Policy Research Working Paper 10580 Does Informality Depress Investments and Job Recovery? Firm-Level Evidence from the COVID-19 Crisis in South Asia Arti Grover Mariana Pereira-López International Finance Corporation & South Asia Region October 2023 Policy Research Working Paper 10580 Abstract Using three rounds of the World Bank’s Business Pulse Sur- expectations on recovery and lower ability to predict future veys in South Asia, this paper quantifies the relationship sales, especially the necessity firms. Third, credit constraints between informality and firms’ investment and employment and accentuated uncertainty among informal firms discour- decisions. Accounting for multidimensionality in definition age investments. Finally, while employment growth is slow and the margins of informality, the analysis suggests that and gradual for formal firms as they begin to recover sales, first, informal firms remain credit and liquidity constrained job growth in informal firms does not correspond to the before and during the crisis, especially the necessity firms. recovery. The results suggest that countries with a large In the pre-crisis period, access to finance is correlated with informal sector may face unusually depressed investments the extensive margin of informality, while during the crisis, and jobs recovery and may have to deploy additional policy both margins of informality matter. Second, informal firms levers to accelerate recovery in the post-crisis period. perceive uncertainty to be higher because of pessimistic This paper is a product of the International Finance Corporation and the Office of the Chief Economist, South Asia Region. 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 agrover1@ifc.org and mpereiralopez@ifc.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 Does Informality Depress Investments and Job Recovery? Firm-Level Evidence from the COVID-19 Crisis in South Asia∗ Arti Grover† & Mariana Pereira-Lopez‡ JEL classification: D22, D25, L20, L25, O17 Keywords: COVID-19, Crisis, Firms, Informality, Jobs ∗ Acknowledgments: The authors are grateful to the South Asia Regional Trade Facilitation Program (SARTFP) for funding support and Denis Medvedev for his strategic guidance during the inception of the project. The authors thank several members of the World Bank teams for their efforts in data collection, including Amila Indeewari Dahanayake, Besart Avdiu, Hosna Ferdous Sumi, Mauricio Alejandro Pinzon Latorre, Mohammad Sulaiman Akbari Ojashwi Samser JB Rana, Peter Mousley, Rafay Khan, Ruchita Mangh- nani and Subika Farazi. The team is grateful to Franklin Maduko, Sarah Hebous, and Reyes Aterido for their careful work in data harmonization. The team thanks Maurizio Bussolo and Siddarth Sharma for their helpful feedback. † International Finance Corporation, World Bank Group, email: agrover1@ifc.org ‡ International Finance Corporation, World Bank Group, email: mpereiralopez@ifc.org 1. Introduction Crises have profound, uneven and long-term effects on aggregate productivity and em- ployment recovery (Reinhart and Rogoff, 2009; Dell’Ariccia et al., 2008). The COVID-19 pandemic, in particular, had an extensive and heterogeneous effect on firms’ operations and performance based on their size, sector, age, and other firm attributes (Apedo-Amah et al., 2020; Cirera et al., 2021). The degree of informality in the context of South Asia has also been a key factor contributing to such differences in effects (Grover and Pereira-López, forthcoming), given its salience in the region (Bussolo et al., 2020). Leveraging three rounds of the World Bank’s Business Pulse Surveys (BPS) in South Asia conducted during the pandemic from May 2020 to August 2022, this paper quantifies the relationship between informality and firms’ investment decisions that may have implications for post-crisis job recovery. Our work suggests that lower investments and employment response among informal firms are potentially attributable to their acute credit constraints and heightened perceived uncertainty. Evidence from the Great Recession shows that shocks to the financial system can explain the trade collapse (Cetorelli and Goldberg, 2011; Chor and Manova, 2012) and output declines, especially in industries that rely more on banking finance. Such effects are stronger among firms with fewer assets that can serve as collateral or those with limited access to inter-firm finance (Bricongne et al., 2012; Görg and Spaliara, 2014; Iacovone et al., 2019; Paravisini et al., 2015). Although the COVID-19 crisis did not originate in the financial sector, the pandemic-induced mobility restrictions led to supply and demand shocks, and the conditions slowly morphed into a financial recession (Reinhart, 2022). With the increase in risks and uncertainty during the crisis, creditors become more risk-averse and are reluctant to lend in the face of imperfect information regarding default likelihood (“flight to quality”) (Dell’Ariccia et al., 2008). In the pandemic period, this resulted in excess demand and a short supply of credit (Didier et al., 2021). Although most crises entail uncertainty, the multiple channels through which the pandemic affected firms and consumers generated unprecedented levels of uncertainty (Meyer et al., 2022) due to evolving variants of the virus and the lockdown measures it provoked (Barrero and Bloom, 2020). Uncertainty effectively increases risk, leading to credit crunch and a higher cost of finance. Under uncertainty, investment can be viewed as "real options," and the value of delaying such decisions that cannot easily be reversed is high. This delays investment, borrowing, and hiring decisions (Bloom et al., 2007; Dell’Ariccia et al., 2008; Barrero and Bloom, 2020). For example, firms might prefer hiring part-time workers, as 2 they can be more easily fired if the effects of the crisis become larger. Finally, uncertainty makes firms less sensitive to demand, prices, and productivity (Bloom, 2014). In the context of COVID-19, for instance, firms showed a weaker response to changes in demand or prices (Barrero and Bloom, 2020). Given the elevated importance of finance and uncertainty aspects during crises, it is critical to understand how these changes affect the decisions on investments and employment ad- justments in developing countries that are fraught with a large informal sector. Informality not only affects the extensive and intensive margins of the shock at the micro-level of the firm, it also impacts aggregate productivity and employment growth at the macro-level by influencing firms’ decisions to invest in fixed assets, the adoption of new technologies, as well as the hiring of new employees (See Ulyssea, 2020). Building on a large body of informality-related research in the region and recent literature on the COVID-19 crisis, we empirically test the interplay between informality and investment decisions. Our work suggests that: First, informal firms are both credit and liquidity constrained. Informal firms experience a lower probability of falling into arrears than their formal counterparts, perhaps because of the relatively lower level of outstanding liabilities. The liquidity position of informal firms is significantly lower, with informal firms holding about two and a half fewer months’ worth of cash to cover their operations than formal firms. Consistent with the informality literature, the extensive margin of informality, that is, not being registered or lacking documentation, is possibly a factor behind lower credit and liquidity position. This result is not surprising because even during normal times, informal firms are