers leave of with Workers businesses Workers Workers granted leave absence wages with hours permanently hired laid off of absence with pay reduced reduced closed Characteristics (%) (%) (%) (%) (%) (%) (%) Total 1 20 5 2 3 5 1 Micro (0–4) 3 21  6 3 6 14 3 Small (5–19) 2 19 14 3 7 10 1 Medium (20–99) 2 20  8 4 6  7 1 Large (100+) 0 20  2 1 1  3 0 Agriculture/mining 3 18  6 3 1  6 2 Manufacturing 0 11  6 1 3  7 2 Retail/wholesale 1 7  4 1 2  4 1 Tourism* 1 21 14 2 9 11 3 Other services 1 26  4 3 3  4 0 Nairobi 0 23  1 1 1  1 0 Other regions 1 18  7 3 3  7 1 Young (0–4) 2 23  9 2 7  6 0 Maturing (5–14) 1 21  6 3 4  5 1 Established (15+) 1 19  4 2 2  4 0 Exporter 0 16 15 1 2 16 0 Non-exporter 1 20  4 2 3  4 1 Below 50% female 1 22  4 2 2  3 0 employment Above 50% female 2  9 11 4 5 13 1 employment Note: To estimate the fraction of workers affected, the total number of paid workers is computed adding the number of full-time paid workers and half the number of part-time workers. The total estimated number of workers in each category are presented in Table A2.3 in Appendix 2. *Tourism refers to accommodation and food services activities. 4.  Main Channels of Transmission of the Impact 19.  More than 50 percent of firms reported decreases in demand, cash flow, available finance, hours worked, and available inputs. Fifty-four percent of firms reported being affected by a decrease in the availability of inputs, which is the least often reported transmission channel. Around 65 percent of firms reported decreases in demand, cash flow, and available finance, while 62 percent of firms lamented a decrease in hours worked (Figure 9). The dif- ferent transmission channels affect different types of firms, similarly, when controlling for observable characteristics. Medium-sized firms are less often affected by any of the transmission channels besides a decrease in working hours. ­ The pandemic hit firms in other services and tourism, as well as firms with a larger workforce, mostly through changes Impact of COVID-19 on Businesses in Kenya   11 in working hours. Even though these firms experienced a larger impact, they are less likely to report any of the other shock transmission channels. Exporting firms are more often affected by a decrease in the availability of inputs, com- pared to non-exporting firms (Table A3.2).  FIGURE 9: Fraction of firms affected by transmission channels 80 60 Percent of firms 40 20 0 Decrease in Decrease in Decrease in Decrease in Decrease in hours worked demand cash flow available availability finance of inputs 20.  The different shock transmission channels of the pandemic have differentiated impacts on sales across sec- tors. When controlling for observable firm characteristics, a significant effect of the reduction in cash flow on sales becomes evident in the manufacturing sector. Therefore, access to finance might be vital for the survival of manufac- turing firms. Decreases in demand have a larger impact on sales of retail firms (Figure 10).  FIGURE 10: Estimated correlation between the change in sales and the shocks of COVID-19 Manufacturing Retail/wholesale 30 40 20 30 10 20 0 10 –10 Percent Percent 0 –20 –10 –30 –20 –40 –50 –30 –60 –40 –70 –50 Decrease Decrease Decrease Decrease Decrease Decrease Decrease Decrease Decrease Decrease in hours in demand in cash in in in hours in demand in cash in in worked flow available availability worked flow available availability finance of inputs finance of inputs Note: Estimated coefficient from a regression of the percentage change in sales on dummies for the shocks from COVID-19. In each panel, the regression considers only businesses in the sector. See Table A3.3 in Appendix 3 for a full set of results. 12    Socioeconomic Impacts of COVID-19 in Kenya 5.  Firm Survival Expectations 21.  The median firm is able to remain open for five months under the current circumstances. On average, a firm in Kenya can remain open 18 weeks. The median firm can remain open for 20 weeks (Figure 11). Medium-sized, large, and more matured firms report a higher number of weeks they can remain open under the current circumstances. Larger reserves or better access to credit could make them more resistant than smaller firms. Firms that are only partially open report being able to continue business for only 14 more weeks in the current circumstances. Sectors which faced a larger decline in sales do not exhibit a shorter period of time of survival. For instance, firms in accom- modation and food services don’t report fewer weeks of being able to remain open than firms from other sectors (Figure A2.3). There are no statistically significant differences between firms based on exporting status, region, or female employment.  FIGURE 11: Number of weeks that businesses can remain open in current circumstances 25 20 Number of weeks 15 10 5 0 Average firms Median firms Note: Open and partially open businesses only. 22.  The median firm can continue to cover costs with available cash for about four weeks. There is a large vari- ation between firms regarding the length of survival, with average firms being able to continue to cover costs for 47 days on average, but the median firm for only 30 days (Figure 12). Large firms can cover costs for a much longer period than smaller firms. Older firms and establishments in Nairobi are able to cover costs slightly longer than newer firms. Vulnerable firms can continue to cover costs for less than half as long as open firms. Firms in the accommoda- tion, food, and education sectors can cover costs for the shortest time period (Figure A2.4).  FIGURE 12: Number of days a business can cover costs with available cash 50 40 Number of days 30 20 10 0 Average firms Median firms Impact of COVID-19 on Businesses in Kenya   13 Expectations about the Future and Uncertainty 23.  Firms in Kenya expect sales to continue to massively decline in the next six months. In a regular scenario, Kenyan firms expect sales to decline by 32 percent in the next six months as compared to the previous year. In a more pessimistic scenario, firms anticipate sales to decline by 56 percent which is even more than sales have declined to date (Figure 13). On average, firms expect sales to decline by about 26 percent. There is relatively little variation between firms regarding their expectation about future sales, the standard deviation being less than half as large as the average expectation. The expected decline in sales in Kenya is more than three times the global average, indicating the pessimistic outlook Kenya firms have (Figure 14), with less uncertainty around these expected results compared to other countries for which the BPS data were collected. Large firms and firms in agriculture and manufac- turing are more optimistic about the future. On average, these firms and firms in the retail sector even expect sales to increase in an optimistic scenario (Figure A2.5).  FIGURE 13: Average change in sales expected for the next six months across scenarios 0 –10 –20 Percent –30 –40 –50 –60 Pessimistic scenario— Regular scenario— Optimistic scenario— sales change in the sales change in the sales change in the next six months next six months next six months  FIGURE 14: Expectations and uncertainty about sales growth for the next six months 30 20 18 12 10 Percent 0 –10 –8 –20 –30 –26 Average expectation Average standard deviation Kenya Global Note: For Figure 14, the global estimate is based on the average adjusted mean from a linear regression that controls for country, size, sector, and timing of the survey. Computations use weights equal to the inverse of the number of observations per country and exclude countries where the fraction of missing values in the dependent variable excludes 60 percent. 24.  While expected sale decreases are large, there is relatively little variation in expectation between firms. Weighting the expectation of changes in sales by the likelihood that each firm associates with each of the three sce- narios (regular, pessimistic, and optimistic) of happening, most firms expect sales to modestly decline, while few firms expect an increase in sales. Some firms, however, even expect declines to go up to 100 percent (Figure 15). Expecting 14  FIGURE 15: Distribution of expectations about  FIGURE 16: Distribution of uncertainty about growth growth in sales in sales .015 .06 .010 .04 Probability Probability .005 .02 0 0 –100 –50 0 50 100 0 20 40 60 80 Expectation about growth in sales Standard deviation of growth in sales large declines may additionally contribute to the economic downturn, as firms are less likely to invest. The standard deviation in expected sales growth is however relatively low (Figure 16). Thus, most firms have similar expectations about future sales developments. 25.  