WPS5571 Policy Research Working Paper 5571 Gender and Finance in Sub-Saharan Africa Are Women Disadvantaged? Reyes Aterido Thorsten Beck Leonardo Iacovone The World Bank Africa Region Private and Financial Sector Development February 2011 Policy Research Working Paper 5571 Abstract This paper assesses whether there is a gender gap in the sole proprietorships than men, and firms with female use of financial services by businesses and individuals in ownership participation are smaller, but more likely to Sub-Saharan Africa. The authors do not find evidence innovate. In the case of individuals, the lower use of of gender discrimination or lower inherent demand for formal financial services by women can be explained financial services by enterprises with female ownership by gender gaps in other dimensions related to the use participation or by female individuals when key of financial services, such as their lower level of income characteristics of the enterprises or individuals are taken and education, and by their household and employment into account. In the case of enterprises, they explain this status. finding with selection bias--females are less likely to run This paper is a product of the Private and Financial Sector Development, Africa 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://econ.worldbank.org. The authors may be contacted at raterido@worldbank.org, Liacovone@worldbank.org, and T.Beck@uvt.nl. 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 Gender and Finance in Sub-Saharan Africa: Are Women Disadvantaged? Reyes Aterido, Thorsten Beck and Leonardo Iacovone* JEL Classification: G21; J16 Keywords: Access to financial services; Gender; Entrepreneurship * Aterido and Iacovone: World Bank (raterido@worldbank.org and Liacovone@worldbank.org); Beck: CentER and EBC, Tilburg University and CEPR, corresponding author: T.Beck@uvt.nl. We are grateful to Judith Frickenstein, Mary Hallward-Driemeier and Amrik Heyer for useful comments and suggestions. We would like to thank the Finmark Trust in South Africa and the Financial Sector Deepening Trusts in Kenya and Tanzania for sharing data with us. We are grateful for financial support from the BNPP Trust Fund. This paper was prepared for the World Bank regional flagship "Expanding Opportunities for Women Entrepreneurs in Africa." This papers findings, interpretations, and conclusions are entirely those of the authors and do not necessarily represent the views of the World Bank, its Executive Directors, or the countries they represent. 1. Introduction Access to and use of financial services by both enterprises and households is of increasing concern to policy makers across Africa and the developing world. Recent data collection efforts on both the enterprise and household levels have enabled a more rigorous analysis (World Bank, 2007). One important dimension in the access to finance debate, which has been less analyzed, is the gender gap. Specifically, it has often been argued that lack of access to finance impedes female entrepreneurship and prevents women from participating in the modern market economy. Given the overall lack of financial service provision, with fewer than one in five households having access to formal financial services, this problem is potentially more pressing in Sub-Saharan Africa than in other developing regions of the world (Honohan and Beck, 2007). As documented by an extensive and still growing literature, access to credit is important for firm growth, especially that of small firms (Beck, Demirguc-Kunt and Maksimovic, 2005), and for new business creation (Klapper, Laeven and Rajan, 2006). Country-specific studies and randomized field experiments confirm that access to capital can be critical for firm growth (Banerjee and Duflo, 2008; De Mel, McKenzie and Woodruff, 2008). However credit is not the only financial service that seems to matter. Recent evidence shows that access to savings services can also increase enterprise investment, especially among female entrepreneurs (Dupas and Robinson, 2009). Broad access to financial services is not only important for individuals, but also for the economy at large; credit constraints reduce the efficiency of capital allocation and intensify income inequality by impeding the flow of capital to poor individuals with investment opportunities with high expected returns (Galor and Zeira, 1993; Aghion and Bolton, 1997; Galor and Moav, 2004, Beck, Demirguc-Kunt and Levine, 2007; Lopez and Serven, 2009). Gender differences in 1 access to financial services can therefore have direct negative repercussions not only for female entrepreneurs and individuals but for the overall economy. This paper analyzes gender differences in access to credit by enterprises and use of formal and informal financial services by individuals in Sub-Saharan Africa. Specifically, we use enterprise surveys to assess whether female entrepreneurs in Sub-Saharan Africa are less likely to rely on formal bank finance, compared to male entrepreneurs and compared to female entrepreneurs outside Sub-Saharan Africa. We also use recent FinScope and FinAccess surveys across nine countries in Southern and East Africa to assess gender differences in the use of formal and informal financial services by individuals. This paper relates to a growing literature on the gender gap in access to credit (see Klapper and Parker, 2010, for a survey). Cross-country studies have shown that women are less likely to get financing from a formal financial institution or are charged higher interest rate than men (Muravyev, et al., 2007) and generally raise less formal and informal venture capital than men (Brush, et al., 2004). Bruhn (2009), on the other hand, does not find any evidence for Latin America of a gender gap in access to credit by enterprises. Richardson, Howarth and Finnegan (2004) find for Sub-Saharan Africa that women entrepreneurs are more likely than male entrepreneurs to rely on internal or informal financing. This gender gap is also reflected in higher financing obstacles reported by women. The literature has also explored the reasons behind such a gender gap. Buvinic and Berger (1990) find that female entrepreneurs struggle more with loan applications, while Lusardi and Tufano (2009) find lower overall financial literacy among women. However, behavioral differences might also be important, leading to taste rather than statistical discrimination (Beck, Behr and Madestam, 2011). Evidence from Africa shows that in many instances, only male heads of households are able to successfully receive formal credit (Johnson, 2004). 