particularly credit constrained and left out of the formal financial and banking system (Busso et al., 2012; La Porta and Shleifer, 2014; Perry and Maloney, 2007) due to lack of collateral and documentation pertaining to accounting and financial records. They also have lower management capabilities to put together a credible loan application (La Porta and Shleifer, 2014). Second, informal firms have lower managerial ability such that their forecasts of future sales have a higher prediction error. This further accentuates their perception of an uncertain operating environment. The extensive margin of informality is associated with higher levels of uncertainty, while high prediction errors further accentuate perceived uncertainty. The latter result is slightly less sharp in the case of necessity firms that have a lower response to such errors. Third, investment in digital technologies and other fixed assets declines sharply with the 3 degree of informality. This is not surprising, given the financial constraints coupled with heightened uncertainty faced by informal firms during the crisis. Fully informal firms have a 25 percentage-point lower probability of investing in technologies and assets than formal firms. Fourth, the elasticity of job growth in informal firms with respect to sales recovery is significantly lower than that of formal firms. While informal firms are expected to have a lower probability of employment adjustment on both the extensive (lay-offs) and the intensive margins (cuts in salaries, benefits, and hours of operation), the result of lower responsiveness to employment growth illustrates that a higher share of informality could become a drag on the economy and contribute to jobless recovery post-crisis. This paper makes two critical contributions to the literature on informality. First, our unique and detailed data allows us to compute a multidimensional definition of informality. Using this measure, we quantify how financial constraints, uncertainty, and adjustment decisions vary in response to a shock by the margins of and motivations for informality. Second, our detailed metrics on firm-level crisis response enables us to empirically test hypotheses and priors about informal firms in a crisis context, for instance, flexibility to make adjustments and, therefore, ability to rebound faster (Alfaro et al., 2020). The rest of the paper is structured as follows. In Section 2, we explain the dataset’s char- acteristics, the construction and descriptive analysis of the main variables of interest, and the methodology implemented. Section 3 presents the main results of our analysis of the heterogeneity of firm-level constraints and decisions according to informality, and Section 4 concludes. 2. Data and Methodology 2.1 Data To analyze the relationship between informality, financial constraints, uncertainty, and firm-level decisions, similar to previous studies like Apedo-Amah et al. (2020), Torres et al. (2023), Cirera et al. (2021), Avalos et al. (2022), Avalos et al. (2023), and Constantinescu et al. (2022), we rely on three rounds of data from the World Bank Business Pulse Survey (BPS).1 In addition to the large set of variables analyzing firm performance, supply chain and international trade-related disruptions, financial fragility, and the different margins 1 The Business Pulse Survey, which finished collecting data in 2022, includes data for around 150,000 firms in 87 countries. 4 of adjustment, including implementation of health protocols, operation hours, number of employees, new financial liabilities, and adjustments, among others, the survey instruments for South Asia also included a module to account for informality through different self- reported measures used in the literature. Through six questions, the survey was able to capture whether the firm was registered, had a Value-Added Tax (VAT) registration number, and had a license for business, which are factors more related to what Ulyssea (2018) and Ulyssea (2020) call the extensive margin of informality and two other variables that indicate if the firm contributes to the employees’ provident fund or social security of the employees, as well as if the firm has a bank account and, conditional on having one if it keeps a separate accounting for the business and for personal matters. Similar to the analysis presented in Grover and Pereira-López (forthcoming), we assume that firms that are informal in the second and third rounds of the survey were informal from the beginning of the pandemic. This is a plausible assumption because, first, we practically do not observe transitions between informality over these two last waves of the survey where we measure informality, and secondly, in these markets, it is not easy to go through all the administrative processes to close (or de-register) a business formally. Another limitation of our data is that it was not possible to implement the third round of data collection for Afghanistan and Sri Lanka due to the political situation in those countries, so considering the trends and characteristics observed regarding lockdown restrictions in previous rounds, the data available mostly misses the whole recovery phase for these two countries. 2.2 Measuring the Margins and Motivations for Informality To measure informality, we follow Grover and Pereira-López (forthcoming) and construct indicators based on the qualitative self-reported questions collected through the BPS. These responses allow us to not merely classify firms as formal or informal but also construct measures that can capture the degree of informality in a continuous manner. To build our continuous measure of informality, we take the first component of a Tetrachoric Principal Component Analysis (PCA) of the five variables of informality: 1) The establishment is not registered for a business license, 2) The establishment is not registered under company law as a sole proprietorship, partnership or limited liability company, 3) The establishment does not contribute to the employees’ provident fund or government employee insurance scheme for its employees, 4) The establishment does not have a value-added tax (VAT) registration number, 5) Neither the owner nor the business has a bank account (see Grover and Pereira-López, forthcoming, for details on tetrachoric correlations, eigenvalues, and 5 factor loadings for the informality indicators. The questions used to build this indicator are shown in Appendix C).2 Building on Grover and Pereira-López (forthcoming), we dig deeper into the differences between the extensive and the intensive margins of informality which, as explained by Ulyssea (2018, 2020) have very different implications in terms of policies to encourage formality. To this end, we construct two sub-indices. PCA 1 restricts to operational indicators and, therefore, accounts for the extensive margin of informality (no business license, no VAT registration number, no registration under company law). PCA 2 focuses on the intensive margin of informality and more direct measures (no contributions to social insurance schemes and no bank account). Thus, a firm could be formal on the extensive margin, pay taxes and have a registration but still not comply with the regulations regarding social security for employees, for example. 3 In our analysis, we also account for heterogeneity among informal firms according to their motivation for informality. Following the methodology in Grover and Pereira-López (forthcoming), we classify informal firms into three categories: necessity, De Soto, and parasite.4 This breakdown of informal firms allows us to test how different are some of the constraints in the context of crises within informal firms as well as if these motivations can also condition the responses of firms in terms of investment decisions and margins of adjustment. 