Policy Research Working Paper 9851 Competition and Firm Recovery Post-COVID-19 Miriam Bruhn Asli Demirguc-Kunt Dorothe Singer Europe and Central Asia Region & Development Research Group November 2021 Policy Research Working Paper 9851 Abstract This paper examines the impact of the COVID-19 crisis environment experienced more reallocation from less pro- on the reallocation of economic activity across firms, and ductive to more productive firms than countries with a weak whether this reallocation depends on the competition competition environment. The evidence also suggests that environment. The paper uses the World Bank’s Enterprise reallocation from low- to high-productivity firms during Surveys COVID-19 Follow-up Surveys for about 8,000 the COVID-19 crisis was stronger compared with pre-crisis firms in 23 emerging and developing countries in Europe times. Finally, the analysis shows that government support and Central Asia, matched with 2019 Enterprise Surveys measures implemented in response to the crisis may have data. It finds that during the COVID-19 crisis, economic adverse effects on competition and productivity growth activity was reallocated toward firms with higher pre-crisis since support went to less productive and larger firms, labor productivity. Countries with a strong competition regardless of their pre-crisis innovation. This paper is a product of the Office of the Chief Economist, Europe and Central Asia Region and the Development Research Group, Development Economics. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at mbruhn@worldbank. org, ademirguckunt@worldbank.org, and dsinger@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Competition and Firm Recovery Post-COVID-19 * Miriam Bruhn,a Asli Demirguc-Kunt,b Dorothe Singerc Keywords: Creative destruction, productivity, competition, firms, COVID-19, government support, Europe and Central Asia (ECA) JEL codes: D22, D24, H81, O47 * We thank Meghana Ayyagari, David C. Francis, Leonardo Iacovone, Georgiana Pop, Jesica Torres Coronado, and Trang Thu Tran for their comments and suggestions. Paul Nicholas Villaflor Pacheco provided excellent research assistance. a mbruhn@worldbank.org, World Bank and IZA. b ademirguckunt@worldbank.org, World Bank. c dsinger@worldbank.org, World Bank. 1. Introduction The COVID-19 pandemic resulted in the deepest global recession since World War II (World Bank 2020a). The rapid spread of the disease and movement restrictions to contain the virus created sudden supply and demand shocks, leading to declines in output (Demirgüç-Kunt, Lokshin, and Torre 2021). These shocks were compounded by widespread uncertainty among businesses and concerns about the risks of increasing financial sector distress limiting access to credit (Bloom and others 2021; Buchheim, Krolage, and Link 2020; and World Bank 2021b). How did firms fare? Previous work has documented the initial impacts of the COVID-19 crisis on businesses around the world (Aga and Maemir 2021; Balleer and others 2020; Bartik and others 2020; Bennedsen and others 2020; Crane and others 2021; Fairlie 2020; Liu, Wei, and Xu 2021; Wagner 2021). Across 51 countries, firms reported an average drop in sales of 49 percent (Apedo-Amah and others 2020), but the effects of the crisis were heterogenous. For example, small firms experienced larger drops in sales than large firms (Brucal, Grover, and Reyes Ortega 2021; De Nicola and others 2021; Karalashvili and Viganola 2021). As argued by Schumpeter (1942), crises can result in “creative destruction.” That is, they can play a cleansing role, by speeding up the reallocation of economic activity from less productive to more productive firms, leading to greater productivity and economic growth in the longer run (Caballero and Hammour 1994). However, it is not obvious whether a pattern of creative destruction will emerge from the COVID-19 crisis. The broad-based shock may simply result in widespread destruction (Bosio and others 2020; De Nicola and others 2021). Or, instead of shifting economic activity to the most productive firms, the crisis may reallocate activity to firms with greater market power or political connections, which could be detrimental to productivity growth (Di Mauro and Syverson 2020). The question of the impact of COVID-19 on firms and its implications for long-run economic growth is thus an empirical one. This paper makes four contributions. First, we examine the process of reallocation across a large set of emerging and developing economies during the COVID-19 crisis. We use data from the World Bank’s Enterprise Surveys (ES) COVID-19 Follow-up Surveys for about 8,000 firms in 23 countries in Europe and Central Asia (ECA), matched with 2019 ES data, to study the relationship between pre-COVID-19 labor productivity and performance during the crisis as measured by changes in sales and employment, anticipating falling into arrears on outstanding liabilities, increased online business activity, and increase in remote work arrangements. Second, we investigate the role of the competition environment in the process of reallocation during a crisis. When markets are competitive, they do a better job of allocating resources toward more productive firms (Arnold, Nicoletti, and Scarpetta 2011; Brown and Earle 2002; Caballero 2008), which may be particularly relevant in a crisis. For example, in countries with a weak competition environment, market power and political connections, instead of productivity, may influence which firms do better during a crisis. Weak competition can also limit innovation and the ability to respond to shocks. Evidence from the United States suggests that firms respond more sluggishly to shocks now than they did decades ago, and regulatory barriers to business entry and expansion are important reasons for the increased sluggishness (Barrero, Bloom, and Davis 2020). We use the Bertelsmann Stiftung Transformation Index for 2019 to measure a country’s competition environment – and several alternative measures in our robustness checks – and interact this index with pre-COVID-19 labor productivity in our analysis. 2 Third, we ask how reallocation during the COVID-19 crisis compares with reallocation pre-crisis. We use additional data from the 2019 ES to study the relationship between labor productivity in 2017 and growth of sales and employment between 2017 and 2018 and compare the findings with our results on reallocation during the COVID-19 crisis. Fourth, we examine what type of firms received government support during the COVID-19 crisis and assess whether the distribution of support measures may hamper competition and slow the recovery process. Many governments implemented broad support schemes to address the initial economic fall- out from the COVID-19 crisis and preserve jobs and firm-specific intangible capital (World Bank 2021b). The large economic shocks required quick action, making it difficult to target government support. The ES COVID-19 Follow-up Surveys show that the reach of government support measures varied widely across countries, but on average half the firms in our sample reported having received some government support during the pandemic. Our paper contributes to several strands of the literature. First, it contributes to the nascent literature on productivity and firm performance during the COVID-19 crisis. Muzi and others (2021) examine ES data for 31 countries and find that higher productivity firms were less likely to exit during the COVID-19 crisis. Kozeniauskas, Moreira, and Santos (2020) analyze data from a survey of Portuguese firms conducted between April and July 2020, showing that higher productivity firms were more likely to remain open and less likely to cut employment. Bloom and others (2021) use data from a survey of UK firms, collected between July 2020 and August 2021. They find that higher productivity firms contracted hours worked less than lower productivity firms did. Our paper also contributes to the literature that studies how economic crises more broadly defined affect the reallocation of economic activity between firms. Studies of earlier economic crises have found mixed evidence on reallocation of economic activity from less to more productive firms. Foster, Grim, and Haltiwanger (2016) study establishment-level data from the US Census Bureau for 1981 to 2010. They find increased reallocation from less productive to more productive establishments in recessions prior to the Great Recession, but the cleansing impact of earlier recessions attenuated during the Great Recession. Similarly, Hallward-Driemeier and Rjikers (2013) use Indonesian manufacturing census data from 1991 to 2001 to show that more productive firms were less likely to exit before the East Asian crisis, but this relationship weakened during the crisis. Using data for Colombian manufacturing firms from 1995 to 2004, Eslava and others (2010) show that credit constraints can limit the reallocation of economic activity from less to more productive firms during a financial crisis. Finally, our paper contributes to the nascent literature on the implications of government support measures enacted in the wake of the COVID-19 pandemic. Preliminary evidence from the literature suggests that government support measures helped firms weather the crisis and decreased the failure rates of small and medium-size enterprises (SMEs) (Albagli, Fernández, and Huneeus 2021; Cespedes, Chang, and Velasco 2021; Cirera and others 2021; Gourinchas and others 2021). However, there is little evidence on whether the support was allocated to the most viable firms. Our findings can be summarized as follows: - The results show that economic activity was reallocated toward more productive firms during the COVID-19 crisis. Firms with high pre-crisis labor productivity experienced smaller drops in 3 sales and employment than firms with low pre-crisis labor productivity. More productive firms were also more likely to adapt to the crisis by increasing online activity and remote work. - Countries with a strong competition environment experienced more reallocation from less productive to more productive firms than countries with a weak competition environment. This result is robust to using alternative measures of the competition environment. In countries with high competition, firms at the 10th percentile of the pre-crisis labor productivity distribution experienced an 18 percentage point larger drop in sales than firms at the 90th percentile; this difference is only 10 percentage points in countries with low competition. Similarly, in countries with high competition, firms in the 10th percentile of the labor productivity distribution were 15 percentage points more likely to decrease employment and 16 percentage points more likely to anticipate falling into arrears than firms in the 90th percentile. The corresponding differences in countries with low competition were 5 and 8 percentage points, respectively. In countries with high competition, the lowest productivity firms were 19 percentage points less likely to have increased remote work than the highest productivity firms; this difference was only 5 percentage points in countries with low competition. - Investigating how reallocation during the COVID-19 crisis compares with pre-crisis reallocation, we find that the relationship between productivity and employment growth was stronger during COVID-19. More productive firms experienced lower sales growth than less productive firms pre-COVID-19, a pattern that was reversed during the crisis. Compared with less productive firms, more productive firms experienced higher employment growth pre-crisis, but to a lesser or equal extent than during the crisis. These findings are subject to the caveat that there are some limits to the comparability of the growth rates of sales and employment. - Examining the determinants of receiving government support, we have three main findings. First, productive firms were less likely to receive any type of government support, even controlling for the drop in sales or employment experienced by the firm. Second, larger firms were more likely than smaller firms to receive some types of government support, which could indicate that support was more likely to go to politically connected firms. Third, governments provided support to firms regardless of their pre-crisis level of innovation. The paper is organized as follows. Section 2 presents the data and summary statistics. Section 3 describes the empirical strategy. Section 4 discusses the main results. Section 5 includes several robustness checks, and section 6 concludes. 