Firms anticipate employment to decline at a slightly lower rate than sales. In a regular scenario, firms on average expect employment to decrease by 20 percent in the next six months compared to the previous year. In a pessimistic scenario, firms anticipate employment to decline by 39 percent and in an optimistic scenario by 8 percent (Figure 17). On average, firms expect employment to decrease by close to 20 percent. There is relatively little uncer- tainty about employment changes (Figure 18). Similar to expectations about sales, smaller firms and firms in tourism and other services are more pessimistic about employment changes (Figure A2.6).  FIGURE 17: Average change in employment expected for the next six months across scenarios 0 –10 Percent –20 –30 –40 Pessimistic scenario— Regular scenario— Optimistic scenario— full-time staff changes full-time staff changes full-time staff changes in the next six months in the next six months in the next six months  FIGURE 18: Expectations and uncertainty about employment growth for the next six months 10 8 5 0 Percent –5 –10 –15 –20 –20 –25 Average employment expectation Average employment standard deviation Expectations about the Future and Uncertainty   15 26.  Fewer firms expect large employment decreases than large sales decreases. The distribution of expectations about growth in employment is less dispersed than for sales. Most firms expect employment to decline between 0 and 30 percentage points (Figure 19). Likewise, few firms expect larger increases in employment. Similar to sales, the standard deviation in expected employment growth is low (Figure 20).  FIGURE 19: Distribution of expectations about  FIGURE 20: Distribution of uncertainty about growth growth in employment in employment .10 .03 .08 .02 Probability .06 Probability .04 .01 .02 0 0 0 20 40 60 80 –100 –50 0 50 Standard deviation of growth in employment Expectation about growth in employment 16    Socioeconomic Impacts of COVID-19 in Kenya Responses to the Shock: Digital Adoption and Innovation 27.  In response to the COVID-19 outbreak, close to 50 percent of firms with five or more employees are starting to use, or increase the use of, digital platforms. Fewer firms are investing in software or digital equipment (13 per- cent), changing their product mix (18 percent), or increasing working from home (12 percent) than investing in digital platforms (49 percent) (Figure 21). Large- and medium-sized firms more often use digital platforms and invest in digital equipment than small-sized firms. Firms in tourism least often use digital platforms (Figure 22). Firms in Nairobi are not more likely to resort to digital solutions than firms in other regions. Exporting firms and firms in agriculture are less likely to repackage their product mix, most likely because they cannot quickly adjust their products to shifts in demand (Table A3.5). The movement toward increased digitalization may contribute to productivity gains and poten- tial positive reallocation within Kenya.  FIGURE 21: Business responses to the COVID-19 shock 60 50 49% 40 Percent of firms 30 20 18% 13% 12% 10 0 Increased use of Investment in Repackage Increased working digital platforms digital solutions product mix from home Note: Business responses do not include micro-sized firms. See Table A3.5 in Appendix 3 for regression results.7  FIGURE 22: Predictive effect of firm characteristics on responses 80 60 Percent of firms 40 20 0 Agriculture/ Manufacturing Retail/ Tourism Other Small Medium Large mining wholesale services (5–19) (20–99) (100+) Sector Size Increased use of digital platforms Investment in digital solutions Repackage product mix Note: Business responses do not include micro-sized firms. See Table A3.5 in Appendix 3 for regression results. 7  A different question was asked to micro-sized firms, which imposed restrictions for the comparability. 17 28.  Seventy-five percent of firms use digital platforms for business administration. Close to 50 percent of firms that started to use or increased their use of digital platforms did so for marketing activities and roughly 40 percent for service delivery (Figure 23). While business administration is the most reported type of usage of digital plat- forms across all firm characteristics, there is substantial variation in the way firms otherwise use digital platforms. For instance, controlling for observable characteristics, large firms are more likely to use digital platforms for supply chain management, marketing, sales, payments, or service delivery than small firms. In turn, younger firms more often make use of supply chain management, marketing, and payment methods than established firms, which more often use digital service delivery tools. There aren’t many differences between sectors; firms in other services less often use digital sales or marketing tools (Table A3.6). Firms in the tourism sector, which especially suffer from lockdown policies or capacity caps, could make use of service delivery platforms to mitigate decreases in sales.  FIGURE 23: Type of digital platform function used 80 60 Percent of firms 40 20 0 Business Production Supply chain Marketing Sales Payment Service administration planning management methods delivery Note: Share of firms among those which indicated an increased use of digital platforms. Micro-sized firms were not included. See Table A3.5 in Appendix 3 for regression results. 29.  Overall, these results could contribute to higher productivity growth and positive reallocation. Evidence from other countries suggests that firms with a higher level of technology are more productive, and digital technolo- gies associated with business administration are positively correlated with overall technology.8 The increase in the demand for digital technologies in Kenya could provide some opportunities for further initiatives to support overall improvement on firm technological capabilities. 30.  Kenya ranks highly among African countries for the probability of firms increasing their use of digital plat- forms. The average probability of a firm increasing the use of digital platforms in Kenya is 53 percent. This probability ranks Kenya third out of 11 African countries with comparable results and is much higher than countries such as Chad (5 percent), Niger (5 percent), and Tanzania and Zambia (18 percent) (Figure 24). 8  Cirera, et al. 2020. “Technology Within and Across Firms.” 18    Socioeconomic Impacts of COVID-19 in Kenya  FIGURE 24: Average adjusted probability of starting or increasing the use of digital platforms 70 58% 54% 60 53% 42% 49% 50 34% 40 Percent 32% 21% 30 18% 18% 20 5% 5% 10 0 TCD NER TZA ZMB GIN ZWE TGO SEN NGA KEN MDG ZAF Source: World Bank—Business Pulse Survey Sub-Saharan Africa Results. Note: Average adjusted probability of starting or increasing the use of digital platforms from a Probit that controls for country, size, sector, and timing of the survey. Computations use weights equal to the inverse of the number of observations per country, and exclude countries where the fraction of missing values in the dependent variable exceeds 60 percent. Responses to the Shock: Digital Adoption and Innovation   19 The Role of Policy 31. In response to the outbreak, the Government of Kenya introduced a range of containment policies. On March 15, 2020 all schools were mandated to close, public and private sector workers were directed to work from home, and social and religious gatherings were banned. Cashless transactions were encouraged, while hospitals and shopping malls were required to provide soap and water, as well as hand sanitizers. A nationwide curfew was intro- duced, followed by restaurants being restricted to takeaway services only, and bars being forced to close. Entry into Kenya was limited to citizens and residents, with quarantine required for 14 days. International flights were banned, and although they resumed on August 1 travelers were required to have a negative COVID-19 test to enter the coun- try. Movement in and out of Nairobi Metropolitan Area, Mombasa, Kilifi, Kwale, and Mandera was restricted from April until early July. As of September 2020, hotels can sell alcohol, but restaurants are mandated to close by 8 p.m. and must not sell alcohol until the end of September. Bars remain closed until further notice. 32.  The government’s immediate mitigation actions have included a range of measures focused on strengthen- ing the health system and delivering direct assistance to households. Authorities have provided in-kind assistance including soap and food aid, mainly in Nairobi’s poorest areas, complemented by assistance from the UN World Food Programme.9 Similarly, cash transfers have been delivered via mobile payments to households in low-income informal settlements in Kenya’s urban centers.10 While schools remain closed, the Kenya Ministry of Education shared guidelines for enhancing teaching and learning through four main platforms: (i) daily radio programs, (ii) education television broadcasts, (iii) KICD’s EduTV Kenya YouTube channel, and (iv) digital learning resources from the Kenya Education Cloud.11 33.  