2 Among institutional factors explaining gender differences in access to credit might be property right restrictions for women. Such restrictions might include requirements for married women to obtain their husbands signature and approval for all banking transactions.1 Women can also be affected by a husbands adverse credit history, which might require his wife to repay the debt or be denied credit (Naidoo and Hilton, 2006). The observation of a gender gap has led many NGO supported microcredit institutions to focus on women rather than men. Given the limitations of microcredit, both in volume and in outreach, however, it is important to understand differences in the use of formal banking services (Honohan, 2004). In addition to formal financial services, many individuals and enterprises across the developing world use informal financial services, ranging from money lenders to ROSCAS. In our empirical work, we will therefore also consider the use of informal financial services by households. The empirical findings of the enterprise analyses show that firms with female ownership participation do not seem to have worse access to credit than firms with purely male ownership, neither within Sub-Saharan Africa nor across the world. However, we do find that African enterprises are more financially constrained and that larger firms have more access to financial services. We argue that the apparently surprising results of lack of gender difference in access to finance can be actually explained by the existence of a selection process. First, female entrepreneurs are less likely to own sole proprietorships than men and face higher regulatory burden than men, especially in Africa. Second, enterprises with female ownership are smaller and therefore have less access to the financial markets, and women are less likely to be entrepreneurs. Third, female owned enterprises are more likely to innovate what could be explained by the fact that female entrepreneurs need to be especially capable in order to be able to enter the formal sector and in fact have characteristics that make them 1 See Women, Business and the Law; "Improving the Legal Investment Climate for Women in Africa" (Hallward- Driemeier, 2011`) for more detailed information and broader coverage of countries in Sub-Saharan Africa. 3 more attractive for financial institutions. Finally, we find some limited support for the hypothesis of a "sectoral selection" as female ownership tend to be more prevalent in sectors that, on average across countries, tend to rely less on access to external finance. This is consistent with Gajigo and Hallward-Driemeier (2010) who find suggestive evidence, in four African countries, that there is a gender gap in capital at the start up. Although start-up capital gender difference is higher along sector than by gender, the median capital for male entrepreneurs is more than twice that of the female entrepreneurs. These selection biases might explain that once we focus on a sample of existing enterprises and control for firm size, sector and other firm characteristics we do not find any significant effect of female ownership participation. The household analysis shows that while unconditional comparisons present a lower use of formal banking services by women, there is no significant gender difference once we control for other individual characteristics, including education, income, work status, geographic location, and education. While gender differences in the use of informal financial services vary across countries, women are not more or less likely to be excluded from any financial service than men, at least in our sample of nine countries. Lower income and education, a lower likelihood to be formally employed and their role within the household explain why, prima facie, women are less likely to use formal financial services. These barriers that women face as individuals to access formal financial services might also explain the selection bias among female entrepreneurs mentioned above, though we cannot formally test for it. On the other hand, we find that women in several countries are more likely to use informal financial services. Our paper contributes to the literature on access to finance along several dimensions. First, while most studies so far have been limited to one country, this is a cross-country exploration of gender differences in access to and use of financial services, both across 4 enterprises and households.2 Second, this paper considers access to and use of all financial services, not just credit as done in large parts of the literature. In addition, it also looks beyond formal financial services to informal financial services. Third, this paper contributes to the literature on gender differences in Sub-Saharan Africa. As rigorous analysis for Sub- Saharan African is often impeded by the lack of appropriate data, the data compilations used in this paper offer a unique opportunity to explore gender differences in the participation in formal and informal economies in Sub-Saharan Africa. The remainder of this paper is organized as follows. The next section focuses on the gender gap in enterprise finance using a large cross-section of firm-level surveys. Section 3 presents results on the gender gap in the use of financial services by individuals and section 4 concludes. 2. Gender and Enterprise Access to Credit This section assesses whether there is a gender gap in enterprises access to and use of bank finance. Specifically, using firm-level survey data, we will assess whether businesses with female ownership participation (i) are less likely to use a formal financing channel (e.g. overdraft or loan), (ii) have a lower share of investment financed by financial institutions, and (iii) have a lower share of working capital financed by financial institutions. The section first describes the data and then turns to multivariate regressions. 2.1. Data To explore the relationship between gender and enterprise access to credit, we use the World Bank-IFC Enterprise Surveys. The Enterprise Surveys have been conducted over the 2 See World Bank (2007) for an overview of studies on access to finance. 