2. Data and Summary Statistics 2.1 Enterprise Surveys Data To measure firm performance and government support received during the COVID-19 crisis, we use data from the ES COVID-19 Follow-up Surveys. 1 The World Bank has collected one or two rounds of these surveys in 44 countries starting in May 2020. We use the latest available survey round for a country as of June 28, 2021. We conduct our main analysis on a sample of 23 emerging markets and developing 1 More information on the ES COVID-19 Follow-up Surveys is available at https://www.enterprisesurveys.org/en/covid-19. 4 countries in ECA to focus on a homogenous environment (see table A.1). As a robustness check, we replicate the results for a global sample of 30 countries (see section 5.4). The sampling frame for the ES COVID-19 Follow-up Surveys includes all firms that replied to the latest pre-COVID-19 ES, making it possible to link performance during the COVID-19 pandemic back to firm characteristics collected through the ES. The ES cover formal (registered) firms with five or more employees and are designed to be nationally representative for the manufacturing and service sectors. For the countries in our sample, the most recent ES was conducted in 2019 or early 2020. The ES COVID-19 Follow-up Surveys ask firms about changes in performance during the crisis, as well as expectations for the next months. Table A.2 provides the definitions of the variables included in our analysis. Table 1 presents the summary statistics. The number of observations varies across variables due to nonresponse. Firms reported that their sales dropped by an average of 24 percent during the COVID-19 crisis. The decline in the number of permanent full-time workers (13 percent) was more muted than the drop in sales, a pattern that was also observed by Apedo-Amah and others (2020). They argue that firms reduced their workforces proportionally less than their drop in sales in part because they relied on other mechanisms to adjust output and labor costs, including granting leave, reducing hours, and lowering wages. Close to 34 percent of the firms have decreased their number of workers since December 2019. About 26 percent reported that they anticipated falling into arrears on outstanding liabilities in the next six months. To respond to the crisis, 28 percent of the firms started or increased online business activity, and 34 percent started or increased remote work arrangements. The ES COVID-19 Follow-up Surveys also include questions on government support received during the pandemic. Governments in ECA enacted numerous measures to support firms, and many were directed to SMEs, according to the World Bank’s SME-Support Measure Dashboard. 2 The five most common types of measures were new credit (26 percent of measures), fiscal relief (20 percent), wage subsidies (17 percent), payment deferrals (12 percent), and cash grants (7 percent). These five types of measures are also the ones captured in the ES COVID-19 Follow-up Surveys. These surveys show that the measure reaching by far the largest number of firms was wage subsidies, with 36 percent of firms receiving such subsidies (table 1). The second most wide-reaching measure was cash transfers, followed by fiscal relief and payment deferrals. Although new credit was the most used measure, according to the World Bank’s SME-Support Measure Dashboard, it reached only 5 percent of the firms. Altogether, 45 percent of the firms reported receiving some type of government support. From the pre-COVID ES, we obtain our measure of productivity, as well as firms’ background characteristics. We use labor productivity, defined as (log of) sales over total number of permanent full- time employees. 3 Labor productivity is preferred to total factor productivity due to limitations in measuring non-labor inputs in the ES (World Bank 2021a). Panel B in table 1 displays summary statistics 2 See https://dataviz.worldbank.org/authoring/SME-COVID19/Overview. 3 In calculating labor productivity, the outliers are eliminated by first log-transforming total annual sales and the number of permanent full-time employees, and then trimming at plus and minus three standard deviations from the mean, as described in World Bank (2021a). 5 for the pre-COVID ES data. The average firm in our sample had 31 employees (2.65 in log units) and was 19 years old (2.79 in log units). Only 21 percent of the firms had a female top manager. The sample includes few firms with at least 10 percent state or foreign ownership (1 and 8 percent, respectively). Less than half the firms reported that they innovated a product or process during 2017 to 2019 (40 percent) or had a loan (40 percent), but 65 percent owned a website. Most firms reported that their main market was local (44 percent) or national (44 percent) versus international (12 percent). Slightly over half the firms are in the service sector (53 percent), followed by manufacturing (26 percent) and retail (21 percent). To examine how reallocation from less productive to more productive firms during the COVID-19 crisis compares with pre-crisis reallocation, we use additional data from the latest pre-COVID-19 ES. These surveys collected sales and employment data for the year of the survey (typically 2019) and two years prior (typically 2017). We calculate labor productivity for 2017, following the same procedure as for 2019. We extrapolate 2018 sales and employment as the simple average of the 2017 and 2019 values, to then calculate one-year growth rates, which are comparable to the sales and employment growth rates collected in the ES COVID-19 Follow-up Surveys. Table 1 shows that the average log of labor productivity in 2017 was only slightly lower than in 2019 (10.47 and 10.51, respectively). On average, firms experienced positive sales growth from 2017 to 2018, at a rate of 9.21 percent. Employment also grew on average, but at a lower rate than sales (4.07 percent). 2.2 Competition Environment We use one main measure of the competition environment and report results for additional measures in the section on robustness checks. Our main measure of market organization is the 2020 Bertelsmann Stiftung Transformation Index (BTI), which surveys political and economic transformation around the world based on expert opinion for 2019. The measure captures the extent to which the fundamentals of market-based competition have been developed, with a maximum score (10) indicating that “market competition is consistently defined and implemented both macroeconomically and microeconomically. There are state-guaranteed rules for market competition with equal opportunities for all market participants. The informal sector is very small” (BTI 2020, 27). The minimum score (1) indicates that “market competition is present only in small segments of the economy and its institutional framework is rudimentary. Rules for market participants are unreliable and frequently set arbitrarily. The informal sector is large” (BTI 2020, 27). We use three additional measures of the competition environment in the section on robustness checks. The first alternative measure is BTI competition policy, which captures the extent to which safeguards exist to protect competition and are enforced, with the maximum score (10) indicating that “comprehensive competition laws to prevent monopolistic structures and conduct exist and are strictly enforced” (BTI 2020, 28). A score of 1 indicates that “no legal or political measures are taken to prevent monopolistic structures and conduct” (BTI 2020, 28). As such, it is a more narrowly defined measure of the competition environment. The average score for ECA is 7 for BTI market organization and 8 for competition policy, with significant variation across countries (table 2). A score of 7 corresponds to an assessment that “market competition 6 has a strong institutional framework, but the rules for market competition are not consistent or always uniform for all market participants. The informal sector is small” and “competition laws to prevent monopolistic structures and conduct exist, but are enforced inconsistently” (BTI 2020, 27-28). The Czech Republic, Estonia, and Slovenia all achieve the maximum score of 10 for both BTI measures. Croatia, Latvia, Lithuania, and Poland achieve a score of 10 for the competition policy measure. The competitive environment is least developed in Azerbaijan, which scores poorly (4) on both BTI measures. Belarus scores similarly poorly on market organization. We use BTI market organization data for 2017 when examining reallocation in pre-crisis times. Overall, the scores are similar to the ones in 2019, with an average score of 7 for ECA and a similar pattern of variation across countries. The Czech Republic and Estonia achieve the maximum score of 10, while Azerbaijan and Belarus score poorly (4). We also use two proxy measures for the size and influence of the state in the economy as alternative competition variables since they may influence the competition environment. Specifically, countries in which the state plays a relatively large and influential role in the economy may have weaker competition environments. One proxy is the share of public sector employees in total employment, from the Worldwide Bureaucracy Indicators (World Bank 2020b). 4 This share averages 23.6 percent in ECA and ranges from 13.5 percent in Romania to 39.3 percent in Belarus. The other proxy is the share of public banks in total banking sector assets, available from Anginer, Demirgüç-Kunt, and Mare (2020). A public bank is defined as a commercial bank in which the state has more than 50 percent ownership stake. A large body of literature suggests that public banks may be subject to government influence and that their existence may distort credit allocation, introducing inefficiencies (World Bank 2012). The 12 percent average for the region masks huge variation across countries. The share is as high as 66 percent in the Russian Federation and 64 percent in Belarus, but the share is 0 in 10 of the 23 countries in the sample. Finally, in the section on robustness checks, we use log gross domestic product (GDP) per capita for 2019, from the World Bank’s World Development Indicators database, to test whether the competitive environment just captures the level of economic development. On average, the ECA countries in our sample have a GDP per capita of about US$10,400 (9.25 in log units), ranging from about US$4,500 (8.14 in log units) in Moldova to about US$26,000 (10.16 in log units) in Slovenia. 3. Empirical Strategy 3.1 Firm Performance, Productivity, and Competition We examine the relationship between firms’ pre-crisis labor productivity and firm performance during the crisis by estimating the following regression specification: (1) = β0 + β1 + β2 + ℎ + + + ε 4 Local sources were used where Worldwide Bureaucracy Indicators were not available or outdated (Armenia, Azerbaijan, Belarus, Montenegro, North Macedonia, and the Russian Federation). 