A series of relief tax measures were enacted to help lessen the immediate financial burden on Kenya’s cit- izens and businesses.12 The Tax Law (Amendment) Act 2020 went into effect on April 25. Tax measures include a reduction of the value added tax (VAT) rate from 16 percent to 14 percent; a reduction of the personal income tax top rate from 30 percent to 25 percent; a reduction of the turnover tax rate for micro, small, and medium enterprises from 3 percent to 1 percent; and 100 percent tax relief for persons earning up to KSh 24,000. In addition, the government enacted a temporary suspension of the listing of loan defaulters for any person; micro, small, and medium enter- prises; and corporate entities whose loan account is in arrears as of April 1, 2020. 34.  The Government of Kenya also implemented a series of economic stimulus measures.13 A Central Bank order for banks to waive fees for individuals who move money between their bank account and mobile wallet came into effect on March 17. The upper limit for mobile money transfers was increased. Authorities reached a deal with 9  World Food Programme, “WFP Supplements Government Support to Poor Families in Kenya Hit by COVID-19.” https://www.wfp.org/news/wfp- supplements-government-support-poor-families-kenya-hit-covid-19 10 Capital News, “250,000 Households Identified for Cash Support in the Wake of COVID-19.” https://www.capitalfm.co.ke/ news/2020/05/250000-households-identified-for-cash-support-in-the-wake-of-covid-19/ 11  World Bank. 2020c. “How Countries Are Using Edtech (Including Online Learning, Radio, Television, Texting) to Support Access to Remote Learning during the COVID-19 Pandemic.” https://www.worldbank.org/en/topic/edutech/brief/how-countries-are-using-edtech-to-support- remote-learning-during-the-covid-19-pandemic 12 KPMG. “Kenya: Government and institution measures in response to COVID-19.” https://home.kpmg/xx/en/home/insights/2020/04/kenya- government-and-institution-measures-in-response-to-covid.html 13  KPMG. “Kenya: Government and institution measures in response to COVID-19.” https://home.kpmg/xx/en/home/insights/2020/04/kenya- government-and-institution-measures-in-response-to-covid.html 20 commercial banks to restructure nonperforming loans caused by COVID-19 layoffs. Additionally, loans and grants are available through government and private funds, including the National Business Compact on CoVid19 (NBCC), as well as special loans through Stanbic Bank and Standard Chartered Bank, among others. The Government of Kenya also disbursed KSh 1 billion for the health care sector and US$5 million for the tourism sector. In addition, the Inter- national Finance Corporation disbursed a US$50 million loan to the Equity Bank Kenya to support small and medium enterprises (SMEs). 35.  One in five firms in Kenya has received public support during the COVID-19 pandemic. Firms based in Nai- robi were more likely to receive assistance than firms in other regions (Figure 25, panel a). Firms in agriculture and in health/social work sectors on average most often received assistance, though the differences are not statistically significant (Figure 25, panel b). Of the firms that got any assistance, close to 36 percent received monetary transfers and 33 percent received tax deferrals (Figure 26). Within the “other assistance” measure, firms most often state the provision of sanitizers and masks. 36.  Micro-sized and small firms were less likely to receive public support, compared to larger firms. While 27 per- cent of large firms got assistance, only 16 percent of micro-sized reported having access to assistance measures (Figure 25, panel a).  FIGURE 25: Distribution of expectations about growth in employment a. All, by size and region b. By sector Other services Nairobi Region Health/social work Other regions Education Finance/real estate Large (100+) Information/communication Food services Medium (20–99) Accommodation Size Transportation/storage Small (5–19) Retail/wholesale Micro (0–4) Construction/utilities Manufacturing All Agriculture/mining 0 10 20 30 0 10 20 30 40 50 Percent of firms Percent of firms The Role of Policy   21  FIGURE 26: Type of assistance received 40 30 Percent 20 10 0 Monetary Deferral of Deferral of Access to Loans with Fiscal Tax Wage Other transfers rent/mortgage/ credit new credit subsidized exemptions deferrals subsidies utilities payments/ interest or reduction suspension of rates interest payments Note: The other option includes provision of sanitizer, masks, and social distancing or precautionary measures. 37.  Information gaps are the main reason for not receiving public support. Among firms that did not receive any assistance, over 80 percent reported not having received assistance because they were not aware of any govern- ment measures (Figure 27). Information campaigns are thus vital to ensure that assistance reaches those that need it most. Awareness over existing government programs is similar over firm characteristics (Figure 28).  FIGURE 27: Reason for not receiving assistance  FIGURE 28: Not being aware of programs by firm characteristic Other services Other 1.9% Tourism Sector Applied but Retail/wholesale not received Not 6.0% aware Manufacturing 81.5% Agriculture/mining Not elegible 3.9% Large (100+) Too difficult Medium (20–99) Size to apply 6.8% Small (5–19) Micro (0–4) 0 20 40 60 80 100 Percent 22    Socioeconomic Impacts of COVID-19 in Kenya 38.  Loans with subsidized interest rates are the most needed policy response according to Kenyan firms. Forty-­ two percent of firms in Kenya call for monetary transfers from the government and 25 percent for tax deferrals, while 50 percent of firms refer to loans with subsidized interest rates as one of the three most needed policies (Figure 29). The type of most-needed assistance, however, varies between different firm characteristics. For instance, exporters disproportionately demand tax deferrals (Table A3.7). This could either reflect higher tax and custom duties or the fact that exporting firms do not face liquidity constraints or lack of credit, and therefore do not call for other policy measures. Agricultural firms and firms in retail are most likely to call for monetary transfers. Firms in the tourism and manufacturing sectors are more likely to call for loans with subsidized rates, indicating liquidity constraints.  FIGURE 29: Self-reported most needed public policies to support businesses 60 40 Percent 20 0 Monetary Deferral Deferral of Access to Loans with Fiscal Tax Wage Other transfers of rent/ credit new subsidized exemptions deferrals subsidies mortgage/ payments/ credit interest or reduction utilities suspension rates of interest payments 39.  Firm size plays an important role with regard to the type of assistance that firms require. Larger firms more often report fiscal exemptions and deferral of loan payments as the most needed policy responses, as these likely have larger duties and open loans. As by definition micro-sized firms have only few or no employees, wage subsidies are the least often reported and most needed policy. In turn, micro-sized and small firms more often call for deferral of rent, mortgages, and utilities, since these make up a large proportion of these firms’ running costs (Table A3.7). 40.  Firms call for different policies depending on the type of shock they experienced. While the majority of firms report loans with subsidized rates to be among the three most needed policies, the type of shock a firm experienced has an effect on the policies that it calls for. Firms hit by a decrease in demand are around 24 percentage points more likely to demand cash transfers (Figure A2.7). In turn, firms that report a decrease in hours worked, i.e., that are more liquidity constrained, more likely call for subsidized loans or tax deferrals. The Role of Policy   23 Policy Recommendations 1.  General Recommendations 41.  To help mitigate the adverse impacts of COVID-19 on firms, this report suggests policy response options divided into four areas: ensure the liquidity of viable firms, enhance firm capabilities, promote access to new markets, and reduce uncertainty by improving access to new information. As the crisis continues to evolve, poli- cies must find a balance between short-term interventions to help businesses “keep the lights on” and a sustainable recovery plan that facilitates the selection of the most productive firms. In the recovery phase, policies should be geared toward supporting growth-oriented enterprises, promoting the reallocation of resources to more efficient companies, and avoiding measures that risk propping up zombie firms (i.e., inefficient firms that survive).14 42.  