5 past eight years in over 100 countries with a consistent survey instrument.3 The surveys try to capture business perceptions on the most important obstacles to enterprise operation and growth, but also include detailed information on management and financing arrangements of companies. Sample sizes vary between 250 and 1,500 companies per country and data are collected using either simple random or random stratified sampling. The sample includes formal and informal enterprises of all sizes, different ownership types and across nine sectors in manufacturing, construction, services and transportation. Firms from different locations, such as capital city, major cities and small towns, are included. Our empirical work relies on data for formal enterprises from 37 African countries and 49 countries from the rest of the world covering in total 35,000 firms during 2006 to 2009 (Appendix Table A1).4 In addition, we focus on informal mostly micro-firms for a sample of 25 African countries. The Enterprise Surveys offer several advantages for our purpose. First, the surveys collect comparable information for several firm characteristics across all the countries. This comparability allows us to document cross-country and within-country profiles of firms that have female ownership participation. Second, the surveys collect information on financing at the firm level as well as several other relevant firm characteristics. These include the firm size and age, human capital composition of the workforce, measures of technology adoption and firms international openness, i.e. export activity and sources of capital. In addition, there is also detailed information on the firms geographical location and its sector of activity (3- digit-ISIC classification). Third, the surveys reach a substantial number of countries across all 3 See www.enterpriseseurveys.org for more details. Similar surveys were previously conducted under the leadership of the World Bank and other IFIs in Africa (RPED), and world-wide in 2000 (World Business Environment Survey). Enterprise Surveys still go under different names in some regions, e.g. BEEPS in the Central Asia and Eastern European transition economies had first been launched in 1999 and was then modified to be comparable with the broader global initiative. 4 We lose 25 countries when using the share of working capital financed with external finance, mostly in Eastern and Central Europe and East Asia. There are a total of 95 surveys but some countries enter the dataset in multiple years. Information collected is standardized across countries and covers formal establishments with 5 or more employees. However, Africa, East Europe and Central Asia, and Latin America include establishments with fewer than 5 employees. In Africa, there are an additional 25 standardized surveys covering micro-firms and targeting informal establishments. 6 the regions of the world. 30.4 percent of our formal sample is in Africa, 33.5 percent in Eastern and Central Europe, 32.4 percent in Latin America, and 3.7 percent in Asia. Thirty- six percent of sample firms have female ownership participation; 28 percent in Africa and 39 percent in the rest of the world. Table 1 Panel A provides descriptive statistics of the variables used in the analysis. Thirty-six percent of firms in our sample have at least one female owner. Thirteen percent of formal firms across the sample have 5 or fewer employees, while in Africa, these represent 21 percent of the firms. In the overall sample, 30 percent of enterprises are sole proprietorships, 13 percent of firms export at least 10 percent of their sales and 10 percent have foreign owners holding at least 10 percent of capital. In terms of location, 40 percent are located in the capital or cities with populations of 1 million or more. The average age of firms is 16 years. We also find that enterprises finance, on average, 18 percent of their fixed asset investment with bank finance and 9 percent of their working capital needs with bank finance, while, on average, 55 percent of firms have a bank loan or overdraft facility. Table 1 Panel B shows the sample distribution by gender and type of company. If we focus on formal companies we confirm that in Africa only 28 percent of companies have female ownership participation, while the percentage in the entire sample (including Africa) is 36 percent. The percentage of informal companies with female ownership participation in Africa is 33 percent.5 Finally, when focusing on formal sole proprietorships (often micro and small ones) the difference in prevalence of female ownership participation between Africa and the entire sample is only three percentage points (24 percent in Africa versus 27.1 percent in the entire sample). In our multivariate analysis, we will present results for the whole sample as well as for a subsample of sole proprietorships, as this allows us to isolate enterprises that are completely run by a female owner, as opposed to other firms where some 5 We cannot compare this to the rest of the world because informal sector surveys where not performed in other regions outside Africa. 7 but not necessarily all of the owners are female. As Hallward-Driemeier and Aterido (2010) show, up to half of the firms that have multiple owners of which at least one is female, do not have women among their prime decisions makers. So ,,female participation in ownership defines a wider circle of firms as ,,female, and thus may lead to lower bound estimates on the extent to which gender may matter. Looking at sole proprietors does address the ownership- decision maker distinction, but the average size of the firm is smaller. 