7 where subscript i denotes each firm; j the sector of the firm; and k the country in which the firm is located. Similar specifications have been used by Kozeniauskas, Moreira, and Santos (2020) and Muzi and others (2021) to assess the impact of COVID-19 on firm performance. The dependent variable Firm performance is captured by one of six variables: percentage change in sales; percentage change in employment; decreased employment; anticipate falling into arrears; increased online activity; and increased remote work. We estimate all regressions using ordinary least squares. Our two main variables of interest are Productivity and Productivity x Competition, an interaction term between labor productivity and country-level measures of competition, to examine whether the relationship between labor productivity and firm performance during the crisis varied with market competition. Productivity is proxied by pre-COVID-19 labor productivity, measured by sales per worker. Competition is captured by the variable BTI market organization (and additional variables in the robustness section). The country-level measures of competition are de-meaned, so that a value of zero corresponds to the average value of competition across countries. The coefficient of labor productivity in table 3 thus represents the relationship between productivity and firm performance for a country with the average value of competition. Firm Characteristics is a vector of firm-level controls and includes size (number of employees), age, gender of the top manager, innovated a product or process during 2017–19, state ownership, foreign ownership, line of credit or loan, owns a website, and location of main market (local, national, or international). The set of firm characteristics comprises a standard set of variables related to a firm’s performance during crisis. 5 On the one hand, larger and older firms may weather a crisis better because they are more established and have more resources, financial and organizational, to do so. They may also have strong political connections. On the other hand, smaller and younger firms may be nimbler and can thus more easily adjust their operations when faced with a crisis. More innovative firms may be more likely to adapt during a crisis. Firms with state or foreign ownership may be more resilient to crises by having access to more resources, connections, or know-how. At the same time, state ownership may make firms less adaptable in times of crisis. Firms with a line of credit or loan pre-crisis may have easier access to finance during a crisis, allowing them to mitigate its impacts. We control for the gender of the top manager as Liu, Wei, and Xu (2021) document that women-led businesses were more likely to close, close for a longer period, and anticipate falling into arrears than men-led businesses during the COVID- 19 pandemic. We also control for owning a website. Wagner (2021) shows that firms with a website were more likely to survive during the pandemic, potentially because having an online presence allows firms to be in contact with potential customers even when in-person contact is restricted. Finally, we control for the location of the firm’s main market (local, national, or international) as movement restrictions due to the pandemic might have impacted firms with different main markets differently. Moreover, export-oriented firms may have been able to substitute sales at home with sales abroad during the negative demand shock at home given that, while global in scale, the COVID-19 crisis affected countries differently (Muzi and others 2021). Similarly, firms whose main market is national rather than local may have substituted sales as the COVID-19 crisis often affected regions within a country differently. 5 See, for example, Muzi and others (2021). 8 The variables S and C represent sector and country fixed effects, respectively. The error term ε is clustered at the country level. The specification is estimated using sampling weights provided in the ES COVID-19 Follow-up Surveys to correct for unequal probability of selection as well as ineligibility. We rescale the sampling weights to give equal weight to all countries since the focus of our analysis is on competition, which is measured at the country level. A potential challenge in our analysis of the relationship between firms’ pre-crisis labor productivity and performance during the pandemic is endogeneity. Firm performance during the crisis might be a predictor of a firm’s labor productivity. As noted by Muzi and others (2021), the fact that all our explanatory variables, including labor productivity, come from the matched pre-COVID-19 ES data set should alleviate this concern. Another potential challenge is that a country’s competitive environment may be correlated with other country characteristics that could influence the reallocation of economic activity from less to more productive firms. In addition to including alternative measures of competition (see section 2.2), we therefore interact productivity with log GDP per capita instead of our measures of competition as a robustness test to check whether we just capture the level of economic development. 3.2 Government Assistance, Productivity, and Firm Characteristics We examine the relationship between receiving any government assistance or different types of government assistance and firm characteristics by estimating the following regression specification: (2) = β0 + β1 + ℎ + + + ε where subscript i denotes each firm; j is the sector of the firm; and k is the country in which the firm is located. The dependent variable is Government Assistance received by a firm, which is captured by one of six binary variables: received any government support and types of government assistance received (cash transfer, payment deferrals, new credit, fiscal relief, or wage subsidies). We estimate the regression using ordinary least squares. Here our main variable of interest is Productivity, which is proxied as in specification (1) by pre-COVID- 19 labor productivity, measured by sales per worker. Firm Characteristics is the same vector of firm-level controls as above, as we expect the firm characteristics that are correlated with firm performance during the COVID-19 crisis to be correlated with receiving government assistance to weather the crisis. The variables S and C represent sector and country fixed effects, respectively, and ε is the error term, which is clustered at the country level. The specification is estimated using rescaled sampling weights provided in the ES COVID-19 Follow-up Surveys to correct for unequal probability of selection and ineligibility. The challenge in our analysis of the relationship between government assistance, productivity, and firm characteristics is that the data do not include detailed information on when exactly firms received government support, making it difficult to assess whether government support may have influenced firm performance. Furthermore, although we know which firms received government assistance, we do 9 not know the criteria that made these firms eligible for such support. As a robustness check, we include additional controls for the percentage change in sales and employment, to control for the possibility that the firms that were impacted the most were more likely to receive support. 4. Results 4.1 Reallocation from Less Productive to More Productive Firms The evidence in this section suggests that economic activity in ECA was reallocated toward more productive firms during the crisis, consistent with creative destruction. Table 3 displays the results of estimating equation 1. The coefficients on labor productivity show that firms with high pre-crisis labor productivity experienced statistically significantly smaller drops in sales and employment than firms with low pre-crisis labor productivity. More productive firms were also less likely to anticipate falling into arrears and more likely to adapt to the crisis by increasing remote work. The magnitudes of the coefficients on labor productivity in table 3 reveal that firms in the 10th percentile of the pre-crisis labor productivity distribution experienced a 13 percentage point larger drop in sales and a 9 percentage point larger drop in employment than firms in the 90th percentile. 6 Firms in the 10th percentile were also 11 percentage points less likely to have increased remote work than firms in the 90th percentile. Firms in the 10th percentile of the pre-crisis labor productivity distribution were 11 percentage points more likely to anticipate falling into arrears than firms in the 90th percentile. Table 3 further shows that larger firms weathered the crisis better. Firms with female top managers were hit harder by the crisis than firms with male top managers. State-owned firms experienced smaller drops in sales than privately-owned firms, although this is only statistically significant at the 10 percent level. A more significant difference appears in state-owned firms’ lower ability to adjust to the crisis by increasing online activity and remote work. Firms that were more innovative in the pre-crisis period continued to innovate during the crisis, since they were more likely to increase their online activity and remote work. Finally, firms that primarily sell to the national market did better than firms that sell to the local market, and those with access to international markets were better able to increase remote work and anticipated lower arrears. 4.2 The Role of Competition The importance of competition has been evident during the COVID-19 crisis. The coefficients on the interaction terms in table 3 show how the relationship between labor productivity and firm performance depends on market competition. For example, the positive coefficient on “Log(labor productivity)*BTI market organization” in the first column in table 3 indicates that the relationship between labor productivity and the drop in sales was stronger in countries with greater competition. 6 The 10th percentile of log labor productivity is 8.87, whereas the 90th percentile is 12.04 (table 1). We multiply the difference between these two (3.18) by the coefficients on log labor productivity in table 3 (4.138 and 2.776), to obtain a 13 percentage point difference in the change in sales and a 9 percentage point difference in the change in employment. 10 Overall, the coefficients on the interaction terms in table 3 suggest that in ECA countries with more competitive markets, there was more reallocation of economic activity from less productive to more productive firms. Figure 1 illustrates the findings from table 3. In countries with high competition, the relationship between labor productivity and firm performance during the crisis is steeper than in countries with low competition. In countries with high competition, firms at the 10th percentile of the pre-crisis labor productivity distribution experienced an 18 percentage point larger drop in sales than firms at the 90th percentile; this difference is only 10 percentage points in countries with low competition. Similarly, in countries with high competition, firms in the 10th percentile of the labor productivity distribution were 15 percentage points more likely to decrease employment and 16 percentage points more likely to anticipate falling into arrears than firms in the 90th percentile. The corresponding differences in countries with low competition were 5 and 8 percentage points, respectively. In countries with high competition, the lowest productivity firms were 19 percentage points less likely to have increased remote work than the highest productivity firms; this difference was only 5 percentage points in countries with low competition. 