2.3 Estimation Methodology To analyze the relationship between informality, firm-level constraints, and decisions during the pandemic, we estimate the specification in equation 2: 2 On a related note, Medvedev and Oviedo (2016) build a formality index for Ecuador using seven different aspects of formality (public notary, official registry, written contracts, municipal license, social security, receipts for purchases, and taxpayer number). 3 A key difference of this indicator against the definitions from Ulyssea (2018, 2020), is that we are not conditioning the intensive margin on being formal on the extensive margin. To test the robustness of our results, we implement the non-linear approach and construct an intensive informality indicator conditional on being formal on the extensive margin. 4 For this classification, first, we estimate the following equation for labor productivity: log (sales/worker)it0 = β0 + β1 ageit0 + β2 age2 it0 + β3 f emalei + γs + ϕc + νj + uit (1) Where γs are sector fixed effects; ϕc are country fixed effects; and νj are size effects, where j =small, medium, and large. Then, we calculate the 25th and the 75th percentiles of the distribution of this conditional productivity by country and sector for fully formal firms. Those informal firms that exhibit productivity below the 25th percentile are classified as necessity firms, those between the 25th and the 75th percentiles as De Soto firms, and those above the 75th are parasites (see Appendix Figure A1 for the distributions). 6 yit = β0 + β1 Inf ormality + βj Xit + βk Zit + uit (2) The main controls included in Xit are indicator variables to identify microenterprises (fewer than nine employees), age, sectors, and a female-leadership. We also control for country and time-fixed effects, the severity of the lockdown (using the Google mobility indicator),5 and a dummy variable indicating if the firm is a panel firm to control for recall biases in the case of retrospective questions. Zit are additional controls to account for the size of the shock. For example, when we analyze uncertainty, we control for the level of change in sales reported against the level of 2019 because a big shock can lead to higher uncertainty. In the case of investment in digital technologies, we control for the score of digital readiness, as used in previous papers (Cirera et al., 2021; Avalos et al., 2023; Grover and Pereira-López, forthcoming) based on the BPS, because this is a factor that could condition the need to invest in digital technologies. The following are the firm outcomes, yit : Financial fragility: This is measured based on whether the responding firm is already in arrears or expects to fall into arrears in any of its outstanding liabilities (Apedo-Amah et al., 2020; Cirera et al., 2021). On average, over the three waves of data collection, 48% of the observations have a value of one in this dichotomous variable (Table B1), with some variation over time.6 In addition, we measure liquidity constraints with self-reported data on the number of weeks that the firm can cover its operating costs. The average in our sample is around 14 weeks (3.5 months) of costs covered, and half of the observations exhibit less than one month of liquidity.7 Uncertainty: Following Altig et al. (2022), we rely on a measure of uncertainty computed using firm owners’ and directors’ forecasts about their own outcomes over different sce- narios (e.g., normal, optimistic, and pessimistic) as well as the probability distributions associated with these scenarios. This information is used to construct two indices: the first 5 The Mobility Index is based on Google mobility reports around transit stations. Specifically, and following Apedo-Amah et al. (2020) and Cirera et al. (2021), we construct a smoothed index as the weighted average of 30-day periods from the start of the pandemic until the date of the survey. The 30-day period average just before the survey has a weight of 1, the average from day 31-60 has a weight of 1/2, the average from day 61 to 90 has a weight of 1/3, and so on until the start of the pandemic. 6 In the first wave of data collection, the share of firms that reported falling or expecting to fall into arrears was 74%. 7 Across the two rounds where this variable was collected, the variation is high. In the first round, firms reported around 8 weeks, while in the second, it increased to around 16. 7 moment, expectations about changes in sales; and the second moment, uncertainty.8 Such measures are also used in Avalos et al. (2022) and Cirera et al. (2021).9 On average, the expectations about change in sales are close to zero (-0.76) even when the recovery phase is included, while the average uncertainty remained relatively high (a standard deviation of 22 percentage points), driven mainly by the second round, where many lockdown restrictions were still in place in the region (Table B1). Investments: For investment in digital technologies, we rely on a binary variable that indicates if the firm invested in these technologies. The average across all observations in the sample is 20% (Table B1). Though the share of firms that invested in these technologies increased over time, it peaked in the third round of data collection, where almost 22% reported investing in digital technologies. Information on investment in other assets, a variable that we use for robustness checks, was collected only in the third round of the survey. On average, 33% of the firms invested in other assets. Employment adjustments: Our survey allows us to directly observe employment adjustments on both the extensive margin (firing) and the intensive margin (reduction of hours of operation, salaries, and other benefits). The former is observed in 21%, while the latter in 41% of observations. The share of firms adjusting on the intensive margin was particularly higher during the first wave, where almost 70% of the firms reported these adjustments. 3. Results 3.1 Financial fragility and credit constraints are heightened in informal firms both prior to and during the crisis and especially those emerging from necessity. Credit constraints and financial fragility are widely cited as the most binding constraint for informal firms. Applying this to the context of a crisis, we observe a negative (but weak) relationship between the probability of falling into arrears and informality (PCA5, see Figure 1, panel a). The result is intuitive because informal firms are less likely to run into problems complying with their financial obligations because they have lower access to 8 These moments are computed as follows: 3 E [yi ] = i=1 pi yi 3 1 σ = [ i=1 pi (yi − E [yi ])2 ] 2 , where i are the three scenarios: pessimistic, regular, and optimistic, pi , are, therefore, the probabilities of each scenario occurring, and yi is the performance variable that respondents are asked to forecast; in this case, change in sales. 