ENTERPRISE S URVEYS ENTERPRISE NOTE S ERIES GENDER Do crises hit female-managed and male-managed firms differently? Evidence from the 2008 financial crisis 2020 1 Tanima Ahmed, Silvia Muzi, Kohei Ueda W hile e orts are currently in place to collect data on the economic consequences of the COVID-19 pandemic, this note looks at the experience of the 2008 nancial crisis to gain insights on possible di erential e ects of crises on female and male entrepreneurs. Speci cally, the note uses rm-level data collected immediately after the 2008 nancial crisis in six countries in Eastern Europe and Central Asia (Bulgaria, Hungary, Latvia, Lithuania, Romania, and Turkey) to look at two aspects of the di erential e ect of the crisis. First, whether there is a di erence in ENTERPRISE NOTE No . 39 the exit rate for rms with male vs. female top managers; and second, whether, among rms that stayed in business, female-managed rms are a ected disproportionally. Results show that rms run by female top managers are more likely to exit the market. Secondly, when able to stay in business, male and female-managed rms su ered a similar impact in the short term; however, female-managed rms su ered more than male-managed rms in the longer term. Introduction social and personal services, trade, and wholesale were hit e COVID-19 pandemic and the consequent less hard. Similar results are found by Hallward-Driemeier transformation of the day-to-day life worldwide are et al. (2017) that examined how variations in rms’ posing unprecedented economic challenges to the private responses during the East Asian crisis in Indonesia a ected sector. Firms are closing or reducing their activities men’s and women’s relative employment vulnerability. because of government restrictions, lack of demand, or ey found that women experienced higher job losses lack of critical inputs or raw materials. While all rms are than their male colleagues within the same rm. However, impacted, there is a growing debate about the possibility the aggregate adverse e ect of such di erential treatment of disproportionally negative e ects of the crisis on was more than o set by women being disproportionately women entrepreneurs. As the World Bank and other employed in rms hit relatively less by the crisis. institutions are collecting rm-level data to answer this When moving out from the labor market lens and question and, more generally, to assess the e ects of the looking at the gender dimension of rms’ adjustments COVID-19 crisis on the private sector, we looked back at WORLD BANK GROUP during crises, the literature is even more limited. Cesaroni the experience of the 2008 nancial crisis to get some et al. (2015) analyzed the response of Italian insights on possible di erential e ects of crises on micro-entrepreneurs to the 2008 economic crisis. ey female-managed and male-managed rms. e results ll found that women entrepreneurs mainly dealt with the an evident gap in the literature and o er useful insights for nancial crisis with a defensive attitude and preferred the upcoming analysis of the COVID-19 data. downsizing their activities and increasing e ciency. e literature on gender-disaggregated e ects of crises Moreover, in Italy, female-owned rms were found to face has, so far, mainly focused on labor market outcomes with the bulk of the studies looking at the e ects on the supply a more pronounced tightening in credit supply during the of labor from individuals (Benería, 2015; Seguino, 2010; crisis with respect to male-owned rms (Cesaroni et al. Lahey and de Villota, 2013). Fewer studies have 2013). Additional studies looked at the e ect of the investigated the demand side. Lim (2000) looked at the nancial crisis on rms overall. Hallward-Driemeier and e ects of the East Asian crisis in the Philippines and found Rijkers (2013), found that the crisis led to a spike in exits that while women were displaced more in the independently of the productivity levels of rms driven manufacturing sector, they overall su ered less from the out of business. Moreover, they found a slowdown in crisis as sectors with high female employment such as entry and excessive employment reallocation. Firms more vulnerable to changing credit market conditions were Before the crisis, in the six countries analyzed, on much more likely to exit during the crisis. Firm size has average 42 percent of rms had at least one woman among also been investigated as potential contributor to rm exit the owners and 