4.3 Reallocation Pre-COVID-19 We now examine how reallocation during the COVID-19 crisis compares with pre-crisis reallocation. Panel A in table 4 replicates the regressions in the first two columns in table 3 for the pre-COVID-19 period. The specification here includes fewer controls since most of the controls in table 3 were not collected for 2017. We control for number of employees, firm age, and sector.7 To facilitate the comparison of the pre-COVID-19 and during COVID-19 effects, panel B in table 4 shows the regressions from table 3 with the reduced set of controls. The results in panel A in table 4 show that more productive firms experienced lower sales growth pre- COVID-19, but they experienced higher employment growth. Comparing panels A and B, we find that the relationship between productivity and employment growth was weaker pre-COVID-19 than during COVID-19. This finding is subject to the caveat that the growth rates of employment do not always cover the same period since the growth rate during the pandemic is relative to employment in December 2019 and the ES COVID-19 Follow-up Surveys were conducted in different months during 2020 and 2021 (see table A.1). Nevertheless, unlike studies that compare reallocation during previous crises to non-crisis times in the United States and Indonesia (Foster, Grim, and Haltiwanger 2016; Hallward-Driemeier and Rjikers 2013), we find no evidence that reallocation of sales and employment from less to more productive firms decreased during COVID-19. On the contrary, this reallocation in ECA appears to have increased during COVID-19. Other notable results in table 4 are that younger and smaller firms in ECA grew faster pre-COVID-19 (consistent with the findings of Ayyagari, Demirguc-Kunt, and Maksimovic (2020) and Haltiwanger, Jarmin, and Miranda (2013)), while larger firms performed better than small firms and age was not correlated with growth during the pandemic. Countries with more competitive markets experienced more reallocation from less productive to more productive firms during COVID-19 but not before COVID- 19. 7 We assume that the sectors did not change between 2017 and 2019. 11 4.4 Government Assistance and Firm Characteristics Table 5 shows the results from estimating equation 2 to examine the correlates of receiving government assistance. Three main findings emerge from this analysis. First, more productive firms were less likely to receive any type of government support. Firms in the 10th percentile of the pre-crisis labor productivity distribution were 8 percentage points more likely to receive government support than firms in the 90th percentile. This result is consistent with Kozeniauskas, Moreira, and Santos (2020), who find that higher productivity firms in Portugal were less likely to receive government support, controlling for age, total income, and an indicator for whether the firm is located in the Lisbon region. Second, larger firms were more likely than smaller firms to receive payment deferrals and fiscal relief. This result is in line with Cirera and others (2021), who find that across 60 countries, support was more limited for smaller firms than larger firms. This finding could indicate that support was more likely to go to politically connected firms, which tend to be much larger than firms that are not politically connected. Before COVID-19, politically connected firms were more likely to receive subsidies (Francis, Hussain, and Schiffbauer 2018). Third, pre-crisis innovation is not correlated with receiving any type of government support. Table 5 further shows that firms with a female top manager, firms that had a line of credit or loan, and firms that own a website were more likely to receive government support. Taken together, we find that the broad support governments provided disproportionately went to less productive and larger firms, irrespective of their pre-crisis innovativeness. Less productive firms that stay in the market with the help of government support can stifle innovation and productivity growth in the recovery phase and beyond. The provision of more support to large firms can also increase industry concentration and market power. 5. Robustness Checks 5.1 Measurement of Competition A challenge when estimating equation 1 is that competition is likely correlated with other country characteristics that could also influence reallocation from less to more productive firms. That is, the interaction term in equation 1 may not specifically capture the role of competition; it may capture the role of other characteristics, such as the level of economic development. We conduct two robustness checks to address this concern. First, we show that our findings are similar when using three alternative measures of the competitive environment. Table A.3 shows the results of estimating equation 1 when using the BTI competition policy measure. The coefficients on the interaction terms are similar to using the BTI market organization measure but only statistically significant for decreased employment and anticipating falling into arrears. Public sector employment and ownership of banks can also affect competition. The results for these two measures are reported in tables A.4 and A.5, respectively. Labor markets may be less competitive and 12 dynamic when the share of public sector employment is high. Using the share of public employment as a proxy for competition reveals that a larger share of employment in the public sector was associated with reduced reallocation from lower productivity to higher productivity firms during the COVID-19 crisis. Public ownership of banks can limit dynamism in the economy by leading to less efficient allocation of credit. The coefficients on the interaction terms suggest that reallocation from lower productivity to higher productivity firms during the COVID-19 crisis was lower in countries with a larger share of government-owned banking sector assets. Second, we use log GDP per capita instead of the BTI in the interaction term with labor productivity to test whether the BTI measures of competition just capture the level of economic development. The results in table A.6 provide little evidence that reallocation from less to more productive firms was greater in countries with higher GDP per capita. The interaction term is only statistically significant at the 10 percent level for one outcome variable: decreased employment. Higher productivity firms were less likely to decrease employment in countries with higher GDP per capita. However, GDP per capita does not mediate the correlation between productivity and change in sales, anticipation of falling into arrears, or probability of increased remote work. The robustness checks in tables A.3 to A.6 thus suggest that competition instead of economic development more broadly played a role in increasing reallocation from less to more productive firms during the COVID-19 crisis. 5.2 Controlling for the Severity of COVID-19 Restrictions Another identification concern is that measures of competition may be correlated with a country’s response to the crisis. For example, movement restrictions may have been more stringent in countries with a more competitive environment. We thus run a robustness test where we add an interaction term of labor productivity and a COVID-19 restrictions index to the specification in table 3. Here, COVID-19 restrictions are measured by the number of days the Oxford COVID-19 Government Response Tracker’s Stringency Index—which captures the strictness of “lockdown-style” policies that primarily restrict people’s behavior along nine dimensions ranging from 0 to 100, with higher values indicating more restrictions 8—was above 70 preceding the fieldwork start day of the ES COVID-19 Follow-up Surveys module in each country. Table A.7 shows the results. The sample is smaller than in table 3 because the COVID-19 restrictions index is not available for Armenia, Montenegro, and North Macedonia. Most of the main results in table A.7 are similar to those in table 3, with the exception that the interaction of labor productivity and the BTI market organization measure is not statistically significantly correlated with the probability of increasing remote work. Thus, overall, the relationship between competition and reallocation from less to more productive firms during the crisis is not driven by differences in COVID-19 restrictions across countries. 5.3 Endogeneity of Government Assistance and Productivity A potential source of endogeneity when estimating the relationship between government assistance and productivity in table 5 is that more productive firms may have been less likely to receive government support than less productive firms since the former experienced smaller drops in sales and employment. The ES data do not include detailed information on when firms received government 8 Details on how the index is calculated are available at https://github.com/OxCGRT/covid-policy- tracker/blob/master/documentation/index_methodology.md. 13 support, making it difficult to assess whether a drop in sales/employment or government support came first. This timing issue is the reason why we do not control for drop in sales or employment in equation 2 (which we estimate in table 5). However, we conduct robustness checks where we add change in sales or employment to equation 2. The results are in tables A.8 and A.9. They show that firms with larger drops in sales were more likely to receive government support, but for a given change in sales or employment, firms with lower labor productivity were still more likely to receive support. 5.4 Global Sample To examine whether our findings are specific to ECA countries, we replicate tables 3 and 5 for the global sample of 30 countries for which all the data we use are available, thus adding seven countries outside ECA (listed in panel B in table A.1). This sample includes all countries for which pre-COVID ES data are available for 2018 to 2020 (with fieldwork completion dates in 2019 to 2020). While ES data are available for five additional countries (Cyprus, Greece, Italy, Malta, and Portugal), these countries are not covered by the BTI competition measure. Table A.10 replicates table 3 with the global sample, and table A.11 replicates table 5. Most of the main findings are similar. The global data show that economic activity was reallocated from less productive to more productive firms during the COVID-19 crisis. This reallocation was greater in countries with more competitive markets (table A.10). Government support went to firms with lower labor productivity, irrespective of their innovativeness (table A.11). The only different main finding in the global sample is that firm size is not correlated with the probability of receiving government support. 6. Conclusion We use ES data to show that during the COVID-19 crisis, economic activity was reallocated toward more productive firms. More productive firms were also more likely to adapt to the crisis by increasing remote work. These findings suggest that the COVID-19 crisis has resulted in “creative destruction.” However, as of June 2021, even highly productive firms reported losses in sales and employees (although these were smaller in magnitude than for low-productivity firms). Thus, the crisis will have a silver lining only if the release of resources by low-productivity firms is followed by increased growth in more productive firms during the recovery (Caballero and Hammour 2005). We find that reallocation from low- to high-productivity firms was greater in countries with more competitive markets. This result is consistent with the previous literature showing that competitive markets are better at allocating resources toward more productive firms (Arnold, Nicoletti, and Scarpetta 2011; Brown and Earle 2002; Caballero 2008). We also find that reallocation of sales and employment from low- to high-productivity firms during the COVID-19 crisis was stronger compared with pre-crisis times. This finding is subject to the caveat that there are limits to the comparability of the growth rates of employment for the two periods. Nevertheless, these results contrast with previous studies that find that reallocation from less to more productive firms during recent (non-COVID-19) crises was weaker than during non-crisis times (Foster, Grim, and Haltiwanger 2016; Hallward-Driemeier and Rjikers 2013). 