Ensure the liquidity of viable firms: A key priority in the short term is to alleviate the restriction of cash flows due to lower demand. Direct measures to address liquidity pressures may encompass accelerated depreciation on some or all categories of assets, which would reduce taxable income.15 (i) FinTech solutions should be promoted. Kenya is known for its innovative solutions regarding digital financial services. Digital technology offers an unprecedented opportunity to mitigate the impact of the COVID-19 crisis on micro, small, and medium enterprises (MSME) financing. Simplified loan application processes and the use of alternative data for credit decisions could be leveraged by banks to reduce turnaround times for MSME loans. (ii) The efficacy of the emergency tax reduction and deferral measures should be assessed. The Government of Kenya has implemented a package of tax measures, which includes reduction of the base corporate income tax rate from 30 to 25 percent, reduction of the turnover tax rate on small business, from 3 to 1 percent, and a reduction of the standard VAT rate from 16 to 14 percent.16 Assessing the impact of these measures on the robustness of firms, as well as on tax revenue is critical, so that evidence-based decisions can be made about whether they should continue. (iii) Conditions must be established to prevent the insolvency of healthy firms due to temporary illiquidity. For micro and small businesses, this could mean increasing the debt threshold required for a creditor to initiate bankruptcy proceedings against a debtor or limiting access in modern personal bankruptcy systems to a debt- or’s petitions alone. Enacting these measures for a fixed time period would prevent the system from becoming one of debt collection during a pandemic, as well as help control the number of cases entering the overbur- dened court system. The Central Bank suspended, for six months, the listing of negative credit information for 14  This section is based on the overall policy guidance described in World Bank 2020a, and Cirera et al. (2021). “Assessing the impact and policy responses in support of private-sector firms in the context of the COVID-19 pandemic.” While the COV-BPS findings indicate some gaps and opportunities for policy actions, these recommendations come from a menu of options based on a broader World Bank knowledge of Kenya and other engagements. 15 These recommendations related to liquidity of viable firms are based on information beyond COV-BPS results. See World Bank (2020b) for further details. A second wave of the COV-BPS in Kenya will focus further on issues related to liquidity and solvency to provide additional evidence for this debate. 16  IMF policy tracker. https://www.imf.org/en/Topics/imf-and-covid19/Policy-Responses-to-COVID-19 24 borrowers whose loans became nonperforming after April 1, 2020.17 These measures need to be reassessed and potentially expanded based on the extent and duration of the COVID-19 crisis, while keeping in mind the risks they present for financial sector sustainability. (iv) De-risking financial institutions will be important for increasing access to finance for healthy firms. Risk aversion is an important factor limiting the willingness of financial intermediaries to increase lending, particularly to MSMEs. The National Treasury is setting up a credit guarantee scheme to issue partial credit for commercial bank loans to MSMEs. (v) The liquidity constraints of microenterprises should be alleviated. Providing liquidity channeled through microfinance institutions, Savings and Credit Cooperative Organisations (SACCOs), and digital platforms can help address the liquidity constraints faced by these institutions and their ability to extend credit to micro and small firms. However, any initiative on this front needs to be combined with rigorous risk management processes. (vi) Government arrears on payments to MSMEs must be addressed.18 This can be accomplished by setting up a receivables financing platform that would allow financial institutions to refinance these receivables through an invoice and receivables discounting scheme. To give comfort to financial institutions, the scheme will be sup- ported by the guarantee product. (vii) Early-stage companies should not be left out of safety net provisions. Public policies to help vulnerable but viable firms stay in business and maintain employment should also include start-ups. The provision of a cash lump sum for firms to stay afloat could help overcome the immediate challenges brought on by the pandemic. Keeping this sum reasonably small would make it feasible from a fiscal perspective while ensuring that it is still relevant for start-ups. If employment retention is crucial to keep the business alive, then an immediate cash injection, either through grant or loan or guarantee, could be explored. If market failures are clearly iden- tified, support for publicly funded venture capital companies and funds to inject equity could be explored.19 Loan or equity injections into venture funds can help them survive through a period when they cannot realize any returns, and can ease the pressure on them to liquidate companies in which they have invested in the short term. 43.  Enhance firm capabilities: The pressure to react to the crisis may offer an opportunity to improve overall mana- gerial and technological capabilities throughout firms in Kenya. The COV-BPS results suggest that firms are respond- ing to the crisis with the adoption of digital technologies, which can be useful for improving their overall capabilities. (viii) Using this crisis as an opportunity to accelerate digital technologies can help increase firm efficiency. Evi- dence across countries suggests that a large proportion of firms are starting to use or increase the use of digital technologies for business purposes. In the case of Kenya the response has been relatively larger than in other developing countries, but there is still a potential for expansion, particularly among small firms. Facilitating the adoption of digital technologies that can be applied to general business functions, such as business planning, marketing, payment, and sales, will be critical for helping firms cope with the COVID-19 crisis and for improving 17  IMFpolicy tracker. https://www.imf.org/en/Topics/imf-and-covid19/Policy-Responses-to-COVID-19 18  The government was in significant arrears to suppliers and contractors, estimated at 0.7 percent of GDP (~KSh 65 billion) in FY2018/19. Paying arrears will be critical to enhance firms’ liquidity during the crisis. The GoK has already taken steps and allocated KSh 13.8 billion to clear arrears and KSh 10 billion for VAT refunds as part of its policy responses to the pandemic (World Bank Kenya Economic Update. April 2020). 19  For example, France and Germany have a long tradition of using these instruments through state development banks to provide risk capital to MSMEs. Policy Recommendations   25 their capabilities going forward. Among the functions with a higher potential for easy adoption are technologies related to supply chain management and sales. (ix) Providing business development services, such as general business training, specific technical training, and management advice, could enhance firm resilience. International metrics suggest that Kenyan firms are lagging behind with respect to the adoption of good managerial practices, which is strongly associated with firm performance measures such as productivity and exports.20 Evidence on business training focusing on improving business practices for SMEs across countries suggests an average impact of 10 percent on profits.21 Indeed, previous experiments with micro and small firms in Kenya suggest that interventions to improve busi- ness practices through mentorship can lead to an increase in profits of 20 percent on average.22 While firms are facing significant challenges associated with a reduction in demand and a shortage of cash, the need for innovative business solutions inherent to such crises can be used as an opportunity to better prepare firms for the recovery process. Ongoing interventions already aimed at supporting SMEs, such as the Kenya Industry and Entrepreneurship Project, should be accelerated and scaled up. Digital solutions can be leveraged here too, through online tools for delivering support for managerial capabilities at a lower cost. 44.  Promote access to new markets: The development of the COVID-19 crisis has caused disruption in several value chains, which may not only create challenges but also opportunities for Kenyan firms. (x) Providing information to firms can support them in prospecting new markets. The variation in the devel- opment of the COVID-19 crisis in different countries can generate significant variations in how global value chains are disrupted, which can in turn create business opportunities.23 Providing information to local produc- ers regarding opportunities in international markets, particularly in exporting sectors such as agricultural com- modities (e.g., coffee, tea, fruits), processed food, and apparel, could help boost export potential during a period of global crisis. Such activities could be conducted by the Kenya Export Promotion Agency. 