9 Since these questions are time-intensive, they were not included in Pakistan and Sri Lanka in the first round of data collection, which reduces our sample size. 8 formal types of financing before the crisis. Separating informality by motivation, there are no significant differences across necessity, De Soto, and parasite firms in terms of financial fragility, although fragility itself declined across rounds as firms recovered from the crisis (Figure 1, panel b). These results are also confirmed in a regression framework (Table 1, panels A and B columns 1-4), further suggesting that the negative coefficient of informality is mainly driven by the extensive margin, that is, measures relating to registration for a business license, having a VAT number, or being registered under company law (Table 1, panel C). Our results confirm the presence of liquidity constraints among informal firms even during the crisis. We observe a tight negative correlation between informality and cash availability with firms to pay for operational expenses (Figure 1, panel c). While all types of informal firms were more affected relative to formal firms, the impact on necessity firms was mod- erately larger (Table 1, panels A and B, columns 5-8). Even though the extensive margin of informality is still more important (Table 1, panel C columns 5-6), informality on the intensive margin is also a relevant factor conditioning firm liquidity.10 10 Our results are robust to additionally controlling for the size of the shock itself, measured as the accu- mulated change in sales from the onset of the crisis to the reference period in each data collection (Table B2). 9 Figure 1: Informality vs. financial fragility (a) Probability of falling into arrears vs. Informality index (b) Predicted probability of falling into arrears by type of informality .6 90 80 Probability of falling into arrears .55 71 68 Probability of falling into arrears 70 67 64 61 60 56 58 53 .5 50 40 33 29 30 .45 30 27 20 .4 10 0 -.2 0 .2 .4 .6 .8 Fully formal Necessity De Soto Parasite Informality index Round 1 Round 2 Round 3 (c) Weeks to continue operating with cash available vs. Infor- (d) Weeks to continue operating with cash available by type mality index of informality 25 18 Weeks to continue operating with cash available Weeks to continue operating with cash available 14 14 20 13 12 10 9 9 15 8 5 10 0 5 Fully formal Necessity De Soto Parasite -.2 0 .2 .4 .6 .8 Informality index Round 1 Round 2 Notes: Both panels control for sector, age, and country. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). 10 Table 1: Financial fragility and informality Probability of falling into arrears Weeks to continue operating with cash available (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -0.0388* -0.0552** -10.35** -9.803** (0.0230) (0.0278) (1.209) (1.461) Panel B. By type of informality Fully formal Baseline Baseline Baselne Baseline Necessity 0.00240 -0.0296 -5.381** -5.011** (0.0175) (0.0193) (0.920) (1.049) De Soto 0.0222 -0.00555 -4.178** -4.342** (0.0156) (0.0170) (0.856) (0.933) Parasite -0.00502 -0.0125 -4.155** -3.314** (0.0161) (0.0227) (0.873) (1.206) Panel C. By margin of informality PCA 1 -0.0351* -0.0466** -6.451** -6.677** 11 (0.0190) (0.0229) (1.029) (1.198) PCA 2 -0.000449 -0.00302 -3.834** -2.953** (0.0200) (0.0235) (1.049) (1.222) Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Initial productivity No Yes No Yes No Yes No Yes Observations 14,999 11,814 15,306 11,950 7,945 6,074 8,150 6,161 Notes: All continuous variables are winsorized at the 1% and 99%. Weeks to continue operating with cash available was only included as a question in the first two rounds of data collection. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 -No statistically significant differences across types of informality in Panel B, cols. 3 and 4. -Differences between necessity firms and the other two types are statistically significant at the 10% level in Panel B, cols. 7 and 8. -The coefficients of PCA 1 and PCA 2 are statistically different at the 10% level in Panel C, cols. 1 and 2; and at the 5% level in panel C, cols. 5 and 6. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). 3.2 Not only are informal firms pessimistic about future sales, they also perceive uncertainty to be higher than formal firms. Our baseline results, using expectations of change in sales as the dependent variable, suggest that informal firms have relatively lower expectations compared to formal firms (Table 2, panel A columns 1-2). Parasite firms have the lowest expectations compared to formal firms, followed by De Soto. These differences by the type of informal firms remain significant even after controlling for labor productivity (Table 2, panel B columns 3-4). The overall result of lower expectations among informal firms is driven by the intensive margin of informality (Table 2, panel C columns 1-2), suggesting that operational issues rather than constraints to entry affect firms’ expectations. Turning to uncertainty, we observe that informal firms perceive uncertainty to be signif- icantly higher than formal firms (Table 2, panel A columns 5-6). Among the types of informal firms, necessity and De Soto firms perceive uncertainty to be larger than parasite firms when initial productivity is not accounted for (Table 2, panel B column 7), while parasite firms perceive uncertainty to be highest when controlling for initial productivity (Table 2, panel B column 8). The latter difference across types of informal firms are, however, not statistically significant. Firms at the extensive margin of informality perceive the levels of uncertainty to be higher, implying that not fulfilling registration formalities implicitly increases the variation in expectations across scenarios. Uncertainty and expectations are innately linked. In line with the global results from BPS data (Avalos et al., 2022), this relationship is V-shaped for South-Asia (Figure 2, panel a) with discontinuity around zero. This suggests that positive expectations about their future sales can be associated with a significant drop in uncertainty. Nevertheless, during the recovery period, when sales are picking up, uncertainty is higher because the expectations of sales in a pessimistic scenario are still negative, thereby leading to a larger variance. 12 Table 2: Expectations and uncertainty Expectations Uncertainty (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -3.801** -2.522 2.877** 3.826** (1.630) (1.984) (0.951) (1.153) Panel B. By type of informality Fully formal Baseline Baseline Baseline Baseline Necessity -1.977* -1.855 1.828** 1.641** (1.111) (1.229) (0.688) (0.756) De Soto -3.234** -3.700** 1.853** 1.709** (0.929) (1.015) (0.600) (0.650) Parasite -4.224** -4.514** 0.751 2.020** (1.001) (1.513) (0.615) (0.858) Panel C. By margin of informality 13 PCA 1 0.0475 0.983 1.848** 2.628** (1.362) (1.660) (0.793) (0.950) PCA 2 -4.509** -4.247** 0.914 0.998 (1.337) (1.578) (0.786) (0.928) Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Initial productivity No Yes No Yes No Yes No Yes Observations 13066 10391 13303 10501 13066 10391 13303 10501 Notes: All continuous variables are winsorized at the 1% and 99%. in sales across rounds of data collection. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 -Expectations are higher for necessity firms with respect to the other two types of informal at the 10% level of significance in Panel B, cols. 3 and 4. -Uncertainty is lower for parasite firms at the 5% level of significance in Panel B, col. 7, but no significant differences are observed in col. 8. -The coefficients of PCA 1 and PCA 2 are statistically different at the 5% level in Panel C, cols. 1 and 2; and at the 10% level in panel C, cols. 5 and 6. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Not only are expectations associated with uncertainty, but the accuracy of these predictions relative to actual sales reflects the firm’s managerial ability to forecast reasonably and take action to reduce further uncertainty (Avalos et al., 2022). Our results confirm that expectations error, computed as the absolute value of the difference in the expectation of sales in the previous wave relative to the actual, 11 is lower for formal firms relative to informal firms (Figure 2, panel b). While the distributions across the types of informal firms do not differ much, necessity and De Soto firms have slightly higher expectation errors. Given the positive relationship between expectations error and uncertainty (Figure 2, panel c) and that informal firms have larger errors, it is not surprising that perceived uncertainty among informal firms is higher (Figure 2, panel d), especially for the necessity firms. Figure 2: Expectation errors, uncertainty and informality (a) Uncertainty vs. Expectations (b) Absolute expectation error 40 .025 Necessity Necessity De Soto De Soto Parasite Parasite Fully formal .02 Fully formal 30 Kernel density .015 Uncertainty 20 .01 10 .005 0 0 -100 -50 0 50 100 0 50 100 150 200 Expectations about sales x (c) Uncertainty and expectations error (d) Uncertainty by types of Informality Necessity 25 De Soto Parasite Fully formal 20 Uncertainty 15 10 5 0 50 100 Absolute expectation error 3.3 Due to credit constraints and higher perceived uncertainty, informal firms are less likely to invest in productive technologies and assets. Given that informal firms have higher financial constraints, perceived uncertainty, and lower managerial capabilities to reasonably predict future sales, it also affects their investment 11 Absolute expectations error, Error =| yi − E (yi) |. 14 decisions on both time and volume. In particular, investment in digital technologies, which helped firms mitigate the negative demand shock from the pandemic, has a negative correlation with informality (Figure 3, panel a). While the probability of investing in digital technologies over the course of the pandemic improved relative to the onset of the crisis (round 1), the negative correlation with informality prevailed.12 On average, fully informal firms (a value of one in the index of informality) exhibit a 25 percentage points lower probability of investing in digital technologies compared to formal firms (Table 3, panel A column 1). Controlling for initial productivity, the magnitude of the negative association for investment in digital assets is the largest among necessity and De Soto-type firms, suggesting that parasite firms may have more flexibility to make such investments (Table 3, panel B, column 4).13 What contributes to lower investments among informal firms? Consistent with the literature, we document a negative correlation between uncertainty and investments (Figure 3, panel b), especially for fully formal firms and, to a lower extent, for necessity firms. For De Soto and parasites, there is basically no clear correlation. When uncertainty is very high, the probability of investing becomes inelastic, implying that most firms delay and postpone these decisions, irrespective of the level and type of informality. Shapley decomposition of the probability of investing in digital technologies suggests that while firm constraints, including financial and perceived uncertainty, explain 8.5% of the outcome on investments, informality plays a key role in both the peak of the crisis (45%) and the recovery phase (25%) (Figure 4). 12 During the recovery phase (third round), the curve moves back down, and the slope becomes larger in absolute terms. This indicates that firms could have temporarily overreacted and then moved back to a new level where formal firms maintain a similar probability of keeping investment in these technologies, while informal firms, consistent with their initial constraints, show a much lower probability of investing. 13 These results are robust to additional control of change in sales in the last 30 days and average uncertainty about prediction, see Appendix Table B3. In addition, the result holds for investment in other assets as well, although the number of observations is fewer because the variable was collected only in the third round (Table 3, panel A column 5). 15 Table 3: Investment in digital technologies and fixed assets Has invested in digital Has invested in fixed assets (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -0.251** -0.288** -0.144** -0.126** (0.0201) (0.0333) (0.0343) (0.0407) Panel B. By type of informality Fully formal Baseline Baseline Baseline Baseline Necessity -0.0976** -0.118** 0.000347 -0.0195 (0.0133) (0.0165) (0.0263) (0.0294) De Soto -0.0996** -0.107** -0.00524 -0.0221 (0.0114) (0.0142) (0.0225) (0.0248) Parasite -0.129** -0.0564** -0.102** -0.0607* (0.0119) (0.0203) (0.0240) (0.0327) Digital readiness 0.0537** 0.0539** 16 (0.00495) (0.00494) Panel C. By margin of informality PCA 1 -0.0988** -0.145** -0.0109 0.00416 (0.0167) (0.0291) (0.0272) (0.0321) PCA 2 -0.168** -0.146** -0.187** -0.186** (0.0157) (0.0215) (0.0298) (0.0349) Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Initial productivity No Yes No Yes No Yes No Yes Digital readiness No Yes No Yes No No No No Observations 14532 9702 14816 9813 6391 5222 6477 5279 Notes: All continuous variables are winsorized at the 1% and 99%. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 -Parasite firms exhibit a statistically significant lower probability of investing at the 5% level of significance in Panel B, col. 3, but, controlling for initial productivity, a higher probability than other types of informal at the 1% level in col. 4. -In col. 7, parasite firms show a lower probability vs. other types of informal at the 1% level. In col. 8, the differences are not statistically significant. -The coefficients of PCA 1 and PCA 2 are statistically different at the 1% level in Panel C, cols. 1, 5, and 6; and are not statistically different in col. 2. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Figure 3: Investment in digital, informality, and uncertainty (a) Investment in digital vs. informality (b) Investment in digital vs. uncertainty by type of informality .4 .35 Round 1 Necessity Probability of investment in digital technologies Probability of investing in digital technologies Round 2 De Soto Round 3 Parasite .3 Fully formal .3 .2 .25 .1 .2 0 .15 -.1 0 .2 .4 .6 .8 -20 0 20 40 60 Informality (PCA score) Uncertainty Notes: Controls for size, sector, age, gender of the CEO, country, panel firms, and initial digital readiness. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Figure 4: Shapley decomposition investment digital 60 Relative Shapley Value (%) - Probability of investing in digital 45.2 Pseudo R2 explained by group 40 28.9 24.5 24.9 22 20.1 20 15.7 8.5 8.5 1.6 0 Round 2 Round 3 Informality External attributes Severity of the lockdown (mobility) Severity of demand shock Constraints Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). 