20 percent of rms had a female top during an economic crisis. Results for Portugal suggest manager. Both gures are in line or higher than the that large rms su er a greater increase in exit hazard average in the other countries in Europe and Central Asia during downturns than smaller rms, although small (ECA) and equal to the average in all countries with ES rms remain generally more likely to exit. Size is less data available ( gure 1). Unfortunately, before 2009, the helpful in avoiding exits in a crisis context, as a large size ES didn’t systematically collect the percentage of female may be responsible for rm inertia and an inability to ownership in respondent rms. e absence of this key adapt optimally to an adverse environment. Neither of the information prevents us from assessing the degree of last two studies, however, looked at the gender dimension. female ownership in these rms. Because of this, the When we shift our focus to the broader literature on rm following analysis focuses on rms with a female top exit, research through the gender lens is also very limited.2 manager.4 Klanins and Williams (2014) examined survival time of Before the crisis, overall rms with a female top male-owned and female-owned businesses. Using Texas manager were as productive as rms with a male top sales tax data, they found that female-owned businesses manager. When looking at the full sample for all countries, labor productivity for female-managed rms (FMF) and out-survive male-owned businesses in certain industries male-managed rms (MMF), de ned as annual sales per and geographic areas although generally female-owned employee, was $49,982 and $50,347, respectively. e businesses tend to be more short-lived. None of these labor productivity of female-managed rms and studies have a cross-country perspective and it is clear that male-managed rms was very similar for rms in retail and more research through the gender lens is needed. We try other services sectors and for small (5-19 employees) and to ll this gap with the analysis presented in this note. medium rms (20-99 employees). However, rms with male top managers were more productive than rms with Data and selected pre-crisis stylized facts female top managers among large rms (100+ employees) is analysis is based on a unique dataset collected to ( gure 2) and manufacturing rms.5 measure the e ects of the 2008 nancial crisis on rms. Access to nance is another key factor to consider as it e Financial Crisis Survey (FCS) was conducted using enables rms to face downturns. Firms that have access to phone interviews to re-visit rms that participated in – at external nance may be able to mitigate the impacts of the time recently conducted- 2008 Enterprise Surveys temporary shocks that would otherw ise force them to (ES). Interviews were carried out in three waves in six exit. As presented in gure 3, selected indicators for access countries: Bulgaria, Hungary, Latvia, Lithuania, to nance for the full sample also show some di erences Romania, and Turkey.3 e rst wave was implemented between female-managed and male-managed rms. On during the months of June and July 2009. e other two average, 41 percent of rms with a female top manager waves, wave 2 and wave 3, were conducted in the months had a bank loan or line of credit compared to 49 percent of February and March 2010, and May and June 2010, of rms managed by men6; on the contrary, the respectively. From the information collected during short proportion of rms not needing a loan was very similar for female- and male-managed rms (45 percent and 46 phone interviews, indicators are computed to measure the percent, respectively). ese patterns are con rmed across e ects of the crisis on key elements of the private sectors and size.7 economy: sales, employment, and nances. Using survey Finally, we looked at access to technology as it may weights, all indicators make inferences to the broader lead to– or be a proxy for- levels of adaptability that may population under consideration. be essential for rms to avoid exiting the market during e original ES data also serve as baseline for crises. ES data show that overall, at the time of the crisis, comparisons since most of its questions refer to scal year female-managed rms where less likely to use technology 2007, thus measuring the pre-crisis scenario. e ES than male-managed rms. In 2008, in the full sample, 72 collect information on a representative sample of formal percent of rms with a