2.2. Multivariate Regression Results We start from estimating the equation below where our outcome variable of interest is measured at the level of the firm i in country c belonging to sector s at time t. (1) where y is one of our three financing variables ­ (i) a dummy variable that takes value one if the firm has access to a formal financing channel (e.g. overdraft or loan), (ii) the share of fixed asset investment financed by financial institutions, and (iii) the share of working capital financed by financial institutions. Fem is a dummy variable indicating female ownership participation, Afr is a dummy for countries in Sub-Saharan Africa, and Empl is the log of the number of employees, thus an indicator of enterprise size. In addition, we include industry and time fixed effects in order to control for industry characteristics and time- specific global shocks. We also include macroeconomic characteristics such as the inflation rate, GDP per capita and average growth of GDP in previous three years. In robustness tests, we add an additional array of firm characteristics; the age of the company, a dummy variable identifying location in a large city6, a dummy variable identifying exporters and a dummy variable 6 Defined as capital cities or with population of one million or more. 8 identifying foreign ownership. Standard errors are clustered at the country-year level, thus allowing for unobserved variables driving correlation in financing across enterprises in a specific survey. All models are estimated with OLS due to the difficulty of interpreting the marginal effects of interaction terms in non-linear models (Ai and Norton, 2003). However, re-estimating the model using a Tobit or Probit model generates qualitatively similar results (these results are available upon request). This flexible specification in regression (1) allows us to analyze several questions: 1. The coefficient 1 indicates whether enterprises with female ownership participation are more financially constrained than other companies. 2. The coefficient 4 indicates whether the financing situation of businesses with female ownership participation in Sub-Saharan Africa is different from that of businesses with female ownership participation in other parts of the world. 3. The coefficients 5and 7 indicate whether these effects are different for firms of different sizes across the world (5) and specifically in Africa (7). Table 2 presents the results for the broad cross-country sample, while Tables 3 to 5 present results for African countries only. Given the cross-sectional nature of our results, they should not be interpreted in a causal manner as they only present conditional correlations. The results, reported in Table 2, columns 1-4, show that, on average, there is no statistically significant difference in terms of access to finance between companies with and without female ownership participation, including in a subsample of sole proprietorships. This relationship does not vary between African and non-African countries, but it varies across firms of different size. In fact, size does not only matter for firms with female ownership participation, but it matters in general, as larger enterprises are more likely to have access to external finance. A ten percent increase in firm size translates into a 6 to 6.5 percent higher probability of accessing formal finance across all firms and 9.6 percent higher 9 probability for sole proprietorships. The economic effect is even bigger for firms in Africa compared to firms elsewhere, with firms in Africa facing a 46 percent lower likelihood of accessing formal finance. The size effect is larger for firms with female ownership participation and for firms in Africa as shown by positive interaction terms between the female dummy and employment and the Africa dummy and employment, respectively; however, the latter results does not hold for sole proprietorships. We do not find any differential effect of the size-Africa interaction across firms with and without female ownership participation. We also find that privately-owned firms with limited liability, older firms, companies located in larger cities, and exporting firms are more likely to access external finance, although many of these variables turn insignificant in the sub-sample of sole proprietorships. The results in columns 5-8 of Table 2 show that, while on average there is no statistically significant difference in the share of investment financed by financial institutions between businesses with and without female ownership participation, this is not the case in Africa. In fact, the interaction between Africa and female ownership participation is positive and marginally significant. This is, at first sight surprising, however this result does not hold for a sample of sole proprietorships, suggesting that only those companies where a female is one among various owners tend to finance a larger share of their capital using bank finance, but that this is not the case for companies solely owned by a female. Similarly, we find there is no statistically significant difference in the share of working capital financed by financial institutions between businesses with and without female ownership participation (columns 9- 12). There is some marginal evidence that female sole proprietorships use a higher share of working capital from financial institutions, though this is significant only at the 10 percent level (columns 11 and 12). However, one strong and robust finding is that Sub-Saharan African businesses appear to be less able to access formal financial institutions to finance 10 both investment and working capital, a remarkable difference between six and 14 percentage points compared to businesses in other parts of the world. Consistent with previous research (Beck, Demirguc-Kunt and Maksimovic, 2008), the results also show that larger companies have a significant advantage in terms of accessing financial institutions to finance both their investments and working capital; an increase in a firms size by 10 percent is associated with a 17-18 percentage point increase in the share of its investment or working capital financed by financial institutions. These results, however, turn insignificant for the sample with sole proprietorship because within this sub-sample the size variation is much smaller as this type of companies tends to be characterized by a limited size. Furthermore, we show that companies located in a large city tend to finance a larger share of their investment and working capital with bank finance, potentially driven by supply of finance more available in larger cities as well as "demand" effects driven by the type of companies that locate in larger cities. The share of external finance for investment of companies located in a large city is ­ on average - eight percentage points larger than that of other companies, while it is just four percentage points larger when focusing on sole proprietorships. Similarly, the share of working capital financed externally for companies located in a large city is ­ on average - five percentage points larger for all firms, and 2.5 percentage points for sole proprietorships. Having analyzed the results for the entire sample composed by companies in various regions, we focus on a sample of companies