14 Many governments implemented broad policy support schemes to address the economic fall-out from the COVID-19 crisis promptly and provide immediate relief to protect firms and workers from the worst effects. These government support measures and some exemptions to competition law and policies may be appropriate if they are temporary and designed to minimize the disruptive effects they can have on competition (Akcigit and others 2021; Pop and Amador 2020). Our analysis suggests that the government support measures implemented in ECA may have adverse effects on competition and productivity growth since they went to less productive and larger firms, regardless of their pre-crisis innovation. Government policy measures aimed at supporting firms through the economic fall-out of the pandemic have also reignited concerns about propping up so-called “zombie firms”—firms that are unprofitable but remain in operation only because of cheap credit and debt forbearance (ECB 2021). Zombie firms may reduce economywide productivity by crowding out resources for new, more productive firms. In the wake of the global financial crisis, low interest rates and weak insolvency frameworks contributed to a significant increase in the share of zombie firms in several European countries (Acharya and others 2020). Government support measures enacted in response to the pandemic may have further increased that share. As economies enter the economic recovery phase, it will be important for policy makers to phase out policy support measures as soon as appropriate and focus on fostering a competitive business environment while continuing to protect vulnerable households. Such an environment is key to a strong recovery, resilience to future crises, and sustainable, long-term economic growth. 15 References Acharya, Viral, Matteo Crosignani, Tim Eisert, and Christian Eufinger. 2020. “Zombie Credit and (Dis)inflation: Evidence from Europe.” NBER Working Paper 27158, National Bureau of Economic Research, Cambridge, MA. Aga, Gemechu, and Hibret Maemir. 2021. “COVID-19 and African Firms: Impact and Coping Strategies.” Policy Research Working Paper 9642, World Bank, Washington, DC, https://openknowledge.worldbank.org/handle/10986/35519. Akcigit, Ufuk, Wenjie Chen, Federico J. Diez, Romain A. Duval, Philipp Engler, Jiayue Fan, Chiara Maggi, Marina Mendes Tavares, Daniel A. Schwarz, Ippei Shibata, and Carolina Villegas-Sánchez. 2021. “Rising Corporate Market Power Emerging Policy Issues.” IMF Staff Discussion Note 2021/001, International Monetary Fund, Washington, DC. Albagli, Elías, Andrés Fernández, and Federico Huneeus. 2021. “Firms’ Margins of Adjustment and Public Policies during COVID-19: A Research Agenda.” Presentation at “COVID-19: Economic Implications and Policy Lessons,” Central Bank of Chile, Santiago, January 13. Anginer, Deniz, Asli Demirgüç-Kunt, and Davide S. Mare. 2020. “Bank Capital and Risk in Europe and Central Asia Ten Years after the Crisis.” Policy Research Working Paper 9138, World Bank, Washington, DC. Apedo-Amah, Marie Christine, Besart Avdiu, Xavier Cirera, Marcio Cruz, Elwyn Davies, Arti Grover, Leonardo Iacovone, Umut Kilin, Denis Medvedev, Franklin Okechukwu Maduko, Stavros Poupakis, Jesica Torres, and Trang Thu Tran. 2020. “Unmasking the Impact of COVID-19 on Businesses: Firm Level Evidence from across the World.” Policy Research Working Paper 9434, World Bank, Washington, DC, https://openknowledge.worldbank.org/handle/10986/34626. Arnold, Jens, Giuseppe Nicoletti, and Stefano Scarpetta. 2011. “Regulation, Resource Reallocation and Productivity Growth.” EIB Papers 16 (1): 90–115 (European Investment Bank, Luxembourg). Ayyagari, Meghana, Asli Demirgüç-Kunt, and Vojislav Maksimovic. 2020. "Are Large Firms Born or Made? Evidence from Developing Countries." Small Business Economics 57: 191–219. Balleer, Almut, Sebastian Link, Manuel Menkhoff, and Peter Zorn. 2020. “Demand or Supply? Price Adjustment during the Covid-19 Pandemic.” CESifo Working Paper Series 8394, Center for Economic Studies, University of Munich, Germany, https://ideas.repec.org/p/ces/ceswps/_8394.html. Barrero, Jose Maria, Nicholas Bloom, and Steven J. Davis. 2020. “COVID-19 Is Also a Reallocation Shock.” Working Paper 2020-59, Becker-Friedman Institute for Economics, University of Chicago. Bartik, Alexander W., Marianne Bertrand, Zoe Cullen, Edward L. Glaeser, Michael Luca, and Christopher Stanton. 2020. “The Impact of COVID-19 on Small Business Outcomes and Expectations.” Proceedings of the National Academy of Sciences 117 (30): 17656–66, doi:10.1073/pnas.2006991117. Bennedsen, Morten, Birthe Larsen, Ian Schmutte, and Daniela Scur. 2020. "Preserving Job Matches during the COVID-19 Pandemic: Firm-Level Evidence on the Role of Government Aid." Covid Economics: Vetted and Real-Time Papers 27, Center for Economic and Policy Research, Washington, DC. 16 Bertelsmann Stiftung Transformation Index (BTI). 2020. “BTI 2020 Codebook for Country Assessments.” https://bti-project.org/en/methodology#Index%20aggregation. Bloom, Nicholas, Philip Bunn, Paul Mizen, Pawel Smietanka, and Gregory Thwaites. 2021. “The Impact of Covid-19 on Productivity.” NBER Working Paper 28233, National Bureau of Economic Research, Cambridge, MA. Bosio, Erica, Simeon Djankov, Filip Jolevski, and Rita Ramalho. 2020. “Survival of Firms during Economic Crisis.” Policy Research Working Paper 9239, World Bank, Washington, DC, http://documents.worldbank.org/curated/en/140091589221773664/Survival-of-Firms-during- Economic-Crisis. Brown, J. David, and John S. Earle. 2002. "Gross Job Flows in Russian Industry Before and After Reforms: Has Destruction Become More Creative?" Journal of Comparative Economics 30 (1): 96–133. Brucal, Arlan, Arti Grover, and Santiago Reyes Ortega. 2021. “Damaged by the Disaster: The Impact of COVID-19 on Firms in South Asia.” Policy Research Working Paper 9604, World Bank, Washington, DC, https://openknowledge.worldbank.org/handle/10986/35357. Buchheim, Lukas, Carla Krolage, and Sebastian Link. 2020. “Sudden Stop: When Did Firms Anticipate the Potential Consequences of COVID-19?” IZA Discussion Papers 13457, Institute of Labor Economics, Bonn, Germany. Caballero, Ricardo J. 2008. “Creative Destruction.” In The New Palgrave Dictionary of Economics, second ed., edited by Steven N. Durlauf and Lawrence E. Blume. New York: Palgrave Macmillan. Caballero, Ricardo J., and Mohamad L. Hammour. 1994. “The Cleansing Effect of Recessions.” American Economic Review 84 (5): 1350–68. ———. 2005. “The Cost of Recessions Revisited: A Reverse-Liquidationist View.” Review of Economic Studies 72 (251): 313–41. Céspedes, Luis Felipe, Roberto Chang, and Andrés Velasco. 2021. “Credit, Employment, and the COVID Crisis.” LSE Public Policy Review 1 (4): 3, https://doi.org/10.31389/lseppr.28. Cirera, Xavier, Marcio Cruz, Elwyn Davies, Arti Grover, Leonardo Iacovone, Jose Ernesto Lopez Cordova, Denis Medvedev, Franklin Okechukwu Maduko, Gaurav Nayyar, Santiago Reyes Ortega, and Jesica Torres. 2021. “Policies to Support Businesses through the COVID-19 Shock: A Firm Level Perspective.” World Bank Research Observer 36 (1): 41–66, https://doi.org/10.1093/wbro/lkab001. Crane, Leland D., Ryan A. Decker, Aaron Flaaen, Adrian Hamins-Puertolas, and Christopher Kurz. 2021. “Business Exit During the COVID-19 Pandemic: Non-Traditional Measures in Historical Context.” Finance and Economics Discussion Series 2020-089r1, Board of Governors of the Federal Reserve System, Washington, DC, https://doi.org/10.17016/FEDS.2020.089r1. De Nicola, Francesca, Aaditya Mattoo, Jonathan David Timmis, and Trang Thu Tran. 2021. “Productivity in the Time of COVID-19: Evidence from East Asia and Pacific.” Research & Policy Brief 46, World Bank, Washington, DC, http://documents.worldbank.org/curated/en/577501619186435038/Productivity-in-the-Time-of- COVID-19-Evidence-from-East-Asia-and-Pacific. Demirgüç-Kunt, Asli, Michael Lokshin, and Ivan Torre. 2021. “The Sooner, the Better: The Economic Impact of Non-Pharmaceutical Interventions during the Early Stage of the COVID-19 Pandemic.” Economics of Transition and Institutional Change 29: 551–73, https://doi.org/10.1111/ecot.12284. Di Mauro, Filippo, and Chad Syverson. 2020. “The COVID Crisis and Productivity Growth.” Technical report, VOXEU, April, Center for Economic and Policy Research, Washington, DC, https://voxeu.org/article/covid-crisis-and-productivity-growth. 17 ECB (European Central Bank). 2021. Financial Stability Review, May. Frankfurt am Main. Eslava, Marcela, Arturo Galindo, Marc Hofstetter, and Alejandro Izquierdo. 2010. "Scarring Recessions and Credit Constraints: Evidence from Colombian Plant Dynamics." Documento CEDE No. 2010-27, Universidad de los Andes–CEDE, Bogotá, Colombia, http://dx.doi.org/10.2139/ssrn.1705524. Fairlie, Robert. 2020. “The Impact of COVID-19 on Small Business Owners: Evidence from the First Three Months after Widespread Social-Distancing Restrictions.” Journal of Economics & Management Strategy 29: 727– 40, https://doi.org/10.1111/jems.12400. Foster, Lucia, Cheryl Grim, and John Haltiwanger. 2016. “Reallocation in the Great Recession: Cleansing or Not?” Journal of Labor Economics 34 (S1): S293–S331. Francis, David, Sahar Hussain, and Marc Schiffbauer. 2018. “Do Politically Connected Firms Innovate, Contributing to Long-Term Economic Growth?” Policy Research Working Paper 8502, World Bank, Washington, DC, https://openknowledge.worldbank.org/handle/10986/29975. Gourinchas, Pierre-Olivier, Ṣebnem Kalemli-Ozcan, Veronika Penciakova, and Nick Sander. 2021. “COVID-19 and SME Failures.” NBER Working Paper 27877, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w27877. Hallward-Driemeier, Mary, and Bob Rijkers. 2013. “Do Crises Catalyze Creative Destruction? Firm-level Evidence from Indonesia.” Review of Economics and Statistics 95 (5): 1788–1810. Haltiwanger, John, Ron S. Jarmin, and Javier Miranda. 2013. “Who Creates Jobs? Small versus Large versus Young.” Review of Economics and Statistics 95 (2): 347–61. Karalashvili, Nona, and Domenico Viganola. 2021. “The Evolving Effect of COVID-19 on the Private Sector.” World Bank, Washington, DC, https://openknowledge.worldbank.org/handle/10986/35869. Kozeniauskas, Nicholas, Pedro Moreira, and Cezar Santos. 2020. “COVID-19 and Firms: Productivity and Government Policies.” Centre for Economic Policy Research, London, https://cepr.org/active/publications/discussion_papers/dp.php?dpno=15156. Liu, Yu, Siqi Wei, and Jian Xu. 2021. “COVID-19 and Women-Led Businesses around the World.” Finance Research Letters, https://doi.org/10.1016/j.frl.2021.102012. Muzi, Silvia, Filip Jolevski, Kohei Ueda, and Domenico Viganola. 2021. “Productivity and Firm Exit during the COVID-19 Crisis: Cross-Country Evidence.” Policy Research Working Paper 9671, World Bank, Washington, DC, http://documents.worldbank.org/curated/en/267581621869749674/Productivity- and-Firm-Exit-during-the-COVID-19-Crisis-Cross-Country-Evidence. Pop, Georgiana, and Ana Amador. 2020. “To Aid and How to Aid: Policy Options to Preserve Markets.” European State Aid Law Quarterly 19 (2): 127–36. Schumpeter, Joseph. 1942. Capitalism, Socialism, and Democracy. New York: Harper & Bros. Wagner, Joachim. 2021. "With a Little Help from My Website, Firm Survival and Web Presence in Times of COVID-19: Evidence from 10 European Countries.” Working Paper Series in Economics No. 399, University of Lüneburg, Lower Saxony, Germany. World Bank. 2012. Global Financial Development Report 2013: Rethinking the Role of the State in Finance. Washington, DC: World Bank. ———. 2020a. Global Economic Prospects, June. Washington, DC: World Bank. ———. 2020b. Worldwide Bureaucracy Indicators Version 1.1. Washington, DC: World Bank, https://databank.worldbank.org/source/worldwide-bureaucracy-indicators-(wwbi). ———. 2021a. “Firm Level Productivity Estimates.” Methodological Note. World Bank, Washington, DC. 18 ———. 