female top manager used email in (registered) private rms with at least ve employees their communications with clients as compared to 87 operating in the manufacturing or services sectors. By percent of men, and only 48 percent of female-managed using a global methodology based on strati ed random rms had their own website compared to 60 percent of sampling, the ES data are fully comparable across male-managed rms ( gure 4)8. e disadvantaged countries. Besides covering several aspects of the business position of female-managed rms in the use of technology environment and rms’ outcomes, the ES also collect as measured by the rm having its own website is information on rms’ characteristics, including questions con rmed across size and sectors. When looking at the use on the gender of the rm’s top manager, on whether rms of email to communicate with clients, female-managed have any women among the owners, and on the rms do as well as male-managed rms only among percentage of women in the workforce. manufacturing rms and large rms. 2 Degree of female participation in firm Labor productivity: median annual sales Figure 1 ownership and management Figure 2 per worker ($USD, deflated) 50% 70,000 64,352 42% 60,000 54,621 40% 36% 49,228 49,982 49,091 34% $USD, deflated 50,000 % of firms 38,653 30% 40,000 20% 19% 18% 30,000 20% 20,000 10% 10,000 0% 0 Percent of firms with Percent of firms with a Small (5-19) Medium (20-99) Large (100+) female participation in female top manager ownership FCS ECA All countries FMF MMF Access to finance: selected indicators Use of technology: selected indicators Figure 3 Figure 4 60% 100% 87% 49% 50% 45% 46% 41% 80% 72% % of firms 40% 60% % of firms 60% 48% 30% 40% 20% 10% 20% 0% 0% Percent of firms with Percent of firms not Percent of firms having Percent of firms using a bank loan/line of needing a loan its own website e-mail to credit communicate with clients/suppliers FMF MMF FMF MMF Source: Enterprise Surveys (2008-2009) and Financial Crisis Surveys (2009) Note: “FCS” is the average for the countries with data from the Financial Crisis Surveys; “ECA” is the average for the countries in the Europe and Central Asia with ES data in 2008-2009; "All countries” is the average for all countries with ES data between 2006 and 2009. Are female-managed firms more likely to exit 2008, we can assume that unattainable rms went out of business than male-managed firms? business.9 As Figure 5 shows, female-managed rms have a Having the pre-crisis situation in mind, the rst issue similar or higher exit rate than male counterparts in all to consider is the gender dimension of exit patterns. e countries, except for Turkey.10 As shown in previous available data allow us to determine whether the rms in the full original sample of the ES were still in existence or studies (Agarwal, 1996; Audretsch, 1995; Boeri and if they had become inactive at the time of the rst wave of Bellman, 1995; Ericson and Pakes, 1995; Littunen, 2000, follow-up data collection. e sample available for this Aga and Francis, 2017), rm exit can be driven by several analysis includes 2,499 observations in the six countries, rm characteristics, including sector, years in operation, of which 2,493 rms have information on the gender of a size of the rm, years of experience of the top manager and top manager. A top manager is female for 405 rms and labor productivity. In other words, one may think that a male for 2,088 rms. higher percentage of female-managed rms exit the market Using this information, rm exit rates from 2008 to because they are in more vulnerable sectors, or because 2009 are computed. In this note the de nition of exit they are smaller, or because female managers are less consists of the sum of three types of rms: a) rms that experienced. To account for these factors, we have discontinued business or that are inactive from estimated the probability of exit (using Probit models) that information obtained during the screening process; b) incrementally add these possible explanatory variables. e rms that during the survey indicated that they have led estimations allow us to verify if female-managed rms are for insolvency or bankruptcy; and c) rms that were still more likely to exit the market than male-managed impossible to locate. Since all the rms contacted for the rms, after accounting for these factors and to measure FCS had been interviewed not long before the baseline in how likely they are to do so. We also consider 3 