45.  Improve access to information: Evidence for Kenya and other countries suggests that firms are expecting large declines in sales in the coming six months, with a high degree of uncertainty. Improving transparency and access to information about support currently available for businesses can increase the likelihood of reaching the firms most in need and could help improve expectations overall. (xi) Providing guidance on health protocols could help reduce risks. Widely disseminating information on proto- cols to minimize the risk of transmission of COVID-19 among customers could increase confidence and busi- ness activity. This could also help reduce the risk of outbreaks within a business, a situation that could seriously exacerbate the operational challenges already being faced by firms. In tandem with such information cam- paigns, some financial support to help firms adopt the required sanitary measures could improve compliance. This topic will be analyzed in the next report, which will include questions on adoption of health protocols. (xii) Stronger communication is needed about policy interventions already available to support businesses. A common challenge across many developing countries, including Kenya, is that a very large share of businesses are not aware of the public programs available to support them. In Kenya, about 80 percent of businesses that 20  See evidence across countries, including Kenya, from the World Management Survey. https://worldmanagementsurvey.org/ 21  McKenzie, D. 2020. “Small Business Training to Improve Management Practices in Developing Countries.” Policy Research Working Paper, 9408. World Bank. 22 Brooks, Donovan, and Johnson. 2018. “Mentors or Teachers? Microenterprise Training in Kenya.” American Economic Journal: Applied Economics; Beaman, Lori, Jeremy Magruder, and Jonathan Robinson. 2014. “Minding small change among small firms in Kenya,” Journal of Development Economics 108: 69–86. 23  https://www.wsj.com/articles/high-food-prices-drive-consumers-to-hunt-for-value-11591700401 26    Socioeconomic Impacts of COVID-19 in Kenya did not receive support reported they were not aware of the options available to them. Evidence from a similar survey across countries suggests that firms that are more likely to receive assistance also have better expecta- tions regarding the future of their business. (xiii) Targeted channels should be used to reach different types of firms with information regarding government programs. Kenya has more than 20 national programs in place to support entrepreneurship activities. An ongo- ing assessment conducted by the World Bank suggests that many of those programs provide services related to access to finance. However, information about those policy instruments is not easily available. The Govern- ment of Kenya could consolidate information about all public programs to support businesses, including the expansion of activities specifically related to COVID-19, and facilitate access to this information for businesses. This could be converted into a sustained practice as a way to optimize public resources. 2.  Recommendations on Targeting Firms 46.  The results of the COV-BPS highlight that one of the main challenges faced in implementing business sup- port during COVID-19 is to identify which firms should be targeted. MSMEs seem to be the most vulnerable in general. They are disproportionally more impacted in terms of reductions in sales and they face a higher likelihood that they will close, partially because they tend to have less access to credit. At the same time, the number of layoffs is significantly higher among large firms. Given the ubiquity of the shock across the whole economy, the challenge of targeting specific groups of firms is not very different than it was prior to COVID-19. The cautions that were appli- cable before the pandemic regarding targeting criteria for interventions to support businesses are still valid—the heterogeneity between firms belonging to the same sector and of similar size must be considered. There are a large ­ number of options to improve the targeting of firms. 47.  The challenge of targeting firms: Results of the COV-BPS suggest the impact of COVID-19 on businesses in Kenya is widespread across firms of different size, sector, region, and age. This creates significant challenges for policy makers when defining a specific group of firms to target. (i) A “funnel approach” to assistance can help ensure that the firms with the greatest potential for improve- ment get the most support. This approach might be particularly relevant for interventions that aim to provide business training and financing support through grants, but who are also looking to identify businesses with a high level of commitment and need. The program can provide very basic assistance services for a large number of firms (e.g., some online courses, simple benchmarking information, or short one-hour firm visits). Firms that demonstrate interest and undertake some improvement actions following this first engagement can then be filtered into receiving a second, more intermediate level of business training support or a specific grant. This approach has the political advantage of offering some assistance to a large number of firms while restricting the most costly and time-consuming parts of the program to firms that demonstrate engagement and immediate improvement.24 (ii) Mobile phones can be used to reach women-owned businesses. MSMEs run by women can be disproportion- ally affected by the COVID-19 shock, and specific targeted interventions already conducted in Kenya suggest that they can lead to effective results.25 Policies could include: (i) providing mobile phones to women to facilitate 24  See more details on McKenzie, D. 2020. “Small Business Training to Improve Management Practices in Developing Countries.” Policy Research Working Paper, 9408. World Bank. 25  McKenzie and Puerto. 2017. “Growing Markets through Business Training for Female Entrepreneurs: A Market-Level Randomized Experiment in Kenya,” American Economic Journal: Applied Economics. Policy Recommendations   27 access to financing, and (ii) using customer data on mobile phone and mobile banking transactions to identify women more vulnerable during this crisis to more effectively target relief payments. (iii) The targeting of solutions can leverage big data analytics from digital platforms. Mobile Network Operator (MNO) data that capture financial transactions, such as credit, remittance, and payment data for firms, can be especially useful. Some of the simple metrics to identify these enterprises could be: (i) reduced volume and number of mobile money transactions, (ii) increased uptake of overdraft facilities in the last few months, and (iii) vulnerable informal MSMEs in hard-hit sectors by the COVID-19 pandemic, such as the retail, agribusiness, and manufacturing sectors. (iv) Complementary measures should be considered to offer support for solvency problems among SMEs or strategically large firms.26 Given the extent of the COVID-19 crisis, providing liquidity may be an insufficient remedy, as liquidity does not compensate businesses for their losses. Should the crisis threaten the solvency of MSMEs, governments would have to consider additional measures to complement the emergency actions discussed above. Some options may include: direct compensation through grants for viable firms/sectors that have been significantly impacted;27 support for publicly funded venture capital companies and funds to inject equity if market failures are clearly identified;28 indirect support through loss-sharing mechanisms and other forms of leverage funding; and stimulation of private equity investment.29 The implementation of any of these options should address specific market failures, be reassessed regularly, and remain temporary in nature. These schemes can be controversial if they lead to large-scale nationalizations and can be expensive in terms of fiscal resources. Therefore, they would have to be designed in a transparent way with clear sunset clauses and exit strategies. 26   More details are provided by World Bank. 2020. “Assessing the impact and policy responses in support of private-sector firms in the context of the COVID-19 pandemic.” 