2021b. Supporting Firms in Restructuring and Recovery. EFI Insight: Finance, Competitiveness and Innovation. Washington, DC: World Bank. 19 Figure 1: Firm Performance, Labor Productivity, and BTI Market Organization Index a. Percentage change in monthly sales compared b. Percentage change in permanent, full-time with one year ago workers since December 2019 10 10 0 Avg % change in employment 0 Avg % change in sales -10 -10 -20 -20 -30 -40 -30 8 9 10 11 12 13 8 9 10 11 12 13 Log(labor productivity) Log(labor productivity) High market competition Low market competition High market competition Low market competition c. Percent of firms that decreased number of d. Percent of firms that anticipate falling into permanent workers since December 2019 arrears on outstanding liabilities 50 % of firms expect to fall in arrears 50 40 % of firms decreased employment 40 30 30 20 20 10 8 9 10 11 12 13 10 Log(labor productivity) 8 9 10 11 12 13 Log(labor productivity) High market competition Low market competition High market competition Low market competition e. Percent of firms that started or increased online f. Percent of firms that started or increased remote business activity work 40 50 % of firms increased online activity % of firms increased remote work 35 40 30 30 25 20 20 15 10 8 9 10 11 12 13 8 9 10 11 12 13 Log(labor productivity) Log(labor productivity) High market competition Low market competition High market competition Low market competition Sources: Authors’ calculations based on the most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; Bertelsmann Stiftung Transformation Index (BTI) 2020. Note: Low or high market competition is defined as having a BTI market organization rating below or above the median across countries. BTI market organization is based on responses to the question: “To what level have the fundamentals of market-based competition developed?” The figure shows average values in 30 percentiles of log(labor productivity)—that is, 20 the average y-value plotted against the average productivity in a bin/percentile range of productivity. The plots control for number of employees, firm age, gender of the top manager, innovation, state ownership, foreign ownership, access to credit or loan, ownership of a website, location of the main market (local, national, or international), sector, and country fixed effects—that is, both the y- and x-axis variables are residuals (with the mean added back in). The lines are estimated using the underlying data, not binned data, equivalent to estimating equation 1, but with high/low competition dummies (two interaction terms with productivity) instead of a continuous variable interaction term. 21 Table 1: Summary Statistics from Enterprise Surveys Data Variable Obs Mean SD p10 p90 Min Max a. Enterprise Surveys COVID-19 Follow-up Percentage change in monthly sales relative to one year earlier 7,851 -23.98 33.38 -70.00 0.00 -100.00 300.00 Percentage change in number of permanent full-time workers since December 2019 6,860 -12.73 42.35 -60.87 13.33 -200.00 200.00 Reduced number of permanent full-time workers since December 2019 7,855 34.28 Anticipate falling into arrears on outstanding liabilities in the next 6 months 7,497 25.77 Started or increased online business activity 8,168 27.91 Started or increased remote work arrangements for workforce 8,078 34.49 Received government support 8,124 44.75 Type of government support received Cash transfer 8,093 10.93 Payment deferrals 8,088 7.92 New credit 8,084 4.84 Fiscal relief 8,081 8.98 Wage subsidies 8,108 36.45 b. Enterprise Surveys Baseline Log(labor productivity) 8,234 10.51 1.27 8.87 12.04 3.43 16.45 Log(number of employees) 8,234 2.65 1.03 1.61 4.09 0.00 8.62 Log(firm age) 8,234 2.79 0.59 1.95 3.37 0.69 4.61 Top manager female 8,234 0.21 Innovated during 2017–19 8,234 0.40 State ownership (10%) 8,234 0.01 Foreign ownership (10%) 8,234 0.08 Has line of credit or loan 8,234 0.40 Owns a website 8,234 0.65 Main market Local 8,234 0.44 National 8,234 0.44 22 International 8,234 0.12 Sector Manufacturing 8,234 0.26 Retail 8,234 0.21 Other services 8,234 0.53 c. Enterprise Surveys 2017 data Log(labor productivity) 7,383 10.47 1.27 8.83 11.98 1.88 15.78 Percentage change in sales (2017 to 2018) 7,223 9.21 25.31 -11.40 29.33 -66.59 199.60 Percentage change in employment (2017 to 2018) 6,522 4.07 16.02 -11.76 22.22 -61.22 182.09 Table 2: Market Organization, Competition Policy, Public Sector Employment, Share of Public Banks, and GDP per Capita Variable Obs Mean SD Min Max BTI market organization 23 7.26 1.79 4.00 10.00 BTI market organization (2017) 23 7.30 1.69 4.00 10.00 BTI competition policy 23 7.78 1.88 4.00 10.00 Public sector employment share 23 23.56 6.44 13.50 39.30 Public banks share 23 11.97 20.23 0.00 66.00 Log(GDP per capita) 23 9.25 0.60 8.41 10.16 Sources: Bertelsmann Stiftung Transformation Index (BTI) 2018, 2020; Worldwide Bureaucracy Indicators from World Bank 2020b; local sources where Worldwide Bureaucracy Indicators were not available or outdated (Armenia, Azerbaijan, Belarus, Montenegro, North Macedonia, and the Russian Federation); Anginer, Demirgüç-Kunt, and Mare 2020; World Bank World Development Indicators. 23 Table 3: Firm Performance, Labor Productivity, and BTI Market Organization Percentage Percentage Anticipate Increased Increased Decreased change in change in falling into online remote employment sales employment arrears activity work Log(labor productivity) 4.138*** 2.776*** -2.964*** -3.572*** 0.443 3.402** (0.547) (0.809) (0.853) (0.822) (0.683) (1.375) Log(labor productivity)*BTI 0.835** 0.040 -1.722*** -1.200*** 0.266 1.338* market organization (0.364) (0.530) (0.450) (0.402) (0.334) (0.655) Log(number of employees) 3.117*** 1.907** 2.994** -1.740*** -0.099 5.578*** (0.598) (0.876) (1.179) (0.585) (1.032) (0.895) Log(firm age) 0.545 1.685 -1.877 -4.089* -2.214** 0.977 (1.708) (1.519) (1.701) (2.355) (1.006) (1.631) Top manager female dummy -7.507*** -4.356** 4.057* 2.618 1.176 -0.102 (2.020) (2.089) (2.294) (2.587) (3.511) (2.193) Innovated during 2017–19 1.297 0.419 -1.538 -0.490 8.733*** 7.841*** (1.466) (2.774) (2.736) (1.645) (2.334) (1.909) State ownership (10%) 7.858* 2.089 -4.971 7.744 -7.915** -22.119** (4.307) (2.707) (7.352) (4.664) (3.317) (8.504) Foreign ownership (10%) 1.704 3.510 -2.722 -4.409 -0.337 11.303*** (2.161) (3.036) (3.574) (2.930) (3.311) (2.677) Has line of credit or loan -0.519 0.333 2.340 3.163 3.524** -0.549 (1.752) (1.965) (1.647) (1.993) (1.683) (1.747) Owns a website -2.398 0.294 -1.240 -0.296 12.394*** 8.361*** (1.702) (2.223) (1.972) (1.532) (1.961) (1.918) National market dummy 5.164*** 3.760** -2.159 0.781 6.250*** 10.793*** (1.271) (1.387) (1.874) (2.317) (2.154) (1.916) International market dummy 3.179 1.937 1.956 -5.208* 0.634 8.104** (2.501) (1.823) (3.237) (2.716) (2.512) (3.187) Constant -75.127*** -53.345*** 66.000*** 79.258*** 6.569 -40.074** (6.028) (9.875) (10.198) (10.357) (7.059) (14.710) R2 0.136 0.117 0.069 0.110 0.108 0.183 Number of observations 7,851 6,860 7,855 7,497 8,168 8,078 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; Bertelsmann Stiftung Transformation Index (BTI) 2020. Note: All regressions are ordinary least squares and include sector and country fixed effects. BTI market organization is based on responses to the question: "To what level have the fundamentals of market-based competition developed?" Change in employment is not available for the Russian Federation. BTI market organization is centered on its mean to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 24 Table 4: Firm Performance and Labor Productivity Pre-COVID-19 Percentage Percentage change in change in sales employment Panel A: Pre-COVID-19 (2017 to 2018) Log(labor productivity) -4.568*** 1.841*** (0.876) (0.381) Log(labor productivity)*BTI 0.411 0.164 market organization (0.500) (0.243) Log(number of employees) 0.262 -1.849*** (0.476) (0.407) Log(firm age) -7.926*** -6.063*** (1.319) (0.758) Constant 76.259*** 6.602 (9.133) (4.451) R2 0.096 0.109 Number of observations 7,223 6,522 Panel B: During COVID-19 (2019 to 2020/2021) Log(labor productivity) 4.540*** 3.296*** (0.579) (0.896) Log(labor productivity)*BTI 0.886** 0.101 market organization (0.376) (0.504) Log(number of employees) 3.643*** 2.549*** (0.600) (0.786) Log(firm age) 0.545 1.418 (1.707) (1.568) Constant -80.242*** -57.451*** (5.932) (8.968) R2 0.120 0.112 Number of observations 8,002 6,983 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; Bertelsmann Stiftung Transformation Index (BTI) 2018, 2020. Note: All regressions are ordinary least squares and include sector and country fixed effects. BTI competition policy is based on the question: "To what extent do safeguards exist to protect competition, and to what extent are they enforced?" Change in employment is not available for the Russian Federation. BTI competition policy is centered on its mean to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. Panel B replicates the results from table 3 with the reduced set of controls that is available for the earlier period shown in panel A. *** p < 0.01, ** p < 0.05, * p < 0.1. 25 Table 5: Government Assistance and Firm Characteristics Cash Payment New Fiscal Wage Any type transfers deferrals credit relief subsidies Log(labor productivity) -2.469*** -1.260** -0.583 0.219 -0.649 -2.900*** (0.845) (0.562) (0.412) (0.447) (0.422) (0.855) Log(number of employees) 0.659 -0.603 0.976** -0.394 1.549** 1.296 (1.514) (0.503) (0.382) (0.382) (0.699) (1.382) Log(firm age) -0.905 -0.964 -1.353 0.112 -1.257 -1.200 (1.663) (1.037) (1.010) (0.583) (0.781) (1.960) Top manager female dummy 5.198** 1.926 0.538 -1.035 1.628 5.745** (2.426) (1.892) (1.500) (0.685) (2.060) (2.516) Innovated during 2017–19 -0.472 0.259 -1.355 0.824 0.246 -0.238 (1.777) (0.975) (1.309) (0.934) (1.174) (1.652) State ownership (10%) 5.245 0.351 5.339 -1.019 -1.819 -3.473 (5.002) (2.324) (9.908) (0.968) (1.557) (8.885) Foreign ownership (10%) -2.113 -3.874*** -0.136 -0.305 1.990 0.979 (2.917) (1.219) (1.157) (1.302) (1.721) (3.939) Has line of credit or loan 4.279* 2.030 3.924*** 3.483*** 0.662 3.922* (2.126) (1.369) (1.381) (1.199) (1.409) (2.013) Owns a website 5.559** 0.578 1.472 0.118 0.345 6.185*** (2.225) (1.290) (0.900) (1.088) (1.373) (1.828) National market dummy -1.382 -1.216 -0.372 1.898*** 0.610 -1.932 (1.858) (0.741) (0.802) (0.553) (1.534) (2.114) International market dummy 1.322 0.639 -0.633 1.302 1.180 -4.304* (2.556) (1.654) (2.056) (1.514) (3.034) (2.495) Constant 65.864*** 26.743*** 13.368*** 1.198 11.977* 61.401*** (8.624) (6.368) (3.746) (5.322) (6.267) (9.777) R2 0.215 0.173 0.059 0.055 0.130 0.224 Number of observations 8,124 8,093 8,088 8,084 8,081 8,108 Source: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia. Note: All regressions are ordinary least squares and include sector and country fixed effects. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 26 Appendix Table A.1: Completion Dates of Survey Fieldwork for the World Bank Enterprise Surveys ES COVID-19 ES COVID-19 Country ES Baseline round 1 round 2 a. Baseline Sample Albania May 2019 June 2020 — Armenia December 2020 April 2021 — Azerbaijan December 2019 May 2021 — Belarus April 2019 August 2020 — Bosnia and Herzegovina September 2019 March 2021 — Bulgaria March 2020 September 2020 December 2020 Croatia November 2019 September 2020 January 2021 Czech Republic March 2020 October 2020 February 2021 Estonia January 2020 November 2020 February 2021 Georgia January 2020 June 2020 November 2020 Hungary March 2020 September 2020 February 2021 Kazakhstan October 2019 March 2021 — Latvia January 2020 November 2020 February 2021 Lithuania January 2020 October 2020 February 2021 Moldova November 2019 May 2020 November 2020 Montenegro July 2019 February 2021 — North Macedonia October 2019 November 2020 — Poland December 2019 August 2020 December 2020 Romania June 2020 September 2020 December 2020 Russian Federation July 2019 June 2020 — Serbia October 2019 February 2021 — Slovak Republic March 2020 October 2020 February 2021 Slovenia November 2019 August 2020 December 2020 b. Expanded Sample Chad May 2019 July 2020 — Jordan November 2019 August 2020 January 2021 Lebanon April 2020 December 2020 — Mongolia May 2019 August 2020 February 2021 Morocco January 2020 August 2020 February 2021 Mozambique January 2019 January 2021 — Zambia March 2020 July 2020 February 2021 Note: — Not available. Dates shown are for the last completed survey for each round in a country. 