Firm exits from 2008 to 2009 Proportion of permanent full-time Figure 5 Figure 6 workers that are female 30% 28% 26% 25% 90% 77% 75% 18% 75% 62% 65% Percentage 20% 16% % of firms 15% 60% 54% 15% 13% 48% 11% 45% 37% 35% 9% 10% 31% 33% 32% 10% 6% 24% 30% 4% 4% 5% 15% 0% 0% Bulgaria Hungary Latvia Lithuania Romania Turkey Bulgaria Hungary Latvia Lithuania Romania Turkey FMF MMF FMF MMF Source: Enterprise Surveys (2008-2009) and Financial Crisis Surveys (2009) access to nance and the use of technology, proxied by Main e ect of the nancial crisis having a bank loan or line of credit and by having its own During the rst wave of the FCS, respondents were website, respectively, as alternative explanations of exit. asked to indicate the main e ect of the nancial crisis out Results are con rmed with female-managed rms being of a list of several options. Around 70 percent of both around 7 percentage points more likely to exit the market female and male-managed rms in each of the countries than male-managed rms in all speci cations of the covered in the survey chose "drop in demand for its model.11 product and services" as the dominant e ect of the crisis What still needs to be explored is the reason for the ( gure 7). For female-managed rms, the percentage of patterns identi ed. More research may be needed to rms that indicated "drop in demand“ as the main e ect determine if the lower productivity of female-managed varies from 66 percent in Turkey to 87 percent in rms among large rms may be a factor. Unfortunately, Bulgaria. For the male-managed rms, percentages range with the available data it’s hard to give an answer to this from 71 percent in Hungary to 78 percent in Romania. question given the very small sample of large rms among When looking at the other options, there is no remarkable the exiting rms. di erence between female-managed and male-managed e disproportionate exits of female-managed rms rms’ answers with the only exception of the option from the market because of the nancial crisis is bad news concerning “increase in the debt levels”. In all countries, not only because the reduction of the already few female-managed rms (only 20 percent of all rms at except for Hungary, female-managed rms are less likely baseline) worsens an already inequal situation; it is bad than male-managed rms to report “increase in debt level” news also because it negatively a ects female employment, as the main e ect of the nancial crisis. It is hard to given that female-managed rms hire more women than explain the reasons behind this di erence. On one hand, male-managed rms. On average, the proportion of this may be the result of lower access to nance, while on permanent full-time employees that are women is higher the other hand it may be that female-managed rms may in rms with female top managers than in rms with male not need additional funds.14 Some hints may be found by top managers across countries ( gure 6).12 e fact that looking at indicators of the e ect on rms’ nancing female-managed rms are more likely to exit the market at discussed below. the time of the crisis, therefore, implies that employment opportunities for women are also lost disproportionately. E ects on rm’s nancing During economic downturns, rms are more likely to Are firms with female top managers affected become overdue in their obligations with nancial by the crisis as firms with male top managers? institutions and to be forced to restructure their liabilities e second issue to consider is the e ect of the crisis or even le for bankruptcy. e FCS survey measures the on rms that remain in business. ree main aspects are percentage of rms overdue on obligations with any considered: main e ect reported by rms, e ect on nancial institution in the 12 months prior to the nancing, and the e ect on sales.13 interview. In all countries, except for Bulgaria and 4 Figure 7 Main effect from the crisis: drop in demand 100% 90% 80% 70% 60% % of firms 50% 40% 30% 20% 10% 0% FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF Bulgaria Hungary Latvia Lithuania Romania Turkey Increase in debt level Increase in output cost Reduced access to credit Drop in demand Other Source: Enterprise Surveys (2008-2009) and Financial Crisis Surveys (2009) Hungary, female-managed rms are less likely to be E ect on sales overdue in obligations than male-managed rms in the In the rst wave of the FCS survey, rms were asked if short run as measured in the rst wave of data collection their