located in Africa. With this objective in mind we test a simpler model described by regression (2) below (2) As before Fem is a dummy variable indicating female ownership participation, and Empl is the log of the number of employees. Similarly, as before, we include industry and time fixed effects, as well as macroeconomic control variables. In robustness tests, we add firm 11 characteristics, and always cluster the standard errors at the country-year level to account for autocorrelation. In addition, when focusing on African countries we are also able to separate the sample into formal versus informal enterprises. In the case of informal enterprises, however, we do not include firm size and its interaction with female, as only 2 percent (or 63 firms) of informal enterprises have more than five employees and the sample of informal enterprises presents thus very limited variation. The results in Table 3 show the absence of a statistically significant difference in terms of accessing external finance between companies with and without female ownership participation in Africa. This result holds for both formal and informal companies, as well as for sole proprietorships. The significant positive coefficient in column 5 for informal firms turns insignificant once we control for other firm characteristics in a smaller sample. Furthermore, we confirm once again that larger companies, as well as older companies are more likely to access external finance. A ten percent increase in size translates into a 9.5 to 11 percent higher probability of accessing formal finance in the overall sample and a seven percent higher probability for sole proprietorships. The size effect is stronger for female sole proprietorships. Other firm-level characteristics do not enter consistently with a significant coefficient. The results in Table 4 confirm that there is no statistically significant difference in terms of the share of external finance for investment between companies with and without female ownership participation. These results apply both to formal and informal companies, and are robust in the sub-sample of sole proprietorships. Additionally, as before we find that larger companies tend to be more likely to finance their investments with external resources; an increase in the size of a company by 10 percent is correlated with an increase in the use of external finance for investments by more than 20 percentage points and more than 13 percentage points in the case of sole proprietorships. Unlike in the case of Table 3, the size 12 effect does not vary across sole proprietors of different genders. None of the other firm characteristics enters significantly in Table 4. The results presented in Table 5 confirm our previous findings that firms with female ownership participation use as much external finance for working capital as other firms in Africa, a result that is consistent across different sub-samples of formal and informal firms and of sole proprietorships. As before, we find that larger firms finance a significantly larger share of their working capital from financial institutions. The result that companies with female ownership participation do not tend to be disadvantaged in accessing formal financing channels to cover the costs of investments, or working capital, seems at first rather surprising. One possible reason for this may be the existence of a selection bias. Such a selection bias would imply that females are discriminated against, de facto or de jure, when trying to establish and run a formal company in the first place, so that female entrepreneurs must be particularly capable or, in other words, must have characteristics that set them apart from male entrepreneurs owning companies with similar characteristics. Gajigo and Hallward-Driemeier (2010) find suggestive evidence in four African countries that there is gender differences in capital at the start up. Although differences are higher along sector than by gender, the median capital for male entrepreneurs is more than twice than female entrepreneurs. This is a suggestive indication that female entrepreneurs may face larger entry barriers than their male counterparts, but that once they enter, they do not face larger constraints. We assess the possibility of such a bias in four ways. First, we analyze if indeed it is the case that African countries tend to be characterized by a degree of gender discrimination that is different from other regions. Figure 1 shows how African countries are ranked in terms of gender discrimination as measured by the Women, Business and the Law index with respect to other countries in the world. This index, 13 developed by the World Bank Group, varies between 0 and 1. It is constructed by averaging 9 dummy variables that have a value of 1 if there is gender equality in a specific area and a 0 if there is not; lower values indicate therefore more pronounced gender discrimination. The dimensions include equality in Law regarding ownership rights, inheritance, capacity before the law, rights of married men compared to married women, as well as a set of work related issues such as tax liability, industry or work hours discrimination, and within these issues, discrimination towards women who are pregnant or nursing. Figure 1 shows that the African countries have a Women, Business and the Law index substantially lower than that of the rest of the world, with the average index value for African countries being 0.5, while for the rest of the world it is 0.85. Second, we evaluate to what extent females are likely to establish a formal business. For this purpose, Table 6 shows the share of female and male entrepreneurs across different legal ownership types. The share of female-owned firms is low in the case of sole proprietorships compared with other types of firms. Only 23 percent firms with female ownership participation are sole proprietorships compared to 34 percent of firms with male ownership participation. This seems to indicate that in our sample women are less likely to be entrepreneurs and more likely to be one of several owners of an enterprise. Third, we test whether certain enterprise characteristics differ by gender ownership participation. While we cannot test for gender differences in inherent characteristics, we can test for observable differences, such as size or the tendency to innovate. We therefore estimate the following regression where the dependent