27  For example, the European Commission indicated that direct compensation for damages suffered due to the COVID-19 outbreak for companies active in sectors that have been particularly hit (e.g., transport, tourism, and hospitality or organizers of cancelled events) would be authorized even though they are state aids, which are typically prohibited in the EU. 28  For example, France and Germany have a long tradition of using these instruments through state development banks to provide risk capital to MSMEs. 29  As with lending, guarantees have the potential to provide large-scale effects and subsidy leverage if they are well designed and implemented. The U.S. Small Business Administration’s leverage program for Small Business Investment Companies has a long history and provides a number of lessons. 28    Socioeconomic Impacts of COVID-19 in Kenya References Altig, David, Jose Maria Barrero, Nicholas Bloom, Steven J. Davis, Brent H. Meyer, and Nicholas Parker. 2019. “Survey- ing business uncertainty.” No. W25956. National Bureau of Economic Research. Apedo-Amah, Marie Christine; Avdiu, Besart; Cirera, Xavier; Cruz, Marcio; Davies, Elwyn; Grover, Arti; Iacovone, Leon- ardo; Kilinc, Umut; Medvedev, Denis; Maduko, Franklin Okechukwu; Poupakis, Stavros; Torres, Jesica; and Tran, Trang Thu. 2020. “Unmasking the Impact of COVID-19 on Businesses: Firm Level Evidence from Across the World.” Policy Research Working Paper n. 9434. World Bank, Washington, DC. Beaman, Lori, Jeremy Magruder, and Jonathan Robinson. 2014. “Minding small change among small firms in Kenya,” Journal of Development Economics 108: 69–86. Brooks, Wyatt, Kevin Donovan, and Terence R. Johnson. 2018. “Mentors or teachers? Microenterprise training in Kenya.” American Economic Journal: Applied Economics 10, no. 4: 196–221. Cirera, X., Comin, D., Cruz, M, and Lee, K. M. 2020. “Technology Within and Across Firms.” NBER Working Paper, n. 28080. Cirera, Xavier; Cruz, Marcio; Davies, Elwyn; Grover, Arti; lacovone, Leonardo; Lopez Cordova, Jose Ernesto; Medve- dev, Denis; Okechukwu Maduko, Franklin; Nayyar, Gaurav; Reyes Ortega, Santiago; Torres, Jesica. 2021. Policies to Support Businesses through the COVID-19 Shock: A Firm-Level Perspective. Policy Research Working Paper, No. 9506. World Bank, Washington, DC. McKenzie, D., and Puerto, S. 2017. “Growing markets through business training for female entrepreneurs: a market-­ level randomized experiment in Kenya.” The World Bank. World Bank. 2020a. “Africa’s Pulse: Assessing the Economic Impact of COVID-19 and Policy Responses in Sub-­ Saharan Africa” Washington, DC. World Bank. 2020b. “Assessing the impact and policy responses in support of private-sector firms in the context of the COVID-19 pandemic.” Finance Competitiveness and Innovation Global Practice. Washington, DC. World Bank. 2020c. “How Countries Are Using Edtech (Including Online Learning, Radio, Television, Texting) to Sup- port Access to Remote Learning during the COVID-19 Pandemic.” Washington, DC. 29 Appendix 1. Description of the Sample The Kenya-BPS sample consists of 2,070 firms randomly selected from the universe of 138,186 firms available in the 2017 Census of Establishments from the Kenyan National Bureau of Statistics (KNBS). Table A1.1 shows the distribution of the sample based on the analysis.30 The sample was stratified by size groups (micro: 0–3, small: 4–9; medium-small: 10–49, medium-large: 50–149, and large: 150+ employees) and sector (agriculture, food processing; wearing apparel, other industry; construction and real estate; wholesale and retail trade; repair of motor vehicles and motorcycles; transportation and storage; accommodation; food service activities; information and communication technology (ICT), finance, professional, and administrative services; and education, health and other services).31 The analysis regrouped the size and sectors of activities for the sake of com- parability across countries, as the COV-BPS has been implemented and harmonized across more than 50 countries.32 Phone interviews were conducted between June 10 and August 30, 2020. The team used Computer Assisted Telephone Interviewing (CATI) through a Survey Solutions platform. The response rate was 37 percent. A total of 5,567 firms were reached or attempted to be reached by the survey team. Among those, 1,827 firms were missed after all attempts. Other firms were missed by being unable to participate, declining consent, or unobtainable contact. The sample based on the universe population for each stratum was provided by KNBS from the Census of Establish- ment data (2017). To account for nonresponses and to ensure the weighted sample matches the universe population in each strata, the unadjusted weights were multiplied by an adjustment factor and calculated for each strata by dividing the target sample total by the actual number of firms successfully sampled, which was used in the analysis.  TABLE A1.1: Number of firms by sector, size, region, and exporting status Agriculture/ Retail/ Other mining Manufacturing wholesale Tourism services Total All 142 388 301 298 941 2,070 Micro (0–4) 19 188 154 109 310 780 Small (5–19) 36 91 88 106 313 634 Medium (20–99) 44 57 42 66 235 444 Large (100+) 43 52 17 17 83 212 Young (0–4) 18 42 43 48 92 243 Maturing (5–14) 57 161 172 170 546 1,106 Established (15+) 67 185 86 80 303 721 Other regions 121 307 233 275 699 1,635 Nairobi 21 81 68 23 242 435 Below 50% female employment 101 332 256 184 755 1,628 Above 50% female employment 41 56 45 114 186 442 30  Tourism-related activities refer to the accommodation and food and beverage service activities. 31  The sample was stratified by the groupings listed. In order to match standard firm size classifications, they have been aggregated in Table A1.1. 32  Results for cross-country comparability are available at Apedo-Amah et al. (2020). 30 Appendix 2. Additional Results  FIGURE A2.1: Changes in mobility over time (percent change compared to baseline) Nairobi 60 40 20 0 –20 –40 –60 –80 Feb Feb Mar Mar Mar Apr Apr Apr May May May Jun Jun Jun Jul Jul Jul Aug Aug Aug Sep Sep Sep Oct 15 25 06 16 26 05 15 25 05 15 25 04 14 24 04 14 24 03 13 23 02 12 22 02 Retail/recreation Grocery/pharmacy Workplace Residential Mombasa 40 20 0 –20 –40 –60 –80 Feb Feb Mar Mar Mar Apr Apr Apr May May May Jun Jun Jun Jul Jul Jul Aug Aug Aug Sep Sep Sep Oct 15 25 06 16 26 05 15 25 05 15 25 04 14 24 04 14 24 03 13 23 02 12 22 02 Retail/recreation Grocery/pharmacy Workplace Residential 31 Percent Percent 32  0 5 10 15 20 25 30 35 40 45 0 5 10 15 20 25 30 35   Hired workers Hired workers Fired workers Fired workers June June Granted leave Granted leave of absence of absence Percent July Agriculture July 0 5 10 15 20 25 30 35 40 Granted leave of Granted leave of Retail/wholesale absence with pay absence with pay Hired workers August August Reduced wages Reduced wages Fired workers Reduced Reduced hours worked June hours worked Socioeconomic Impacts of COVID-19 in Kenya Granted leave of absence Percent Percent July 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 Granted leave of Other services absence with pay Hired workers Hired workers August Reduced wages  FIGURE A2.2: Margin of adjustment in employment by month and sector Fired workers Fired workers Reduced June June hours worked Granted leave Granted leave of absence of absence Tourism July July Granted leave of Granted leave of Manufacturing absence with pay absence with pay August August Reduced wages Note: Fraction of businesses reporting at least one employee in each category; excludes businesses that are permanently closed. Reduced wages Reduced Reduced hours worked hours worked  FIGURE A2.3: Predictive effect of size and sector on weeks a business can remain open a. By size, age, and status b. By sector 30 Other services Health/social work 20 Education Finance/real estate 10 Information/communication Food services 0 Micro (0–4) Small (5–19) Medium (20–99) Large (100+) Young (0–4) Maturing (5–14) Established (15+) Open Partially open (mandated by government) Accommodation Transportation/storage Retail/wholesale Construction/utilities Manufacturing Agriculture/mining Size Age Status 0 10 20 30 Note: The figure includes only open and partially open businesses. Marginal effect from a linear regression on dummies for size, formality status, sector, region, age, exporting status, and female employment dummy. Only groups with statistically significant differences are shown. The number of weeks a firm can remain open are winsorized at the 99 percent level, to account for outliers. See Table A3.4 for full set of results.  