sales increased or decreased from June 2008 to June ( gure 8). When rms are overdue in their payments, or 2009 and by what percentage. e percentage change in they want to prevent falling into arrears, they may attempt sales discussed in this section is the net e ect computed to restructure their liabilities. Firms may restructure their using this question. As expected, in all countries both liabilities also to take advantage of better conditions in the female and male-managed rms experienced a decrease in nancial market. e FCS included a measure of the their sales ( gure 10). However, in Bulgaria, Lithuania, percentages of rms that restructured outstanding Romania, and Turkey, the net percentage decrease in sales is liabilities in the previous 12 months. In all countries, lower for female-managed rms than male-managed rms. except for Lithuania and Romania, in the short run (wave e reduction is very similar in Latvia and lower for 1), more female-managed rms have restructured their male-managed rms only in Hungary. In the longer-term nancial liabilities compared to male-managed rms (wave 3), however, female-managed rms experienced a ( gure 9). higher decrease in sales than male-managed rms across all To summarize, the data show that female-managed countries ( gure 11).15 Also in this case, as for the exit rms are less likely to experience an increase in debt due to analysis, the results were tested by estimating the probability the nancial crisis and less likely to be overdue in nancial of exit (using Probit regression analysis) accounting for obligations; however, a higher percentage of rms’ characteristics as potential explanatory variables of female-managed rms restructured their liabilities as these di erences. As in the previous section we use sector, compared to male-managed rms. e combination of years in operation, size of the rms, years of experience of the three indicators seems to suggest that female-managed the top manager and labor productivity as explanatory rms may have restructured their liabilities to take variables. e results con rm that in the short-term (wave advantage of better conditions in the markets (e.g. lower 1), female-managed rms experienced a lower decrease in interest rate) rather than as a reaction to a di cult sales than the male-managed rms by about seven nancial situation. Unfortunately, with the available data, percentage points. ese results are consistent across the it is impossible to verify if this was the case. We nd some subsequent addition of explanatory variables. Moreover, indication of the better market conditions from the similar regression analyses with wave 3 data indicate that in analysis of the long-term data (wave 3) where we nd no the long run, female-managed rms experienced a higher evidence of a di erence in the nancial distress indicators decrease in net sales than the male-managed rms by by female and male-managed rms. around 13 percentage points.16 5 Firms overdue in obligations last year Firms that restructured liabilities Figure 8 (wave 1) Figure 9 last year (wave 1) 35% 32% 32% 31% 40% 30% 27% 35% 35% 25% 28% 28% 20% 21% 30% 25% % of firms 20% 25% 23% % of firms 16% 15% 19% 15% 17% 17% 15% 13% 20% 16% 12% 9% 15% 10% 10% 10% 5% 4% 5% 5% 0% 0% FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF FMF MMF Bulgaria Hungary Latvia Lithuania Romania Turkey Bulgaria Hungary Latvia Lithuania Romania Turkey Net % change in sales in previous month Net % change in sales in previous month Figure 10 compared to same month last year Figure 11 compared to same month last year (wave 1) (wave 3) Bulgaria Hungary Latvia Lithuania Romania Turkey Bulgaria Hungary Latvia Lithuania Romania Turkey MMF MMF MMF MMF MMF MMF MMF MMF MMF MMF MMF MMF FMF FMF FMF FMF FMF FMF FMF FMF FMF FMF FMF FMF 0% 0% -5% -5% -2% -10% -15% -10% Percentage -12% -10% -14% Percentage -20% -15% -15% -15% -12% -11% -25% -20% -23% -30% -26% -25% -19% -28% -25% -35% -23% -23% -40% -36% -30% -24% -28% -45% -40% -42% -35% -31% -43% -33% -40% -36% Source: Enterprise Surveys (2008-2009) and Financial Crisis Surveys (2009) Conclusions don’t seem to su er from a lack of access to nance more than male-managed rms neither in the short term nor in is note summarizes the results of an analysis using the longer term. What deteriorated disproportionally data collected during the 2008 nancial crisis to examine more for female-managed rms were sales, maybe if female-managed rms su ered more than re ecting a more limited