variables are indicators of size, product innovation and process innovation. (3) 14 The results in Table 7 Panel A point to some significant differences across firms with and without female ownership participation in terms of size and tendency to innovate. The results in Columns 1-4 suggest that there are important differences between enterprises with and without female ownership participation in terms of size: all things equal, firms with female ownership participation tend to be significantly smaller, a result that holds both for the overall sample and the sub-sample of sole proprietorships.7 Having female ownership participation reduces the number of employees, on average, between 8 and 14 percent. Having found that size is a key factor positively correlated to access to finance, this may explain how, by being small, enterprises with female ownership participation tend to be less likely to access finance, while at the same time, we do not find any statistically significant difference between businesses with and without female ownership participation once we control for size. Interestingly, African firms are, on average, not smaller, while sole proprietorships are smaller in Africa. The interaction of the African and female dummy variables is insignificant, suggesting that the smaller size of firms with female ownership participation is not more or less pronounced in Africa than in other parts of the world. In addition, the results in columns 5-12 show that enterprises with female ownership participation are significantly more likely to innovate than enterprises without female ownership participation, both with respect to product and process innovation. On average, firms with female ownership participation are three to six percent more likely to innovate than other firms. These results hold not only for the overall sample of enterprises, but also for the sub-sample of sole proprietorships, though the significance of the female dummy drops to 10 percent in the regressions of process innovation (columns 11 and 12). In addition, the interaction terms between the Africa and the female dummy variables enter negatively and significantly in the regressions for sole proprietorships, suggesting that female-owned 7 Bruhn (2009) reports similar findings for a sample of Latin American countries. 15 sole proprietorships in Africa are not more likely to innovate than male sole proprietorships in Africa. We also find that, on average, firms in Africa are significantly less likely to innovate. It is important to interpret these findings with caution, though, as the sample for which we have information available on innovation is significantly smaller than our overall sample, both in terms of countries (31) and in the number of firms within these countries. Finally, we test whether female entrepreneurs are more likely to be active in sectors with lower needs for external finance. Since the regressions in Tables 2 ­ 5 focus on intra- industry variation (as they include industry fixed effects), they are not able to pick up such a selection bias. We therefore run the following regression (4) where ­ as above - Fem is a dummy variable indicating female ownership participation and Finance is the average of all firms in sector s across all sample countries: (i) a dummy variable that takes value one if the firm has access to a formal financing channel (e.g. overdraft or loan), (ii) the share of fixed asset investment financed by financial institutions, and (iii) the share of working capital financed by financial institutions. By using averages across countries, we are able to control for reverse causation to a certain extent. The coefficient thus indicates whether enterprises with female ownership participation are more or less likely to operate in sectors that ­ on average ­ use more external finance. By including country fixed effects, we control for country differences in access to external finance. The results in Table 7 Panel B indicate that firms with female ownership participation are indeed less likely to operate in sectors where a higher share of firms uses a loan or overdraft facility. This finding holds both across the complete sample and the subsample of sole proprietorships. On the other hand, firms with female ownership participation are not less likely to operate in sectors that use a higher share of external finance for investment or working capital. 16 We interpret the results of Tables 6 and 7 as, on the one hand, partially supporting our hypothesis of a selection bias among female entrepreneurs who have to be more capable than their male counterparts in order to be part of the formal enterprise universe. On the other hand, these results might point towards the fact that a key channel that may explain the apparent discrimination between businesses with and without female ownership participation could be size, with businesses with female ownership participation more likely to be smaller and therefore for this reason less likely to access formal external finance. Furthermore, firms with female ownership participation are more likely to operate in sectors where firms ­ on average across countries ­ are less likely to have a loan, which might point to demand side constraints. Concluding, this section shows that while African companies are less likely to access formal financial institutions to finance their investments and working capital needs, enterprises with female ownership participation do not appear more financially constrained than firms without female ownership participation. However, we find that while companies in Africa tend to have substantially more limited access to external finance, larger businesses have systematically better access to external finance, and companies with female ownership participation tend to be smaller than their counterparts owned purely by males. In addition, we find some support towards the fact that this lack of difference in terms of accessing finance could be partially explained by the fact that the female entrepreneurs appear to be a "selected sample" with characteristics that may explain our findings. To start with, females tend to be less likely to be owners of a formal company, and once they are able to break this "glass ceiling" it is because female entrepreneurs appear to be significantly more likely to innovate both in terms of product and process. 