3.4 Informal firms are less likely to make employment adjustments and respond with less job growth even as sales begin to recover. High levels of uncertainty, lack of working capital, and financial constraints also affect firms’ decisions on employment adjustments. On the one hand, informal firms have more flexibility to adjust as they are also informal in their hiring practices and can easily change the terms of informal contracts with employees or even fire without notice. During crises, adjustments on both intensive and extensive margins are, therefore, expected to be more 17 agile. On the other hand, the smaller scale (in terms employee size) of informal firms does not leave much scope for adjustments on either the extensive or the intensive margins. The overall effect depends on the net of the two effects and is an empirical question. Our data suggest that the scale effect dominates such that informal firms have about an 8 and 14 percentage-point lower probability of adjusting on the extensive and intensive margins, respectively (Table 4, panel A column 1 and 5). The elasticity of adjustment is lowest among informal firms driven by necessity and highest among the parasite type, suggesting that the available scale of adjustment could be driving these results.14 Shapley’s decomposition of the probability of employment adjustment at extensive and intensive margins suggests that informality explains 60% and 72%, respectively, of these adjustments (Figure 5, panels a and b). The size of the shock (change in sales) is much more important at the extensive margin, indicating that larger negative effects of the crisis on sales induce more adjustments on the extensive margin. Firm-level constraints (mainly uncertainty) explain 11% of the adjustments on the extensive margin, while a much smaller share on the intensive margin. The overall characteristics of informality explain a larger share of adjustments on the intensive margin, and so do the external attributes such as age, size, and sector. Macro evidence suggests that the unemployment rate rises sharply during recessions, while the recovery is only slow and gradual during expansions. This mismatch of sharp spikes in unemployment during recessions and slow job gains during expansions is known as “cyclical asymmetry” (Andolfatto and Spewak, 2018; Andolfatto, 1997; Ferraro, 2018, 2023). Micro evidence in Cirera et al. (2021) confirms this pattern using BPS data. Our work suggests that the elasticity of change in employment with the drop in sales was comparable for both formal and informal firms in the first phase of the crisis. In the recovery phase, however, informal firms show a nearly inelastic response to employment growth with the recovery in sales. This is in contrast to formal firms where the change in employment is slower but still positive (Figure 6, panel a). Using a sharper definition of fully formal firms, we illustrate that not only do all types of informal firms show signs of relatively slow job recovery, and particularly the parasite ones, but that the fully formal firms are remarkably different (Figure 6, panel b). This implies that the degree of formality matters. Given the presence of a large informal sector in South Asian countries, this does not bode well for job recovery in the post-crisis period. 14 Mean size in our sample is ten employees for necessity firms, 13 for De Soto, 12 for parasite firms, and 48 for fully formal. We test the robustness of our result to adding uncertainty and change in sales as controls in Table B4, and the results do not change. 18 Table 4: Employment adjustments Extensive margin Intensive margin (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -0.0805** -0.115** -0.136** -0.0700** (0.0244) (0.0318) (0.0279) (0.0342) Panel B. By type of informality Fully formal Baseline Baseline Baseline Baseline Necessity -0.0172 -0.0331* -0.0413** -0.0498** (0.0161) (0.0182) (0.0191) (0.0219) De Soto -0.0237* -0.0681** -0.0416** -0.0653** (0.0139) (0.0154) (0.0173) (0.0193) Parasite -0.0715** -0.0964** -0.158** -0.127** (0.0140) (0.0213) (0.0176) (0.0272) Panel C. By margin of informality PCA 1 -0.00889 -0.0183 0.0166 0.0500* (0.0211) (0.0270) (0.0246) (0.0300) PCA 2 -0.0783** -0.106** -0.176** -0.141** (0.0196) (0.0241) (0.0239) (0.0289) 19 Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Initial productivity No Yes No Yes No Yes No Yes Observations 8886 6809 9119 6902 8888 6810 9121 6902 Notes: All continuous variables are winsorized at the 1% and 99%. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 -Parasites exhibit a significantly lower probability of adjustments at the 1% level in Panel B, cols. 3, 7, and 8. Both the Soto and parasite firms adjust less at the 5% level of significance in Panel B, col. 4. -The coefficients of PCA 1 and PCA 2 are statistically different at the 1% level in Panel C, cols. 1, 2, 5, and 6. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Figure 5: Shapley decomposition employment adjustments (a) Extensive margin (b) Intensive margin Relative Shapley Value (%) - Probability of adjusting employment extensive margin Relative Shapley Value (%) - Probability of adjusting employment intensive margin 59.8 72.4 60 60 Pseudo-R2 explained by group Pseudo-R2 explained by group 40 40 22 20 20 20 11 5.6 4.3 1.6 2.3 1 0 0 Informality External attributes Informality External attributes Severity of the lockdown (mobility) Severity of demand shock Severity of the lockdown (mobility) Severity of demand shock Constraints Constraints Notes: Firm attributes include sector, age, gender of the CEO, country, and panel firms. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Figure 6: Employment adjustments in crisis and recovery phases (a) Employment adjustments by formal and informal firms (b) Employment adjustments by types of informal firms 0 0 -10 Change in employment (%) Change in employment (%) -10 -20 -20 -30 -30 -40 -100 -50 0 50 100 -100 -50 0 50 100 Change in sales over waves (p.p) Change in sales over waves (p.p) Formal Informal Necessity De Soto Parasite Fully formal Notes: Firm attributes include sector, age, gender of the CEO, country, and panel firms. Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). 20 4. Conclusions Building on companion research using three rounds of the World Bank’s Business Pulse Surveys in South Asia, this paper quantifies the relationship between informality and firms’ investment and employment decisions. Our work suggests that informal firms face larger constraints on liquidity, access to finance, and perceived uncertainty. Accounting for multidimensionality and the margins of informality, our work suggests that first, informal firms, especially those motivated by necessity and survival objectives, are both liquidity and credit constrained. Second, informal firms perceive higher uncertainty in the context of the crisis. Not surprisingly, the extensive margin of informality, measured via registra- tion or a lack of documentation, is highly correlated with the lack of credit and a higher perceived uncertainty. Third, acute credit constraints and higher perceived uncertainty, especially among necessity firms, lead to a substantially lower probability of investing in digital technologies and fixed assets. Furthermore, informal firms have a lower elasticity of employment as sales begin to recover. Our results suggest that countries with a large informal sector may need to deploy additional levers to accelerate recovery in the post-crisis period.15 First, a lack of access to finance by informal firms can be addressed with additional support from microfinance institutions (MFI).16 At a broader scale, countries with a large informal sector should ideally focus on developing their financial markets to reduce information asymmetries and perceived risks by financial institutions. Second, for dealing with uncertainty, it is important that policy makers clearly communicate the public support program in terms of the scope and time frame, to partially reduce uncertainty and its effect in delaying investment and hiring decisions. Third, policies focused on generating capabilities within firms can contribute to an appreciation for investments in technologies for coping with the crisis. Finally, given that the extensive margin of informality is highly correlated with outcomes on investments, policies aimed at promoting the formalization (e.g., simplified schemes for registration) if accompanied by policies reducing the ongoing costs or increasing the benefits of formality, could allow access to formal financial sources as informal firms try to emerge from the shock. 