ability to adapt their business to male-managed rms. e ndings are very interesting: the crisis. e role of access to technology and ability to rms managed by women are more likely to exit the adapt through innovation and changes in the business market at the onset of the crisis; however, when able to model are interesting patterns to explore further in future stay in business, they endure the crisis as well as analyses as they may be key elements to focus when male-managed rms in the short term, with a designing measures to support female entrepreneurship. deterioration in the relative conditions in the longer term. e ndings from this note not only ll an evident It is di cult to identify the reasons behind these gap in the literature but also provide evidence that may be patterns. Just before the crisis, female-managed rms, useful to understand the potential challenges that women which accounted for 20 percent of all rms, showed less in business may be facing at the time of writing due to the access to nance and used less technology than their male e ects of the COVID-19 crisis. As soon as more data on peers. However, the analysis that takes all these factors the e ect of the COVID-19 pandemic are available for a into consideration still shows a higher probability of exit large enough number of countries, the analysis will be for female-managed rms. ey were also slightly less replicated to verify if ndings hold in a di erent type of productive than male-managed rms (largely driven by crisis. For those interested, the data for the nancial crisis large rms), which may partially explain the higher exit and the COVID-19 crisis are available via rates. When staying in business, female-managed rms https://www.enterprisesurveys.org/en/full-survey. 6 Notes 16 Even though adding labor productivity a ects the signi cance of the e ect of the nancial crisis on net change in sales by female and 1 e note has been committed and funded by the We-Fi Initiative. male-managed rms, the coe cient is consistently negative, 2 e literature on rms exist focuses on mainly on factors such as indicating that female-managed rms experienced a higher decline productivity and age, with higher productive rms (Olley and in sales in the long-term. Regression results are available upon Pakes 1996; Fariñas and Ruano 2005; Aga and Francis, 2017) and request. older rms (Jovanovic 1982; Pakes and Ericson 1998) being at a lower risk of exiting the market. 3 In all cases, participation in the nancial crisis survey was voluntary References for all sectors included in the original sample, except for Turkey Aga, G., Francis, D., 2017. As the market churns: productivity and where only the original manufacturing sub-sample was targeted. rm exit in developing countries, Small Business Economics 49 (2), 4 As robustness check the analysis have been replicated for rms with 379-403 female participation in ownership and results hold. Results are Agarwal, R., 1996. Technological activity and survival of rms. available upon request. Economics Letters 52, 101–108. 5 e di erence is statistically signi cant at 5% level. Audretsch, D., 1995. Innovation, growth and survival. International 6 e di erence is statistically signi cant at 5% level. Journal of Industrial Organization 13, 441–457. (1996), 101–108. 7 Female-managed rms were slightly more likely to have a bank Benería, L., 2015. Gender Perspectives and Gender Impacts of the loan or line or credit and less likely of not needing a loan than Global Economic Crisis, Feminist Economics, 21:4, 206-210 male-managed rms among manufacturing rms; these Boeri, T., Bellman, L., 1995. Post-entry behaviour and the cycle: di erences, though, are not statistically signi cant. evidence from Germany. International Journal of Industrial 8 e di erence is statistically signi cant at 1% level. Organization 13, 483–500. 9 In total, 284 rms are de ned as exiting rms of which 224 are Cesaroni, F., Sentuti, A., Buratti, A., 2015. Same crisis, di erent male-managed and 60 are female-managed rms. Exit rate is the strategies: Italian men and women entrepreneurs in front of the percentage of male and female-managed rms that exit the market economic recession. Journal of Research in Gender Studies, 5(2), as percentage of total male and female-managed rms. 205-231 10 As mentioned earlier, Turkey’s sample is composed mostly of Cesaroni, F.M., Lotti, F. and Mistrulli, P.E., 2013. Female Firms and manufacturing which tends to be a male dominant sector and that Banks’ Lending Behaviour: What happened during the Great can be part of the reasons why it exhibits a di erent pattern. Recession?. Bank of Italy Occasional Paper, (177). 