27 Table A.2: Variable Definitions Variable Definition Percentage change in monthly Percentage change in monthly sales compared to the same month in 2019, as sales relative to one year earlier directly reported by firms. Percentage change in number of Percentage change in permanent full-time workers since December 2019. The permanent full-time workers formula is: ((a1-a0)/[(a1+a0)/2])*100; where a1 = permanent full-time since December 2019 workers, end of last completed month, and a0 = permanent full-time workers, end of December 2019. Reduced number of permanent Equals 100 if the firm reduced number of permanent full-time workers since full-time workers since December 2019 and 0 otherwise. December 2019 Anticipate falling into arrears on Equals 100 if the firm anticipates falling into arrears on outstanding liabilities outstanding liabilities in the next in the next 6 months and 0 otherwise. 6 months Started or increased online Equals 100 if the firm started or increased online business activity and 0 business activity otherwise. Started or increased remote Equals 100 if the firm started or increased remote work arrangements for work arrangements for workforce and 0 otherwise. workforce Received government support Equals 100 if the firm received any national or local government assistance provided in response to the COVID-19 outbreak and 0 otherwise. Cash transfer Equals 100 if the government support received by the firm was a cash transfer and 0 otherwise. Payment deferrals Equals 100 if the government support received by the firm was a payment deferral and 0 otherwise. New Credit Equals 100 if the government support received by the firm was new credit and 0 otherwise. Fiscal relief Equals 100 if the government support received by the firm was fiscal relief and 0 otherwise. Wage subsidies Equals 100 if the government support received by the firm was wage subsidies and 0 otherwise. 28 Log(labor productivity) Log of annual sales divided by the number of full-time permanent employees (2009 US$). Log(number of employees) Log of total number of full-time employees. Log(firm age) Log of number of years the firm has been operating. Top manager female Equals 1 if the firm's top manager is female and 0 otherwise. Innovated during 2017–19 Equals 1 if the firm innovated a product or process in the last 3 years and 0 otherwise. State ownership (10%) Equals 1 if the firm is at least 10% state owned and 0 otherwise. Foreign ownership (10%) Equals 1 if the firm is at least 10% foreign owned and 0 otherwise. Has line of credit or loan Equals 1 if the firm has line of credit or bank loans and 0 otherwise. Owns a website Equals 1 if the firm has its own website and 0 otherwise. Local market Equals 1 if the firm's main product is sold mostly in same municipality where the firm is located and 0 otherwise. National market Equals 1 if the firm's main product is sold mostly across the country where the firm is located and 0 otherwise. International market Equals 1 if the firm's main product is sold mostly internationally and 0 otherwise. Manufacturing Equals 1 if the firm is in the manufacturing sector and 0 otherwise. Retail Equals 1 if the firm is in the retail sector and 0 otherwise. Other services Equals 1 if the firm is in the selected service sector, excluding retail, and 0 otherwise. Percentage change in sales Percentage change in sales from 2017 to 2018. The formula is ((a1- (2017 to 2018) a0)/[(a1+a0)/2])*100, where a1 = sales in 2018, extrapolated as the simple average of 2019 and 2017 sales, and a0 = sales in 2017. Percentage change in Percentage change in permanent full-time workers from 2017 to 2018. The employment (2017 to 2018) formula is ((a1-a0)/[(a1+a0)/2])*100, where a1 = permanent full-time workers 29 in 2018, extrapolated as the simple average of 2019 and 2017 sales, and a0 = permanent full-time workers in 2017. 30 Table A.3: Firm Performance, Labor Productivity, and BTI Competition Policy Percentage Percentage Anticipate Increased Increased Decreased change in change in falling into online remote employment sales employment arrears activity work Log(labor productivity) 4.115*** 2.736*** -2.992*** -3.705*** 0.400 3.559*** (0.635) (0.946) (0.967) (0.827) (0.647) (1.145) Log(labor productivity)*BTI 0.542 -0.105 -1.272* -1.116** 0.073 1.337 competition policy (0.397) (0.655) (0.659) (0.440) (0.378) (0.923) Log(number of employees) 3.108*** 1.918** 3.022** -1.710*** -0.096 5.535*** (0.599) (0.898) (1.183) (0.598) (1.030) (0.892) Log(firm age) 0.484 1.714 -1.720 -3.943 -2.210** 0.756 (1.726) (1.563) (1.698) (2.354) (0.987) (1.640) Top manager female dummy -7.581*** -4.369** 4.190* 2.692 1.148 -0.191 (2.028) (2.084) (2.317) (2.582) (3.510) (2.207) Innovated during 2017–19 1.266 0.402 -1.474 -0.489 8.713*** 7.825*** (1.479) (2.788) (2.744) (1.637) (2.328) (1.914) State ownership (10%) 8.121* 2.020 -5.589 7.116 -7.880** -21.435** (4.345) (2.692) (7.314) (4.622) (3.370) (8.667) Foreign ownership (10%) 1.760 3.580 -2.770 -4.244 -0.262 11.105*** (2.171) (3.177) (3.534) (2.991) (3.328) (2.698) Has line of credit or loan -0.428 0.338 2.190 3.081 3.553** -0.436 (1.742) (1.969) (1.649) (1.977) (1.667) (1.717) Owns a website -2.443 0.293 -1.162 -0.246 12.381*** 8.309*** (1.702) (2.205) (1.980) (1.546) (1.973) (1.907) National market dummy 5.143*** 3.791** -2.107 0.905 6.266*** 10.660*** (1.278) (1.383) (1.860) (2.292) (2.171) (1.881) International market dummy 3.248 1.949 1.806 -5.271* 0.665 8.183** (2.502) (1.825) (3.229) (2.682) (2.509) (3.192) Constant -74.550*** -52.717*** 65.459*** 80.234*** 7.208 -41.321*** (7.037) (10.896) (12.499) (10.614) (6.737) (12.992) R2 0.135 0.117 0.068 0.110 0.108 0.183 Number of observations 7,851 6,860 7,855 7,497 8,168 8,078 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; Bertelsmann Stiftung Transformation Index (BTI) 2020. Note: All regressions are ordinary least squares and include sector and country fixed effects. BTI competition policy is based on the question: "To what extent do safeguards exist to protect competition, and to what extent are they enforced?" Change in employment is not available for the Russian Federation. BTI competition policy is centered on its mean to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 31 Table A.4: Firm Performance, Labor Productivity, and Public Sector Employment Share Percentage Percentage Anticipate Increased Decreased Increased change in change in falling into online employment remote work sales employment arrears activity Log(labor productivity) 4.007*** 2.763*** -2.636*** -3.382*** 0.416 3.337*** (0.573) (0.750) (0.915) (0.795) (0.686) (1.053) Log(labor productivity)*Public -0.149** 0.073 0.185** 0.164* -0.069 -0.460** employment share (0.057) (0.051) (0.084) (0.095) (0.082) (0.203) Log(number of employees) 3.132*** 1.918** 2.936** -1.773*** -0.098 5.577*** (0.604) (0.869) (1.188) (0.585) (1.032) (0.904) Log(firm age) 0.614 1.696 -1.991 -4.210* -2.193** 1.092 (1.718) (1.484) (1.712) (2.377) (1.008) (1.611) Top manager female dummy -7.583*** -4.390** 4.206* 2.705 1.174 -0.069 (2.049) (2.072) (2.317) (2.596) (3.508) (2.189) Innovated during 2017–19 1.259 0.384 -1.402 -0.437 8.733*** 7.902*** (1.466) (2.738) (2.707) (1.633) (2.335) (1.894) State ownership (10%) 7.788* 2.080 -4.842 7.818 -7.938** -22.283** (4.339) (2.692) (7.339) (4.667) (3.316) (8.540) Foreign ownership (10%) 1.962 3.546 -3.364 -4.804 -0.262 11.620*** (2.170) (2.966) (3.546) (2.904) (3.299) (2.702) Has line of credit or loan -0.552 0.389 2.310 3.171 3.498** -0.788 (1.747) (1.978) (1.664) (1.971) (1.679) (1.682) Owns a website -2.468 0.306 -1.145 -0.177 12.365*** 8.188*** (1.676) (2.210) (2.029) (1.504) (1.985) (1.891) National market dummy 5.240*** 3.798** -2.311 0.727 6.249*** 10.747*** (1.299) (1.371) (1.915) (2.328) (2.152) (1.982) International market dummy 3.252 1.962 1.788 -5.288* 0.646 8.127** (2.506) (1.802) (3.228) (2.701) (2.512) (3.233) Constant -72.816*** -52.938*** 60.574*** 75.813*** 7.166 -37.696*** (6.539) (8.643) (10.209) (10.653) (6.930) (12.255) R2 0.135 0.117 0.066 0.109 0.108 0.186 Number of observations 7,851 6,860 7,855 7,497 8,168 8,078 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; Worldwide Bureaucracy Indicators, World Bank 2020b; local sources where Worldwide Bureaucracy Indicators were not available or outdated (Armenia, Azerbaijan, Belarus, Montenegro, North Macedonia, and the Russian Federation). Note: All regressions are ordinary least squares and include sector and country fixed effects. Change in employment is not available for the Russian Federation. Public sector employment share is centered on its mean to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 32 Table A.5: Firm Performance, Labor Productivity, and Public Banks Share Percentage Percentage Anticipate Increased Increased Decreased change in change in falling into online remote employment sales employment arrears activity work Log(labor productivity) 3.910*** 2.641*** -2.510** -3.236*** 0.371 3.022* (0.604) (0.667) (0.943) (0.885) (0.696) (1.598) Log(labor productivity)*Public -0.041** -0.029 0.054** 0.055** -0.003 -0.016 banks share (0.017) (0.028) (0.026) (0.026) (0.024) (0.046) Log(number of employees) 3.160*** 1.909** 2.903** -1.798*** -0.092 5.623*** (0.597) (0.863) (1.187) (0.572) (1.034) (0.913) Log(firm age) 0.608 1.672 -1.999 -4.200* -2.192** 1.096 (1.714) (1.509) (1.695) (2.384) (1.012) (1.648) Top manager female dummy -7.620*** -4.356** 4.260* 2.766 1.145 -0.251 (2.041) (2.075) (2.316) (2.591) (3.505) (2.185) Innovated during 2017–19 1.228 0.431 -1.355 -0.402 8.708*** 7.728*** (1.457) (2.749) (2.711) (1.641) (2.326) (1.888) State ownership (10%) 7.649* 2.012 -4.644 8.014 -7.929** -22.219** (4.347) (2.716) (7.315) (4.748) (3.352) (8.374) Foreign ownership (10%) 1.862 3.463 -3.216 -4.662 -0.232 11.851*** (2.149) (3.008) (3.537) (2.924) (3.329) (2.789) Has line of credit or loan -0.481 0.332 2.241 3.105 3.552** -0.409 (1.743) (1.979) (1.660) (1.994) (1.659) (1.709) Owns a website -2.459 0.283 -1.150 -0.207 12.379*** 8.272*** (1.717) (2.205) (2.014) (1.530) (1.973) (1.926) National market dummy 5.294*** 3.784** -2.396 0.663 6.282*** 10.969*** (1.308) (1.396) (1.917) (2.328) (2.150) (1.970) International market dummy 3.302 1.964 1.718 -5.336* 0.671 8.282** (2.524) (1.806) (3.221) (2.694) (2.498) (3.189) Constant -72.101*** -52.715*** 59.623*** 74.959*** 7.536 -35.211* (6.148) (8.141) (9.568) (10.985) (7.001) (17.167) R2 0.135 0.117 0.066 0.109 0.108 0.180 Number of observations 7,851 6,860 7,855 7,497 8,168 8,078 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; Anginer, Demirgüç-Kunt, and Mare 2020. Note: All regressions are ordinary least squares and include sector and country fixed effects. Change in employment is not available for the Russian Federation. Public banks share is centered on its mean to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 33 Table A.6: Firm Performance, Labor Productivity, and GDP per Capita Anticipate Percentage Percentage Increased Increased Decreased falling change in change in online remote employment into sales employment activity work arrears Log(labor productivity) 4.080*** 2.778*** -2.862*** -3.436*** 0.338 3.184* (0.669) (0.941) (0.904) (0.949) (0.633) (1.578) Log(labor productivity)* 1.705 0.049 -3.711* -1.899 -0.335 1.527 Log(GDP per capita) (1.077) (1.867) (1.973) (1.427) (1.523) (2.152) Log(number of employees) 3.108*** 1.907** 3.021** -1.746*** -0.085 5.586*** (0.599) (0.887) (1.186) (0.591) (1.030) (0.898) Log(firm age) 0.545 1.687 -1.863 -4.136* -2.177** 1.026 (1.707) (1.519) (1.694) (2.370) (0.992) (1.636) Top manager female dummy -7.602*** -4.361** 4.270* 2.786 1.144 -0.260 (2.021) (2.074) (2.321) (2.576) (3.507) (2.190) Innovated during 2017–19 1.269 0.416 -1.464 -0.458 8.699*** 7.751*** (1.470) (2.785) (2.749) (1.642) (2.322) (1.901) State ownership (10%) 8.040* 2.090 -5.373 7.490 -7.958** -21.961** (4.329) (2.691) (7.382) (4.680) (3.366) (8.464) Foreign ownership (10%) 1.765 3.514 -2.845 -4.617 -0.160 11.649*** (2.183) (3.172) (3.545) (3.018) (3.325) (2.786) Has line of credit or loan -0.410 0.335 2.125 3.042 3.557** -0.388 (1.751) (1.976) (1.660) (1.976) (1.664) (1.715) Owns a website -2.372 0.294 -1.312 -0.283 12.368*** 8.341*** (1.696) (2.252) (2.024) (1.535) (1.972) (1.932) National market dummy 5.200*** 3.762** -2.245 0.741 6.295*** 10.902*** (1.272) (1.389) (1.837) (2.319) (2.166) (1.958) International market dummy 3.209 1.939 1.879 -5.271* 0.687 8.209** (2.502) (1.842) (3.203) (2.697) (2.516) (3.207) Constant -74.479*** -53.296*** 64.884*** 77.609*** 8.024 -37.376** (7.314) (11.136) (12.054) (11.402) (6.567) (17.111) R2 0.135 0.117 0.067 0.108 0.108 0.181 Number of observations 7,851 6,860 7,855 7,497 8,168 8,078 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia; World Bank World Development Indicators. Note: All regressions are ordinary least squares and include sector and country fixed effects. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 34 Table A.7: Firm Performance, Labor Productivity, and BTI Market Organization, Controlling for Severity of COVID-19 Restrictions Percentage Percentage Anticipate Increased Increased Decreased change in change in falling into online remote employment sales employment arrears activity work Log(labor productivity) 4.171*** 2.627*** -2.773*** -3.115*** 0.920 3.771*** (0.630) (0.764) (0.900) (0.708) (0.692) (1.126) Log(labor productivity)*BTI 0.914** 0.366 -1.907*** -1.005** 0.126 0.897 market organization (0.329) (0.572) (0.531) (0.386) (0.385) (0.545) Log(labor productivity)*COVID 0.011 0.027* -0.017 0.020*** -0.006 -0.040 restrictions index (0.012) (0.014) (0.012) (0.005) (0.006) (0.024) Log(number of employees) 2.773*** 1.875* 2.586** -1.632** -0.096 5.425*** (0.620) (0.993) (1.023) (0.701) (1.213) (0.976) Log(firm age) 1.496 2.117 -1.780 -5.235* -1.839 0.926 (1.905) (1.690) (1.888) (2.555) (1.119) (1.868) Top manager female dummy -8.068*** -3.182 1.733 3.368 0.951 -0.352 (2.202) (1.996) (1.989) (2.824) (3.961) (2.260) Innovated during 2017–19 0.359 -0.809 -1.012 0.277 9.026*** 8.439*** (1.468) (2.984) (2.784) (1.781) (2.528) (2.054) State ownership (10%) 7.587 1.997 -4.757 8.051 -8.886** -21.033** (4.651) (2.519) (7.169) (4.768) (3.153) (8.932) Foreign ownership (10%) 0.500 3.629 -2.257 -4.290 -0.010 8.230*** (1.928) (3.459) (3.888) (3.127) (3.437) (2.163) Has line of credit or loan -1.066 -0.522 3.487* 3.900 3.571* -1.078 (1.892) (2.171) (1.703) (2.257) (1.863) (1.973) Owns a website -1.377 -0.141 -0.505 -1.129 11.848*** 9.689*** (1.829) (2.615) (2.268) (1.734) (2.198) (1.769) National market dummy 6.050*** 3.272** -1.152 1.742 5.094** 11.472*** (1.261) (1.360) (1.969) (2.255) (2.392) (2.072) International market dummy 4.302 1.620 2.303 -4.867* -1.072 9.776*** (2.738) (1.961) (3.504) (2.679) (2.587) (3.385) Constant -77.081*** -51.520*** 64.654*** 78.998*** 1.236 -45.035*** (6.352) (9.299) (10.564) (10.782) (7.273) (13.339) R2 0.145 0.127 0.075 0.123 0.112 0.172 Number of observations 7,208 6,220 7,215 6,835 7,458 7,367 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 20 countries in Europe and Central Asia; Bertelsmann Stiftung Transformation Index (BTI) 2020; and Oxford COVID-19 Government Response Tracker’s Stringency Index 2021. Note: All regressions are ordinary least squares and include sector and country fixed effects. BTI market organization is based on responses to the question: "To what level have the fundamentals of market-based competition developed?" The COVID-19 restrictions index is based on the number of days the Oxford COVID-19 Government Response Tracker’s Stringency Index— which captures the strictness of “lockdown-style” policies that primarily restrict people’s behavior along nine dimensions ranging from 0 to 100, with higher values indicating more restrictions —was above 70 preceding the fieldwork start day of the Enterprise Surveys COVID-19 Follow-up Surveys module in each country. The COVID restrictions index is not available for Armenia, Montenegro, and North Macedonia. Change in employment is not available for the Russian Federation. BTI market 35 organization and the COVID restrictions index are centered on their respective means to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 36 Table A.8: Government Assistance and Firm Characteristics, Controlling for Drop in Sales Cash Payment New Fiscal Wage Any type transfers deferrals credit relief subsidies Log(labor productivity) -1.880** -1.118* -0.304 0.220 -0.504 -1.984** (0.763) (0.547) (0.360) (0.476) (0.438) (0.719) Log(number of employees) 1.537 -0.444 1.354*** -0.269 1.710** 2.286 (1.631) (0.462) (0.471) (0.463) (0.750) (1.472) Log(firm age) -0.949 -1.254 -1.562 0.006 -1.097 -1.099 (1.471) (1.176) (1.088) (0.571) (0.813) (1.649) Top manager female dummy 3.734 1.169 -0.013 -1.069 1.702 4.247 (2.587) (1.940) (1.318) (0.630) (2.188) (2.676) Innovated during 2017–19 -0.154 0.487 -1.432 1.154 0.256 -0.233 (1.887) (0.942) (1.202) (0.957) (1.261) (1.842) State ownership (10%) 6.272 0.945 5.188 -0.941 -2.173 -2.324 (4.405) (2.529) (9.976) (0.988) (1.825) (8.829) Foreign ownership (10%) -1.863 -3.772*** -0.652 -0.857 2.137 1.366 (2.738) (1.291) (1.138) (1.397) (1.716) (3.793) Has line of credit or loan 4.349* 1.800 3.880*** 3.390*** 0.509 4.208** (2.098) (1.368) (1.205) (1.177) (1.436) (1.958) Owns a website 5.219** 0.024 0.571 0.052 0.669 5.670*** (2.041) (1.086) (1.144) (1.143) (1.365) (1.756) National market dummy -0.546 -1.099 0.013 1.728*** 0.967 -1.147 (1.750) (0.878) (0.727) (0.485) (1.637) (1.974) International market dummy 1.731 1.143 -0.734 1.392 1.147 -4.294* (2.511) (1.836) (1.909) (1.532) (3.099) (2.264) Average change in monthly sales compared -0.212*** -0.051** -0.058** -0.011 -0.032 -0.257*** to one year ago (0.050) (0.020) (0.023) (0.020) (0.024) (0.048) Constant 52.495*** 24.823*** 9.528** 0.838 8.440 43.602*** (8.782) (6.770) (3.725) (5.444) (7.102) (7.925) R2 0.236 0.171 0.066 0.053 0.135 0.256 Number of observations 7,772 7,748 7,744 7,741 7,736 7,760 Source: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia. Note: All regressions are ordinary least squares and include sector and country fixed effects. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 37 Table A.9: Government Assistance and Firm Characteristics, Controlling for Drop in Employment Cash Payment New Fiscal Wage Any type transfers deferrals credit relief subsidies Log(labor productivity) -2.753** -1.586** -0.644 0.176 -0.561 -3.148*** (0.975) (0.682) (0.498) (0.496) (0.482) (0.919) Log(number of employees) 0.717 -0.578 0.860** -0.313 1.427* 1.294 (1.643) (0.434) (0.376) (0.397) (0.741) (1.519) Log(firm age) -0.959 -1.422 -0.897 0.129 -1.501 -1.031 (1.895) (1.244) (1.082) (0.675) (0.923) (2.106) Top manager female dummy 6.753** 2.399 0.704 -1.006 1.579 6.312** (2.773) (2.143) (1.491) (0.736) (2.267) (2.862) Innovated during 2017–19 -0.566 -0.107 -1.587 1.039 0.334 -0.132 (1.923) (1.258) (1.364) (1.059) (1.219) (1.722) State ownership (10%) 4.920 0.424 5.346 -1.114 -1.080 -4.040 (5.015) (2.370) (10.550) (1.092) (1.699) (9.433) Foreign ownership (10%) -3.836 -3.044*** -0.667 -0.844 1.610 -0.638 (2.966) (0.957) (1.293) (1.185) (1.624) (3.729) Has line of credit or loan 4.512** 2.500 4.099*** 3.590** 0.483 3.624* (2.075) (1.566) (1.255) (1.279) (1.515) (2.058) Owns a website 6.230** 0.343 2.277** 0.241 0.628 6.431*** (2.275) (1.462) (0.959) (1.197) (1.574) (1.781) National market dummy -0.101 -1.378 -0.667 1.745** 0.663 -0.715 (1.841) (0.901) (0.896) (0.653) (1.651) (2.141) International market dummy 1.695 0.035 -1.651 1.236 1.268 -3.991 (2.836) (1.699) (1.826) (1.597) (3.094) (2.532) Percentage change of permanent full-time 0.021 0.028** 0.007 -0.008 -0.012 0.001 workers since December 2019 (0.027) (0.012) (0.011) (0.010) (0.021) (0.015) Constant 69.056*** 32.152*** 13.398*** 1.305 11.888 64.081*** (10.976) (7.883) (4.390) (6.071) (7.754) (10.330) R2 0.206 0.178 0.065 0.055 0.137 0.211 Number of observations 6,792 6,771 6,767 6,765 6,757 6,779 Source: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 23 countries in Europe and Central Asia. Note: All regressions are ordinary least squares and include sector and country fixed effects. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 38 Table A.10: Firm Performance, Labor Productivity, and BTI Market Organization, Expanded Country Sample Percentage Percentage Anticipate Increased Increased Decreased change in change in falling into online remote employment sales employment arrears activity work Log(labor productivity) 3.261*** 2.370*** -2.087*** -3.119*** 0.446 2.410* (0.508) (0.631) (0.701) (0.673) (0.568) (1.264) Log(labor productivity)*BTI 0.930*** 0.322 -1.594*** -1.186*** 0.201 1.215** market organization (0.328) (0.442) (0.407) (0.328) (0.263) (0.444) Log(number of employees) 2.867*** 1.480* 2.850** -2.133*** 0.721 5.154*** (0.556) (0.835) (1.053) (0.688) (1.014) (0.794) Log(firm age) 0.755 1.915 -2.413 -3.328 -1.987** 0.057 (1.582) (1.383) (1.563) (2.162) (0.926) (1.544) Top manager female dummy -6.982*** -2.869 1.934 3.080 1.405 0.016 (1.869) (2.133) (2.581) (2.370) (3.235) (2.001) Innovated during 2017–19 0.925 0.510 -2.013 -0.423 8.442*** 8.075*** (1.351) (2.432) (2.496) (1.623) (2.223) (1.862) State ownership (10%) 8.453** 4.637 -7.707 3.178 -4.040 -12.555 (3.764) (3.818) (6.938) (5.811) (7.070) (13.159) Foreign ownership (10%) 1.918 2.450 -1.030 -3.069 0.442 11.589*** (1.856) (2.625) (3.309) (2.635) (2.806) (2.478) Has line of credit or loan -0.483 -0.527 3.076* 4.114** 3.871** -0.719 (1.623) (1.867) (1.590) (1.888) (1.746) (1.551) Owns a website -2.464 0.530 -1.081 -0.924 10.647*** 7.554*** (1.661) (1.938) (1.775) (1.499) (2.145) (1.785) National market dummy 4.993*** 3.641*** -1.155 0.869 5.228** 9.786*** (1.110) (1.226) (1.776) (2.053) (2.050) (1.839) International market dummy 3.649 2.700 1.270 -5.334** -0.624 7.099** (2.395) (1.762) (3.094) (2.576) (2.465) (3.014) Constant -71.327*** -49.012*** 63.574*** 80.196*** 7.054 -27.248* (5.567) (8.989) (9.527) (9.133) (6.145) (13.650) R2 0.150 0.110 0.064 0.180 0.114 0.191 Number of observations 8,931 8,041 9,036 8,626 9,365 9,600 Sources: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 30 countries; Bertelsmann Stiftung Transformation Index (BTI) 2020. Note: All regressions are ordinary least squares and include sector and country fixed effects. BTI competition policy is based on the question: "To what extent do safeguards exist to protect competition, and to what extent are they enforced?" Change in employment is not available for the Russian Federation. BTI competition policy is centered on its mean to facilitate interpretation of the coefficients. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 39 Table A.11: Government Assistance and Firm Characteristics, Expanded Country Sample Cash Payment New Fiscal Wage Any type transfers deferrals credit relief subsidies Log(labor productivity) -1.847** -0.862 -0.437 0.312 -0.533 -2.387*** (0.810) (0.531) (0.369) (0.454) (0.577) (0.779) Log(number of employees) 0.620 -0.407 0.553 -0.631 1.063 1.090 (1.389) (0.497) (0.590) (0.448) (0.882) (1.247) Log(firm age) -0.944 -0.876 -1.009 0.202 -0.863 -1.552 (1.514) (1.032) (0.982) (0.764) (0.860) (1.791) Top manager female dummy 5.894** 2.698 1.932 0.112 3.274 6.518** (2.323) (1.802) (1.910) (1.336) (2.398) (2.439) Innovated during 2017–19 0.987 0.875 -0.094 1.084 1.325 0.582 (1.982) (1.085) (1.577) (0.909) (1.352) (1.540) State ownership (10%) 2.245 -1.266 4.768 -0.659 -2.734 -5.244 (4.804) (2.656) (8.597) (0.968) (2.760) (8.121) Foreign ownership (10%) -2.751 -3.479*** -0.722 -0.551 2.385 0.331 (2.517) (1.038) (1.135) (1.194) (1.537) (3.334) Has line of credit or loan 2.974 1.829 2.518 2.771** -0.282 3.453* (2.413) (1.383) (1.701) (1.256) (1.491) (1.908) Owns a website 5.653** 0.749 1.157 -0.089 0.543 6.371*** (2.067) (1.160) (0.863) (0.979) (1.372) (1.769) National market dummy -2.733 -2.683** -1.343 0.565 -0.827 -3.028 (1.823) (1.285) (1.031) (0.968) (1.629) (1.957) International market dummy -0.457 -0.800 -1.281 0.753 0.299 -6.069** (2.585) (1.795) (2.029) (1.439) (2.989) (2.477) Constant 58.933*** 22.001*** 13.456*** 1.892 12.103* 56.921*** (8.238) (5.843) (3.671) (4.625) (6.965) (9.102) R2 0.213 0.164 0.063 0.054 0.131 0.222 Number of observations 9,309 9,274 9,259 9,255 9,256 9,285 Source: Most recent Enterprise Surveys COVID-19 Follow-up Surveys and Enterprise Surveys for 30 countries. Note: Regressions include sector and country fixed effects. Standard errors are clustered at the country level. *** p < 0.01, ** p < 0.05, * p < 0.1. 40