15 Policies focused on reducing the costs of formalizing (i.e., registration costs) do not yield significant results in inducing firms to formalize (see Ulyssea, 2020; Bruhn and McKenzie, 2014; De Mel et al., 2013). By contrast, monetary incentives can be helpful (De Mel et al., 2013), although the cost-effectiveness of such interventions needs to be considered carefully. 16 Although microfinance may not be transformational, it is still a critical vehicle to channel credit to informal firms, especially during crisis (Malik et al., 2020). Furthermore, innovations like asset-based microfinance can allow for investments in a financially-sustainable way to credit-constrained firms (Cai et al., 2021). 21 References L. Alfaro, O. Becerra, and M. Eslava. EMES and COVID-19: Shutting Down in a World of Informal and Tiny Firms. NBER Working Paper Series, 2020. URL http://www.nber.org/papers/w27360. D. Altig, J. M. Barrero, N. Bloom, S. J. Davis, B. Meyer, and N. Parker. Surveying business uncertainty. Journal of Econometrics, 231(1):282–303, 2022. D. Andolfatto. Evidence and theory on the cyclical asymmetry in unemployment rate fluctuations. Canadian Journal of Economics, pages 709–721, 1997. D. Andolfatto and A. Spewak. Does the yield curve really forecast recession? Economic Synopses, 1 (30):1–2, 2018. M. C. Apedo-Amah, B. Avdiu, X. Cirera, M. Cruz, E. Davies, A. Grover, L. Iacovone, U. Kilinc, D. Medvedev, F. O. Maduko, S. Poupakis, J. Torres, and T. T. Tran. Unmasking the Impact of COVID- 19 on Businesses. Policy Research Working Paper; No. 9434., 10 2020. doi: 10.1596/1813-9450-9434. URL https://openknowledge.worldbank.org/handle/10986/34626. E. Avalos, J. M. Barrero, E. Davies, L. Iacovone, and J. Torres. Business expectations and uncertainty in developing and emerging economies. Available at SSRN 4327860, 2022. E. Avalos, X. Cirera, M. Cruz, L. Iacovone, D. Medvedev, G. Nayyar, and S. R. Ortega. Firms’ Digitalization during the COVID-19 Pandemic. Policy Research Working Paper; No. 10284, 2023. J. M. Barrero and N. Bloom. Economic uncertainty and the recovery. In Navigating the Decade Ahead: Implications for Monetary Policy", Federal Reserve Bank of Kansas City, Economic Policy Symposium Proceedings, pages 255–284, 2020. N. Bloom. Fluctuations in uncertainty. Journal of economic Perspectives, 28(2):153–176, 2014. N. Bloom, S. Bond, and J. Van Reenen. Uncertainty and investment dynamics. The review of economic studies, 74(2):391–415, 2007. J.-C. Bricongne, L. Fontagné, G. Gaulier, D. Taglioni, and V. Vicard. Firms and the global crisis: French exports in the turmoil. Journal of international Economics, 87(1):134–146, 2012. M. Bruhn and D. McKenzie. Entry regulation and the formalization of microenterprises in developing countries. The World Bank Research Observer, 29(2):186–201, 2014. M. Busso, M. Fazio, and S. Algazi. (In) formal and (un) productive: the productivity costs of excessive informality in Mexico. IDB Working Paper No. IDB-WP-341, 2012. URL https://papers. ssrn.com/sol3/papers.cfm?abstract_id=2207240. 22 M. Bussolo, S. Sharma, and H. Timmer. COVID-19 has worsened the woes of South Asia’s informal sector, 2020. URL https://blogs.worldbank.org/endpovertyinsouthasia/ covid-19-has-worsened-woes-south-asias-informal-sector. J. Cai, M. Meki, S. Quinn, E. Field, C. Kinnan, J. Morduch, and F. Said. Microfinance. VoxDevLit, 3 (1):26, 2021. N. Cetorelli and L. S. Goldberg. Global banks and international shock transmission: Evidence from the crisis. IMF Economic review, 59(1):41–76, 2011. D. Chor and K. Manova. Off the cliff and back? credit conditions and international trade during the global financial crisis. Journal of international economics, 87(1):117–133, 2012. X. Cirera, M. Cruz, A. Grover, L. Iacovone, D. Medvedev, M. Pereira-Lopez, and S. Reyes. Firm Recovery during COVID-19. 10 2021. doi: 10.1596/1813-9450-9810. URL https://openknowledge. worldbank.org/handle/10986/36428. C. Constantinescu, A. M. Fernandes, S. Poupakis, A. G. Goswami, and S. Reyes. Globally engaged firms in the covid-19 crisis. 2022. S. De Mel, D. D. McKenzie, and C. Woodruff. The Demand for, and Consequences of, Formalization among Informal Firms in Sri Lanka. American Economic Journal: Applied Economics, 5(2):122–50, 4 2013. ISSN 1945-7782. doi: 10.1257/APP.5.2.122. G. Dell’Ariccia, E. Detragiache, and R. Rajan. The real effect of banking crises. Journal of Financial Intermediation, 17(1):89–112, 2008. T. Didier, F. Huneeus, M. Larrain, and S. L. Schmukler. Financing firms in hibernation during the covid-19 pandemic. Journal of Financial Stability, 53:100837, 2021. D. Ferraro. The asymmetric cyclical behavior of the us labor market. Review of Economic Dynamics, 30:145–162, 2018. D. Ferraro. Fast rises, slow declines: Asymmetric unemployment dynamics with matching frictions. Journal of Money, Credit and Banking, 55(2-3):349–378, 2023. H. Görg and M.-E. Spaliara. Financial health, exports and firm survival: Evidence from uk and french firms. Economica, 81(323):419–444, 2014. A. Grover and M. Pereira-López. Do Shocks Perpetuate Disparities Within and Across Informal Firms? Evidence from the COVID-19 Pandemic in South Asia, forthcoming. L. Iacovone, E. Ferro, M. Pereira-López, and V. Zavacka. Banking crises and exports: Lessons from the past. Journal of Development Economics, 138:192–204, 2019. 23 R. La Porta and A. Shleifer. Informality and development. Journal of Economic Perspectives, 28(3): 109–126, 6 2014. doi: 10.1257/JEP.28.3.109. K. Malik, M. Meki, J. Morduch, T. Ogden, S. Quinn, and F. Said. Covid-19 and the future of microfi- nance: Evidence and insights from pakistan. Oxford Review of Economic Policy, 36(Supplement_1): S138–S168, 2020. D. Medvedev and A. M. Oviedo. Informality and profitability: Evidence from a new firm survey in ecuador. The Journal of Development Studies, 52(3):412–427, 2016. B. Meyer, E. Mihaylov, J. M. Barrero, S. J. Davis, D. Altig, and N. Bloom. Pandemic-era uncertainty. Journal of Risk and Financial Management, 15(8):338, 2022. D. Paravisini, V. Rappoport, P. Schnabl, and D. Wolfenzon. Dissecting the effect of credit supply on trade: Evidence from matched credit-export data. The Review of Economic Studies, 82(1):333–359, 2015. G. Perry and W. F. Maloney. Informality: Exit and exclusion. World Bank Publications, 2007. C. M. Reinhart. From health crisis to financial distress. IMF Economic Review, 70(1):4–31, 2022. C. M. Reinhart and K. S. Rogoff. The aftermath of financial crises. American Economic Review, 99(2): 466–472, 2009. J. Torres, F. Maduko, I. Gaddis, L. Iacovone, and K. Beegle. The impact of the covid-19 pandemic on women-led businesses. The World Bank Research Observer, 38(1):36–72, 2023. G. Ulyssea. Firms, Informality, and Development: Theory and Evidence from Brazil. American Economic Review, 108(8):2015–47, 8 2018. ISSN 0002-8282. doi: 10.1257/AER.20141745. G. Ulyssea. Informality: Causes and Consequences for Development. 2020. doi: 10.1146/ annurev-economics. URL https://doi.org/10.1146/annurev-economics-. 