11 Results are signi cant at 5% and 10% depending on the Ericson, R., Pakes, A., 1995. Markov-perfect industry dynamics: a speci cation. e positive sign is retained but we lose signi cance framework for empirical work. e Review of Economic Studies in the speci cation that controls for productivity. Due to missing 62, 53–82. information in the variables needed to compute productivity, in Fariñas, J., Ruano, S., 2005. Firm productivity, heterogeneity, sunk this speci cation we lose as many as 53 exiters ( rms that exit the costs, and market selection. International Journal of Industrial market) and more than 400 observations compared to the Organization, 23(7–8), 505–534. preceding speci cation. Regression results are available upon Hallward-Driemeier M. and B. Rijkers (2013). Do crisis catalyze request. creative destruction? Firm-level evidence from Indonesia. e 12 ese di erence holds for female and male-managed rms in Review of Economic and Statistics, 95(5), 1788-1810. manufacturing, retail, and other services and for small, medium Hallward‐Driemeier, M., Rijkers, B. and Waxman, A., 2017. Do and large rms. Employers' Responses to Crises Impact Men and Women 13 e sample for this section of the analysis is comprised of a total of Di erently? Firm‐level Evidence from Indonesia. Review of 1,682 observations in wave 1 and 1,392 rms in wave 3. e Development Economics, 21(4), pp.1018-1056. sample distribution in wave 1 of FCS by female and male managed Jovanovic, B., 1982. Selection and the evolution of industry. rms (total) are: 32 and 118 (150) rms in Bulgaria, 17 and 170 Econometrica, 50(3), 649–670. (187) rms in Hungary, 80 and 146 (226) rms in Latvia, 38 and Lahey, K.A., de Villota, P., 2013. Economic Crisis, Gender Equality, 201 (239) rms in Lithuania, 61 and 306 (367) rms in Romania, and Policy Responses in Spain and Canada, Feminist Economics, and 46 and 467 (513) rms in Turkey, respectively. 19:3, 82-107 e sample distribution in wave 3 of FCS by female and male Lim, J.Y., 2000. e e ects of the East Asian crisis on the employment managed rms (total) are: 31 and 121 (152) rms in Bulgaria, 17 of women and men: e Philippine case. World Development, and 134 (151) rms in Hungary, 75 and 131 (206) rms in Latvia, 28(7), pp.1285-1306. 35 and 182 (217) rms in Lithuania, 52 and 250 (302) rms in Littunen, H., 2000. Networks and local environmental characteristics Romania, and 40 and 324 (364) rms in Turkey, respectively. in the survival of new rms. Small Business Economics 15, 59–71. 14 e sensitivity of the responses on the main e ect of the nancial Olley, S., Pakes, A., 1996. e dynamics of productivity in the crisis in increasing the debt levels for female and male-managed telecommunications equipment industry. Econometrica, 64(6), rms was also tested by estimating probit regressions controlling 1263–1297. for relevant characteristics of rms (sector, size, age of rms, years Pakes, A., & Ericson, R. (1998). Empirical implications of alternative of experience of the top manager, and labor productivity). e models of rm dynamics. Journal of Economic eory, 79(1), 1–46. results hold. Seguino, S., 2010. e global economic crisis, its gender and ethnic 15 Results were tested by using t-tests that show di erences in Latvia, implications, and policy responses, Gender and Development, 18:2, Lithuania, and Turkey are statistically signi cant with the net 179-199 percentages decrease in sales being higher for female-managed Varum, C.A., Rocha, V.C.. 2012. " e e ect of crises on rm exit and rms by 18.9, 24.8, and 26.1 percentage points than those of the moderating e ect of rm size," Economics Letters, Elsevier, vol. male-managed rms, respectively. 114(1), pages 94-97. e Enterprise Note Series presents short research reports to encourage the exchange of ideas on business environment issues. e notes present evidence on the relationship between government policies and the ability of businesses to create wealth. e notes carry the names of the authors and should be cited accordingly. e ndings, interpretations, and conclusions expressed in this note are entirely those of the authors. ey do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its a liated organizations, or those of the Executive Directors of the World Bank or the governments they represent. 7