17 3. Gender and Household Access to Financial Services This section explores gender differences in the use of different financial services by individuals. Unlike in the previous section, we therefore focus more on savings and payments than credit services. We consider both formal and informal financial services. This section first discusses the data and simple comparisons in the use of financial services by men and women, before turning to multivariate regressions. 3.1. Data and Ocular Econometrics To explore the relationship between gender and use of financial services, we use 11 household surveys across 9 Sub-Saharan African countries, co-branded as FinScope or FinAccess surveys. Specifically, we have data for Botswana, Kenya, Malawi, Namibia, Rwanda, South Africa, Tanzania, Uganda and Zambia. These surveys, first undertaken in South Africa in 2002, are surveys with up to 7,600 observations and sampled with cluster stratified probability. They are based on individuals rather than households. While this might reduce the accuracy in terms of financial services that the individual has indirect access to through other household members and might reduce the representativeness, it has the advantage that we can focus specifically on the gender gap (Cull and Scott, 2010). For this study, we have a total of 11 surveys available, with Kenya and Tanzania having undertaken surveys twice. While South Africa has undertaken yearly surveys, we only include the one from 2008. All surveys used in this section were undertaken between 2004 and 2009. The FinScope surveys distinguish between four different population segments ­ (i) users of formal banking services, (ii) users of other formal services, such as insurance companies, mobile phone services and regulated MFIs, (iii) users informal financial services, including unregulated SACCOs, ASCAs and ROSCAs, and (iv) individuals excluded from any service. There are two ways to explore this differentiation; first, considering each strand 18 separately, i.e. users of formal banking, other formal and informal financial services and, second, considering the most formal strand. In the latter case, one takes into account that many users of formal financial services also use informal financial services and so focuses on the most formal form of financial service. In the following, we will follow the first approach, focusing on (i) users of formal banking services, (ii) users of informal financial services (who could also use formal banking services) and (iii) individuals excluded from formal and informal services. Figure 2 shows that, on average, women are less likely to use formal financial services than men, while gender differences in the use of informal financial services vary across countries. Here, we graph the share of surveyed in each country that (i) uses formal banking services, (ii) uses informal services, and (iii) is excluded from any financial service, separately for men and women. All observations are weighted according to their representativeness. Panel A shows that women are less likely to use formal banking services across all 11 surveys, although the gender differences vary across countries. In Botswana (2004), the gender gap is less than two percent and not significant, whereas in Kenya (2009), the gender gap is 11 percent, with 32 percent of men using formal banking services, but only 21 percent of women. The Panel A graphs also show the large cross-country variation in use of formal banking services, documented elsewhere, with almost 70 percent of surveyed in South Africa using formal banking services, while only 17 percent using formal banking services in Zambia (2005). The Panel B graphs show that the gender gap in the use of informal financial services varies across countries. In Kenya (2006 and 2009), South Africa (2008) and Tanzania (2006 and 2009), women are more likely to use informal financial services than men, while the reverse holds in Namibia and Rwanda. The graphs also show the wide-spread variation in the 19 use of informal services across the nine sample countries, ranging from 50 percent in Kenya to only 1 percent in Namibia. The Panel C graphs show that women are either as likely or more likely to be excluded from any financial service as men. Specifically, in Malawi, Namibia, Rwanda, Tanzania (2006), Uganda and Zambia, they are more likely to be excluded from any financial service, while in the other countries there is no significant difference between men and women. These graphs also indicate the high degree of financial exclusion across Southern and Eastern Africa, ranging from 80 percent in Zambia to 19 percent in South Africa. These findings are consistent with the hypothesis of a gender gap in the use of formal banking services. However, they do not control for other individual characteristics. Next, we will therefore turn to multi-variate regressions to explore whether this unconditional differences still hold once we control for other factors that can explain the use of formal and informal financial services. 3.2. Multiavariate Results We next turn to multivariate regression analysis to explore whether the gender differences in the use of financial services hold when we control for other characteristics of individuals and households. Specifically, we use probit regressions of the following form: (5) where y is access to financial services measured by the use of (i) formal banking services and (ii) informal financial services. In addition, we use (iii) a dummy variable that indicates whether a person is financially excluded, i.e. uses neither formal nor informal financial services. The regression is weighted and stratified on the rural-urban level. The coefficient of interest is , which indicates whether women are more or less likely to use financial services. We run these regressions both survey-by-survey as well as a pooled version with 20 all available surveys and dummy variables for each survey. While the results from the pooled regressions gives us an indication of the average effect across surveys, they do not allow for slope differences across countries. In addition to using country-specific weights, we weight by the inverse of the respective population in each country in the case of the pooled regression. We include a wide array of other individual characteristics that might explain variation in the use of financial services (see Table 8). Appendix Table A2 presents the descriptive statistics for all characteristics, for