24 A. Online Appendix Figures Figure A1: Distribution of conditional productivity across types of informality .4 .3 Kernel density .2 .1 0 -5 0 5 10 IHS transformation labor productivity conditional on size, sector, age, female ownership, and country Fully formal Necessity De Soto Parasite Informal (all types) Source: Grover and Pereira-López (forthcoming) 25 B. Online Appendix Tables Table B1: Descriptive Statistics Variable Mean sd p10 p50 p90 N Main outcome variables Has fallen into arrears 0.482 0.500 0.000 0.000 1.000 15,022 Weeks to cover costs 14.237 19.367 0.000 4.000 52.000 7,832 Uncertainty about change in sales 22.192 19.679 0.000 17.321 54.443 13,591 Expectations about change in sales -0.760 30.520 -44.815 0.250 32.500 13,591 Absolute error in prediction of change in sales 33.914 28.485 3.300 27.000 75.250 4,674 Has invested in digital technologies 0.208 0.406 0.000 0.000 1.000 14,779 Has invested in fixed assets in 2021 0.334 0.472 0.000 0.000 1.000 6,497 Adjusted labor extensive 0.208 0.406 0.000 0.000 1.000 8,870 Adjusted labor intensive 0.405 0.491 0.000 0.000 1.000 8,871 Control variables Manufacturing 0.418 0.493 0.000 0.000 1.000 19,826 Retail 0.180 0.384 0.000 0.000 1.000 19,826 Services 0.402 0.490 0.000 0.000 1.000 19,826 Micro (9 employees or less) 0.493 0.500 0.000 0.000 1.000 20,160 Young 0.417 0.493 0.000 0.000 1.000 20,092 Mature 0.152 0.359 0.000 0.000 1.000 20,092 Established 0.431 0.495 0.000 0.000 1.000 20,092 Female-led 0.205 0.404 0.000 0.000 1.000 20,160 Change in sales vs. 2019 (%) -35.319 40.734 -100.000 -40.000 15.000 20,160 Score digital readiness (from 0-3) 1.306 1.180 0.000 1.000 3.000 13,624 Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). 26 Table B2: Robustness Financial fragility and informality controlling for change in sales Probability of falling into arrears Weeks to continue operating with cash available (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -0.0602** -0.0738** -9.690** -8.855** (0.0231) (0.0277) (1.232) (1.469) % change in sales during past 30 days -0.00181** -0.00184** -0.00179** -0.00184** 0.0753** 0.0664** 0.0754** 0.0675** (0.000124) (0.000142) (0.000123) (0.000142) (0.00710) (0.00798) (0.00704) (0.00796) log(Initial labor productivity) 0.00113 0.000295 0.159* 0.156* (0.00150) (0.00162) (0.0816) (0.0925) Panel B. By type of informality Fully formal Baseline Baseline Baselne Baseline Necessity -0.0194 -0.0442** -4.298** -3.967** (0.0175) (0.0192) (0.931) (1.053) De Soto 0.00119 -0.0245 -3.406** -3.507** (0.0156) (0.0169) (0.862) (0.931) Parasite -0.0210 -0.0254 -3.462** -3.172** 27 (0.0162) (0.0227) (0.887) (1.206) Panel C. By margin of informality PCA1 -0.0423** -0.0485** -6.440** -6.228** (0.0192) (0.0230) (1.056) (1.217) PCA2 -0.0167 -0.0234 -3.117** -2.461** (0.0200) (0.0233) (1.054) (1.212) % change in sales during past 30 days -0.00180** -0.00184** 0.0759** 0.0669** (0.000124) (0.000142) (0.00711) (0.00799) log(Initial labor productivity) 0.00113 0.161** (0.00150) (0.0817) Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Observations 14532 11581 14809 11709 7620 5916 7802 5997 Notes: All continuous variables are winsorized at the 1% and 99%. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Table B3: Robustness: Investment in digital technologies and fixed assets Investment digital Investment fixed assets (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -0.248** -0.268** -0.136** -0.115** (0.0218) (0.0350) (0.0381) (0.0446) Panel B. By type of informality Fully formal Baseline Baseline Baseline Baseline Necessity -0.0928** -0.112** 0.00820 -0.0185 (0.0143) (0.0175) (0.0280) (0.0313) De Soto -0.0992** -0.104** 0.00144 -0.0194 (0.0120) (0.0146) (0.0236) (0.0260) Parasite -0.120** -0.0498** -0.101** -0.0625* (0.0125) (0.0213) (0.0253) (0.0343) % change in sales during past 30 days 0.00120** 0.00134** 0.00117** 0.00129** 0.00209** 0.00224** 0.00204** 0.00221** (0.000107) (0.000135) (0.000106) (0.000134) (0.000203) (0.000229) (0.000201) (0.000227) 28 Average uncertainty of the prediction -0.000128 0.0000662 -0.000163 -0.0000310 -0.00192** -0.00224** -0.00189** -0.00213** (0.000243) (0.000301) (0.000238) (0.000293) (0.000550) (0.000617) (0.000548) (0.000616) Panel C. By margin of informality PCA 1 -0.104** -0.135** -0.00844 0.0133 (0.0184) (0.0309) (0.0305) (0.0353) PCA 2 -0.156** -0.137** -0.174** -0.180** (0.0167) (0.0225) (0.0320) (0.0371) % change in sales during past 30 days 0.00119** 0.00134** 0.00206** 0.00221** (0.000107) (0.000135) (0.000203) (0.000228) Average uncertainty of the prediction -0.000130 0.0000538 -0.00195** -0.00226** (0.000243) (0.000301) (0.000550) (0.000617) Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Initial productivity No Yes No Yes No Yes No Yes Digital readiness No Yes No Yes No No No No Observations 14532 9702 14816 9813 6391 5222 6477 5279 Notes: All continuous variables are winsorized at the 1% and 99%. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). Table B4: Robustness: Employment adjustments controlling for change in sales and uncertainty Extensive margin Intensive margin (1) (2) (3) (4) (5) (6) (7) (8) Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial Conditional Cont. initial productivity productivity productivity productivity Panel A. Baseline estimations Informality -0.104** -0.146** -0.121** -0.0353 (0.0286) (0.0380) (0.0332) (0.0417) % change in sales during past 30 days 0.0000240 0.000238 -0.000778** -0.000765** (0.000138) (0.000166) (0.000178) (0.000213) Average uncertainty of the prediction 0.0000880 -0.000320 0.000803** 0.000605 (0.000289) (0.000345) (0.000347) (0.000415) Panel B. By type of informality Fully formal Baseline Baseline Baseline Baseline Necessity -0.0247 -0.0486** -0.0341 -0.0402 (0.0190) (0.0218) (0.0223) (0.0255) De Soto -0.0312** -0.0764** -0.0406** -0.0565** (0.0157) (0.0176) (0.0197) (0.0221) Parasite -0.0822** -0.0983** -0.155** -0.113** (0.0161) (0.0251) (0.0200) (0.0319) 29 % change in sales during past 30 days -0.0000440 0.000193 -0.000771** -0.000753** (0.000136) (0.000166) (0.000175) (0.000212) Average uncertainty of the prediction 0.0000560 -0.000329 0.000837** 0.000707* (0.000282) (0.000340) (0.000339) (0.000410) Panel C. By margin of informality PCA 1 -0.0188 -0.0291 0.0568* 0.113** (0.0252) (0.0330) (0.0293) (0.0368) PCA 2 -0.0889** -0.122** -0.202** -0.171** (0.0225) (0.0276) (0.0271) (0.0332) % change in sales during past 30 days 0.00000824 0.000221 -0.000829** -0.000806** (0.000138) (0.000166) (0.000178) (0.000213) Average uncertainty of the prediction 0.0000831 -0.000340 0.000801** 0.000581 (0.000289) (0.000345) (0.000347) (0.000414) Panel controls Yes Yes Yes Yes Yes Yes Yes Yes Other controls Yes Yes Yes Yes Yes Yes Yes Yes Initial productivity No Yes No Yes No Yes No Yes Observations 7117 5464 7277 5523 7118 5465 7278 5523 Notes: All continuous variables are winsorized at the 1% and 99%. Robust standard errors in parentheses * p<0.10, ** p<0.05, ** p<0.01 Source: Authors’ calculations using data from the World Bank Business Pulse Survey (BPS). C. Questionnaire Informality Informality module Is this establishment registered for a business license? Cov6a6 1= Yes 2= No INSTRUCTION: Only ask in countries where sampling -9=Don’t know (spontaneous) frame may include informal firms Does this establishment have a tax registration number? Cov6e10 1= Yes 2= No INSTRUCTION: Only ask in countries where sampling -9=Don’t know (spontaneous) frame may include informal firms Is this Establishment registered under company law as a Cov6e7 1= Yes sole proprietorship, partnership or limited liability 2= No company? -9=Don’t know (spontaneous) INSTRUCTION: Only ask in countries where sampling frame may include informal firms Is this establishment’s business bank account separate Cov6e8 1= Yes, Establishment has a separate from any personal accounts of the owners? bank account 2= No, Establishment has a mixed business/personal account 3= No, Establishment or the owner don’t have a bank account Does this Establishment contribute to the employees’ Cov6e9 1= Yes provident fund or government employee insurance 2= No scheme for its employees? -9=Don’t know (spontaneous) INSTRUCTION: Only ask in countries where sampling frame may include informal firms 30