FIGURE A2.4: Predictive effect of size and sector on days a business can cover costs with available cash a. By size, age, region, and status b. By sector 120 Other services Health/social work 80 Education Finance/real estate 40 Information/communication Food services Accommodation 0 Micro (0–4) Small (5–19) Medium (20–99) Large (100+) Young (0–4) Maturing (5–14) Established (15+) Other regions Nairobi Open Vulnerable Transportation/storage Retail/wholesale Construction/utilities Manufacturing Agriculture/mining Size Age Region Status 0 20 40 60 80 Note: Vulnerable firms are those that are partially open or temporarily closed. Marginal effect from a linear regression on dummies for size, formality status, sector, region, age, exporting status, and female employment dummy. Only groups with statistically significant differences are shown. The number of days a firm covers costs are winsorized at the 99 percent level, to account for outliers. See Table A3.4 for full set of results. Appendix 2. Additional Results   33  FIGURE A2.5: Average change in sales expected for the next six months by sector and size 20 10 0 –10 –20 Percent –30 –40 –50 –60 –70 Small Medium Large Agriculture/ Manufacturing Retail/ Tourism Other (5–19) (20–99) (100+) mining wholesale services Size Sector Pessimistic scenario—sales change in the next six months Regular scenario—sales change in the next six months Optimistic scenario—sales changes in the next six months  FIGURE A2.6: Average change in employment expected for the next six months across scenarios 10 0 –10 Percent –20 –30 –40 –50 Small Medium Large Agriculture/ Manufacturing Retail/ Tourism Other (5–19) (20–99) (100+) mining wholesale services Size Sector Pessimistic scenario—full-time staff changes in the next six months Regular scenario—full-time staff changes in the next six months Optimistic scenario—full-time staff changes in the next six months 34    Socioeconomic Impacts of COVID-19 in Kenya  FIGURE A2.7: Predictive effect of shocks on top three most needed public policies Percent of change in probability 30 15 0 –15 –30 –45 Decrease in hours worked Decrease in demand Decrease in cash flow Decrease in available finance Decrease in availability of inputs Decrease in hours worked Decrease in demand Decrease in cash flow Decrease in available finance Decrease in availability of inputs Decrease in hours worked Decrease in demand Decrease in cash flow Decrease in available finance Decrease in availability of inputs Cash transfer Subsidized loans Tax deferral Note: Marginal effects from a probit regression of dummies whether the business demands each specific policy on dummies for each shock type, controlling for observable characteristics (size, sector, region, age, and exporting status). See Table A3.8 for full set of results.  TABLE A2.1: Expected time to resume operations (fraction of temporarily closed business) More Less than 2–4 1–2 2–6 than Don’t 2 weeks weeks months months 6 months know   (%) (%) (%) (%) (%) (%) Total 0 1 0 7 10 82 Micro (0–4) 1 1 0 3 7 88 Small (5–19) 0 1 0 9 10 80 Medium (20–99) 0 0 0 14 18 68 Large (100+) 0 0 0 1 0 99 Agriculture/mining 0 0 0 1 0 99 Manufacturing 0 0 0 3 2 95 Retail/wholesale 0 1 0 2 5 92 Tourism 1 1 2 7 8 81 Other services 0 0 0 12 15 72 Nairobi 1 0 0 2 6 91 Other regions 0 1 0 8 11 79 Young (0–4) 0 0 1 8 7 84 Maturing (5–14) 0 1 0 6 11 81 Established (15+) 0 0 0 8 10 82 Exporter 0 1 2 0 0 97 Non-exporter 0 1 0 7 10 82 Below 50% female 0 0 0 6 8 85 employment Above 50% female 1 2 1 12 17 68 employment Appendix 2. Additional Results   35  TABLE A2.2: Weeks that business can remain open in current circumstances   Average firm Median firm Total 18 20 Micro (0–4) 17 20 Small (5–19) 18 20 Medium (20–99) 21 20 Large (100+) 21 20 Agriculture/mining 20 20 Manufacturing 22 20 Retail/wholesale 17 18 Tourism 19 20 Other services 19 20 Other regions 18 20 Nairobi 19 20 Young (0–4) 15 19 Maturing (5–14) 18 20 Established (15+) 19 20 Non-exporter 18 20 Exporter 22 21 Below 50% female employment 18 20 Above 50% female employment 19 20 36    Socioeconomic Impacts of COVID-19 in Kenya  TABLE A2.3: Estimated number of workers affected by margin of adjustment Businesses open or temporarily closed Workers Workers granted Workers Workers Workers in granted leave of with with businesses   Workers Workers leave of absence wages hours permanently Characteristics hired laid off absence with pay reduced reduced closed Total 29,399 617,003 152,016 72,066 82,006 145,964 19,326 Micro (0–4) 2,492 17,411 5,280 2,427 4,896 12,010 2,304 Small (5–19) 6,929 63,907 46,888 9,737 25,012 32,896 3,066 Medium (20–99) 12,944 147,174 56,598 30,739 41,556 51,541 5,323 Large (100+) 7,034 388,511 43,249 29,164 10,543 49,516 8,633 Agriculture/mining 8,809 60,600 21,101 8,962 3,687 19,063 5,200 Manufacturing 1,015 27,312 14,264 1,545 6,564 19,275 4,839 Retail/wholesale 4,316 40,961 25,331 3,165 11,704 22,399 5,072 Tourism 985 15,199 10,436 1,796 6,675 7,997 1,881 Other services 14,273 472,930 80,883 56,598 53,376 77,230 2,334 Nairobi 4,593 278,039 16,695 10,931 17,832 16,175 2,252 Other regions 24,806 338,964 135,320 61,135 64,175 129,789 17,074 Young (0–4) 2,596 25,183 10,000 2,630 7,589 7,135 4,45 Maturing (5–14) 16,017 242,407 70,803 32,871 42,083 60,700 15,227 Established (15+) 10,786 349,413 71,213 36,564 32,335 78,128 3,654 Exporter 117 20,966 19,369 1,028 3,107 21,722 0 Non-exporter 29,282 596,037 132,647 71,038 78,899 124,242 19,326 Below 50% female 22,330 575,742 103,942 52,495 57,098 87,999 12,659 employment Above 50% female 7,069 41,261 48,074 19,571 24,908 57,964 6,666 employment Appendix 2. Additional Results   37 Appendix 3. Results from OLS and Probit Regressions  TABLE A3.1: Estimated correlation between employment adjustments and business characteristics (1) (2) (3) (4) (5) (6) Probit Probit Probit Probit Probit Probit Granted Granted leave of Reduced Hired Fired leave of absence Reduced hours workers workers absence with pay wages worked Small (5–19) 0.157 0.289*** 0.682*** 0.197 0.427*** 0.188 (0.19) (0.11) (0.15) (0.19) (0.14) (0.12) Medium (20–99) 0.248 0.325*** 0.733*** 0.450** 0.560*** 0.187 (0.19) (0.12) (0.17) (0.18) (0.16) (0.13) Large (100+) 0.313 0.547*** 0.540** 0.552** –0.095 –0.114 (0.23) (0.20) (0.23) (0.27) (0.20) (0.25) Manufacturing –0.116 0.144 0.455** –0.029 0.542** 0.588*** (0.25) (0.16) (0.21) (0.25) (0.24) (0.20) Retail/wholesale –0.249 0.133 0.494** 0.039 0.229 0.250 (0.24) (0.18) (0.22) (0.26) (0.24) (0.22) Tourism –0.206 0.203 0.581*** –0.037 0.581** 0.577*** (0.25) (0.17) (0.21) (0.26) (0.23) (0.21) Other services –0.393* 0.065 0.199 0.040 0.268 0.164 (0.22) (0.15) (0.19) (0.24) (0.22) (0.20) Nairobi –0.195 –0.076 –0.431*** –0.209 –0.174 –0.183 (0.17) (0.11) (0.16) (0.18) (0.16) (0.13) Maturing (5–14) –0.016 –0.190 –0.213 –0.281 –0.033 0.071 (0.27) (0.15) (0.18) (0.25) (0.18) (0.16) Established (15+) –0.412 –0.318** –0.270 –0.231 –0.257 –0.116 (0.27) (0.16) (0.20) (0.23) (0.20) (0.17) Exporter –0.599* 0.033 0.298 0.105 –0.063 0.565** (0.36) (0.25) (0.26) (0.24) (0.25) (0.26) Above 50% female –0.273* –0.194* 0.301** –0.019 0.050 0.044 employment (0.15) (0.12) (0.13) (0.15) (0.12) (0.12) Observations 2,070 2,070 2,070 2,070 2,070 2,070 Note: Standard errors in parentheses. Weights are applied in all estimations. * p<0.10, ** p<0.05, *** p<0.01 38  TABLE A3.2: Estimated correlation between shocks and business characteristics (1) (2) (3) (4) (5) (6) OLS Probit Probit Probit Probit Probit Reported Reduction Reduction Reduction Reduction change in operating Reduction in cash in finance in availability in sales hours in demand flow available of inputs Small (5–19) 0.125 –0.105 –0.142 –0.100 –0.120 –0.109 (2.54) (0.11) (0.10) (0.10) (0.10) (0.10) Medium 6.651** –0.174 –0.502*** –0.434*** –0.483*** –0.456*** (20–99) (3.04) (0.14) (0.12) (0.12) (0.12) (0.11) Large (100+) 10.718** –0.071 0.237 0.196 0.147 0.068 (5.39) (0.19) (0.18) (0.18) (0.17) (0.17) Manufacturing –8.838** 0.494*** 0.062 0.074 0.155 0.207 (3.59) (0.16) (0.17) (0.17) (0.17) (0.16) Retail/ –9.338** 0.402** –0.162 –0.213 –0.187 –0.014 wholesale (3.77) (0.17) (0.17) (0.18) (0.17) (0.16) Tourism –19.604*** 0.958*** –0.492*** –0.550*** –0.476*** –0.224 (3.84) (0.18) (0.17) (0.17) (0.17) (0.16) Other services –7.901** 0.520*** –0.544*** –0.637*** –0.562*** –0.442*** (3.42) (0.15) (0.15) (0.15) (0.15) (0.14) Nairobi 0.484 –0.081 0.196* 0.178 0.180 0.169 (2.74) (0.11) (0.11) (0.11) (0.11) (0.10) Maturing (5–14) –2.531 0.084 0.097 0.067 0.051 0.004 (3.65) (0.16) (0.14) (0.14) (0.14) (0.14) Established 2.079 0.120 0.068 –0.057 –0.123 –0.137 (15+) (3.96) (0.17) (0.15) (0.15) (0.15) (0.15) Exporter 4.254 –0.048 0.288 0.399 0.492* 0.457* (5.70) (0.25) (0.26) (0.28) (0.28) (0.24) Above 50% 0.286 0.288** –0.377*** –0.398*** –0.352*** –0.329*** female (2.89) (0.12) (0.10) (0.10) (0.10) (0.10) employment Observations 1,565 1,628 2,070 2,070 2,070 2,070 Note: Standard errors in parentheses. Sampling weights applied in all specifications. Base categories are micro-sized firms, the agricultural sector, firms in other regions, young firms, and non-exporters. * p<0,10, ** p<0,05, *** p<0,01 Appendix 3. Results from OLS and Probit Regressions   39  TABLE A3.3: Estimated correlation between sales reductions and shocks (1) (2) (3) (4) (5) (6) OLS OLS OLS OLS OLS OLS Accommodation/ Other Total Agriculture Manufacturing Retail food services Decrease in hours worked –4.206* –8.355 –2.503 –0.534 0.431 –6.580** (2.20) (5.05) (4.22) (4.20) (4.06) (2.87) Decrease in demand –16.107*** –9.542 –3.481 –21.393** –20.382* –14.916*** (4.56) (7.44) (11.51) (9.13) (11.09) (4.83) Decrease in cash flow –15.373*** –17.997* –26.725** 1.979 –21.725 –26.329*** (4.84) (9.44) (12.45) (8.81) (19.68) (5.49) Decrease in available –0.565 2.730 –5.402 –13.933 –15.613** 4.927 finance (3.86) (8.10) (8.00) (9.27) (6.12) (4.26) Decrease in availability –1.192 –11.215* –9.087 10.986 –1.600 –5.596 of inputs (3.39) (6.63) (6.49) (9.10) (5.76) (3.55) Observations 1,565 125 332 255 188 665 Note: Standard errors in parentheses. Sampling weights applied in all specifications. * p<0,10, ** p<0,05, *** p<0,01 40    Socioeconomic Impacts of COVID-19 in Kenya  TABLE A3.4: Estimated correlation between firm survival and business characteristics (1) (2) OLS OLS Days that Weeks that establishment establishment can continue to can remain open pay costs Small (5–19) 0.771 –0.636 (0.88) (4.78) Medium (20–99) 3.981* 10.738 (2.06) (7.30) Large (100+) 3.363** 45.210*** (1.49) (10.59) Manufacturing 3.270** –11.282 (1.64) (9.67) Retail/wholesale –1.303 –23.904** (1.49) (9.82) Tourism 0.591 –34.091*** (1.46) (9.15) Other services 0.199 –28.567*** (1.22) (8.71) Nairobi 0.493 17.238*** (1.13) (6.03) Maturing (5–14) 2.338* 9.169 (1.38) (5.59) Established (15+) 2.275 7.593 (1.48) (5.79) Exporter 2.016 –1.022 (2.05) (11.40) Above 50% female employment 1.195 –7.640* (0.93) (4.58) Observations 774 1,749 Note: Standard errors in parentheses. Sampling weights applied in all specifications. Base categories are micro-sized firms. the agricultural sector, firms in other regions, young firms, and non-exporters. The number of weeks a firm can remain open and the number of days a firm cover costs are winsorized at the 99 percent level, to account for outliers. * p<0.10. ** p<0.05. *** p<0.01 Appendix 3. Results from OLS and Probit Regressions   41  TABLE A3.5: Estimated correlation between responses to the pandemic and business characteristics (1) (2) (3) Probit Probit Probit Increased use of Investment in Repackaging of digital platforms digital solutions product mix Medium (20–99) 0.458*** 0.402** 0.190 (0.13) (0.16) (0.15) Large (100+) 0.612*** 0.842*** 0.408* (0.18) (0.24) (0.23) Manufacturing –0.130 0.387 0.743*** (0.19) (0.24) (0.24) Retail/wholesale 0.056 0.004 0.702** (0.22) (0.32) (0.28) Tourism –0.639*** 0.268 0.710*** (0.21) (0.27) (0.26) Other services 0.097 0.283 0.629*** (0.17) (0.24) (0.23) Nairobi 0.270* 0.224 0.305** (0.14) (0.16) (0.15) Maturing (5–14) –0.099 –0.186 0.181 (0.24) (0.24) (0.25) Established (15+) –0.104 –0.197 0.017 (0.24) (0.25) (0.26) Exporter 0.268 0.230 –0.516** (0.25) (0.29) (0.22) Above 50% female employment 0.097 0.259 0.240 (0.14) (0.16) (0.16) Observations 1,039 1,039 1,039 Note: Standard errors in parentheses. Sampling weights applied in all specifications. Base categories are large firms, the agricultural sector, firms in other regions, young firms, and non-exporters. Micro-sized firms were not included. * p<0.10. ** p<0.05. *** p<0.01 42    Socioeconomic Impacts of COVID-19 in Kenya  TABLE A3.6: Estimated correlation between type of digital platform used and business characteristics (1) (2) (3) (4) (5) (6) (7) Probit Probit Probit Probit Probit Probit Probit Business Production Supply chain Payment Service Marketing Sales administration planning management methods delivery Medium (20–99) 0.471*** 0.783*** 0.478** –0.271 0.167 –0.067 0.362** (0.18) (0.23) (0.23) (0.17) (0.20) (0.19) (0.18) Large (100+) 0.528 0.842** 0.883*** 0.541** 0.429* 0.759*** 0.528* (0.34) (0.33) (0.28) (0.27) (0.25) (0.26) (0.27) Manufacturing –0.449* –0.046 –0.143 0.245 0.318 –0.101 0.286 (0.27) (0.27) (0.29) (0.28) (0.26) (0.26) (0.26) Retail/wholesale –0.061 –0.240 –0.150 –0.303 –0.269 0.042 0.181 (0.34) (0.34) (0.33) (0.32) (0.29) (0.32) (0.32) Tourism 0.154 –0.148 –0.190 0.313 0.116 –0.552 –0.451 (0.34) (0.41) (0.41) (0.35) (0.34) (0.35) (0.37) Other services –0.042 –0.314 –0.196 –0.544** –0.564** –0.399* 0.312 (0.25) (0.24) (0.26) (0.24) (0.24) (0.23) (0.24) Nairobi 0.141 0.168 0.329 0.518*** 0.649*** 0.396** –0.013 (0.20) (0.21) (0.20) (0.18) (0.18) (0.19) (0.18) Maturing (5–14) –0.398 0.172 –0.118 –0.274 0.142 –0.791** 0.919*** (0.43) (0.38) (0.38) (0.37) (0.42) (0.34) (0.35) Established (15+) –0.073 –0.264 –0.796** –0.877** –0.009 –0.897*** 0.975*** (0.45) (0.41) (0.40) (0.38) (0.44) (0.34) (0.36) Exporter 0.153 0.396 0.454* 0.409 –0.110 –0.444* –0.561** (0.36) (0.26) (0.26) (0.44) (0.26) (0.26) (0.25) Above 50% female 0.006 0.145 –0.103 –0.116 –0.215 –0.126 0.297 employment (0.22) (0.22) (0.21) (0.19) (0.23) (0.20) (0.19) Observations 542 542 542 542 542 542 542 Note: Standard errors in parentheses. Sampling weights applied in all specifications. Sample consists of firms which indicated an increased use of digital platforms. Micro-sized firms were not included. * p<0.10, ** p<0.05, *** p<0.01 Appendix 3. Results from OLS and Probit Regressions   43  TABLE A3.7: Estimated correlation between self-reported most needed policies and business characteristics (1) (2) (3) (4) (5) (6) (7) (8) (9) Probit Probit Probit Probit Probit Probit Probit Probit Probit Deferral Deferral Access Loans with Access Cash of rent or of loan to new subsidized Fiscal Tax Wage to any transfer mortgage payments credit rates reductions deferral subsidies assistance Small (5–19) –0.053 –0.078 0.322** 0.221** 0.108 0.219* –0.095 0.406** 0.129 (0.10) (0.13) (0.14) (0.11) (0.10) (0.13) (0.10) (0.20) (0.11) Medium –0.247** –0.452*** 0.297* –0.039 0.059 0.323** –0.198* 0.674*** 0.361*** (20–99) (0.11) (0.16) (0.16) (0.13) (0.11) (0.15) (0.11) (0.19) (0.13) Large (100+) –0.155 –0.602*** 0.486** –0.253 0.006 0.545** –0.045 0.359 0.354** (0.18) (0.18) (0.22) (0.16) (0.18) (0.24) (0.19) (0.31) (0.17) Manufacturing –0.314** 0.128 0.143 1.000*** 0.331** 0.458** –0.127 –0.028 –0.188 (0.15) (0.20) (0.20) (0.22) (0.15) (0.22) (0.16) (0.29) (0.16) Retail/ –0.227 –0.087 –0.059 0.722*** 0.180 0.299 –0.068 –0.086 –0.151 wholesale (0.16) (0.21) (0.23) (0.23) (0.16) (0.24) (0.17) (0.30) (0.17) Tourism –0.279* 0.131 0.006 0.959*** 0.311** 0.250 0.036 –0.260 –0.052 (0.16) (0.20) (0.21) (0.22) (0.16) (0.24) (0.17) (0.28) (0.17) Other services –0.352** –0.073 0.005 0.864*** 0.162 0.483** –0.037 0.273 –0.324** (0.14) (0.18) (0.19) (0.21) (0.14) (0.21) (0.15) (0.25) (0.15) Nairobi –0.194* 0.103 0.015 –0.073 –0.212** 0.280** 0.016 –0.366 0.227** (0.10) (0.14) (0.14) (0.11) (0.10) (0.12) (0.10) (0.26) (0.11) Maturing (5–14) 0.064 –0.271* 0.167 0.079 –0.025 0.338** –0.068 0.001 0.063 (0.13) (0.16) (0.17) (0.14) (0.13) (0.17) (0.14) (0.28) (0.15) Established 0.116 –0.180 0.092 –0.096 –0.171 0.318* –0.137 0.345 –0.025 (15+) (0.14) (0.18) (0.19) (0.16) (0.14) (0.19) (0.15) (0.27) (0.17) Exporter –0.247 –0.127 0.135 –1.022*** –0.069 –1.392*** 0.675*** –0.168 0.172 (0.24) (0.24) (0.27) (0.24) (0.24) (0.27) (0.24) (0.31) (0.22) Above 50% –0.051 0.342*** 0.224 0.174 0.156 0.199 –0.231** 0.145 0.012 female (0.10) (0.12) (0.14) (0.11) (0.10) (0.13) (0.10) (0.17) (0.11) employment Observations 2,070 2,070 2,070 2,070 2,070 2,070 2,070 2,070 2,070 Note: Standard errors in parentheses. Sampling weights applied in all specifications. Base categories are micro-sized firms, the agricultural sector, firms in other regions, young firms, and non-exporters. * p<0,10, ** p<0,05, *** p<0,01 44    Socioeconomic Impacts of COVID-19 in Kenya  TABLE A3.8: Estimated correlation between policy demand and shocks (1) (2) (3) (4) (5) (6) (7) (8) Probit Probit Probit Probit Probit Probit Probit Probit Deferral Deferral Loans with Monetary of rent or of loan Access to subsidized Fiscal Tax Wage transfer mortgage payments new credit rates reductions deferral subsidies Decrease in –0.169* 0.702*** 0.262* 0.268** 0.209** 0.507*** 0.335*** 0.624*** hours worked (0.10) (0.16) (0.13) (0.12) (0.10) (0.13) (0.10) (0.18) Decrease in 0.725*** 0.457** –0.313 –0.263 –0.415** –0.381** –0.071 –0.168 demand (0.21) (0.20) (0.25) (0.17) (0.19) (0.18) (0.19) (0.17) Decrease in 0.079 0.117 –0.460** –0.472** –0.151 –0.642*** 0.148 –0.448** cash flow (0.24) (0.25) (0.22) (0.23) (0.24) (0.21) (0.27) (0.18) Decrease –0.056 0.652** 0.331 0.407* 0.378* 0.007 0.055 –0.696*** in available (0.22) (0.26) (0.22) (0.22) (0.21) (0.20) (0.24) (0.27) finance Decrease in –0.155 0.018 0.222 0.347** 0.193 0.022 –0.058 1.120*** availability of (0.13) (0.23) (0.18) (0.15) (0.12) (0.17) (0.13) (0.33) inputs Observations 1,628 1,628 1,628 1,628 1,628 1,628 1,628 1,628 Note: Standard errors in parentheses. Sampling weights applied in all specifications. * p<0,10, ** p<0,05, *** p<0,01 Appendix 3. Results from OLS and Probit Regressions   45