each country. First, we control for geographic location by including a dummy variable Rural indicating whether the individual lives in a rural district. Geographic barriers such as larger distance to the nearest bank office would suggest a negative relationship between Rural and use of formal financial services, while the use of informal financial services might not necessarily vary across different geographic areas. We control for the education level of individuals, by including dummy variables that indicate whether the individual has (i) no education or less than primary, (ii) primary completed, (iii) secondary level completed, and (iv) at least an undergraduate college degree. We expect individuals with higher levels of education to be more likely to use formal and informal financial services. We also include the Age of the individuals, in logs. While there might be a positive relationship between the age of individuals and the use of financial services, this relationship might be non-linear and turn negative at higher ages when individuals leave the labor market. We also include an income measure where available. With the exception of Kenya 2006 (no income), Kenya 2009 (individual expenses), and Uganda (household income), we include the log of individual monthly income.8 We expect higher-income individuals to be more likely to be able to afford formal financial services, while the relationship of income with the use of informal financial services is not clear a- 8 In the case of the pooled regression, we convert all income measures into USD, using average-year exchange rates. 21 priori. Finally, we include dummy variables indicating what the main income source of the individual is. Specifically, Employed, Self-employed and Agriculture are dummy variables indicating the employment status and sector, with the omitted category being dependent on pension or family. We also control for the ownership of a mobile phone, which might indicate stronger commercial needs and therefore demand for financial services. It might also indicate, however, openness to new technologies and therefore bank delivery channels. We control for the personal circumstances of the individual by including dummy variables for being married, whether the survey respondent is head of household and ­ where available ­ whether the respondent is the main earner and decision taker. All four factors might increase the probability of using financial services, be they formal or informal, as being married and/or being head of household imply stronger economic responsibilities. Where available, we also include a variable indicating risk aversion, which is a positive response to the question: "do you disagree that to get ahead on life one need to take some risks?". We also include an indicator of numeracy, which measures the extent to which the respondent can solve simple calculus problems. We again expect respondents that score higher on numeracy are more likely to use financial services. The results in Table 9 show that - on average ­ women are not significantly more or less likely to use formal financial services. We report first the pooled regression with survey dummies and then 11 survey-specific regressions. The pooled regression includes only variables that are available for all surveys. The regression in column 1 of Table 9 shows an insignificant coefficient on the female dummy. This is confirmed by considering the country- level regressions. Only in the 2008 survey for South Africa does the female dummy enter significantly and negatively, suggesting that females have a 12.2 percent lower probability of using formal financial services. Unlike in the univariate comparisons of Figure 2, we 22 therefore cannot find a gender gap in the use of formal banking services, once we control for other individual characteristics. The use of formal banking services is correlated with an array of other individual factors. Individuals with higher income are more likely to use formal banking services, as are users of mobile phones. Even controlling for the fact that users of mobile phones have, on average, higher incomes, they are, on average across our sample countries, 32.1 percent more likely to use formal banking services. Formally employed individuals are more likely to use formal banking services, while there is no consistent relationship across countries in the case of self-employed and individuals working in agriculture. Perhaps surprisingly, rural individuals are less likely to use formal banking services only in Kenya, Malawi, South Africa, and Tanzania (2006), while there does not seem to be any urban-rural gap in the other countries, with rural inhabitants even more likely to use formal banking services in Botswana. Education is a strong predictor of the use of banking services, with the use increasing linearly in most countries ­ with the notable exception of South Africa -, from individuals with primary education to individuals with secondary education to individuals with tertiary education. Older individuals are more likely to use formal banking services in almost all countries. We find that married individuals are more likely to use formal banking services in Botswana and Namibia. With respect to the household status of individuals, we find that being the household head does not increase the probability of using formal banking services, except in Uganda. Risk aversion is not significantly correlated with the use of banking services except for Botswana where it is negatively correlated, while numeracy is positively associated with the use of formal banking services in most but not all countries; these two variables, however, are not available for all countries. The survey dummies in the pooled regression provide evidence for the cross-country variation in the use of financial services beyond differences in population composition. We 23 find that relative to South Africa ­ the omitted country ­ all countries except for Namibia have lower levels of financial service use. The differences range from 41 percent in Zambia to 20 percent in Botswana. Individuals in Namibia are as likely to use formal banking services as individuals in South Africa. The results in Table 10 show that ­ on average across the 11 surveys ­ women are more likely to use informal financial services. The effect is also economically large, with women being 3.7 percent more likely to use informal financial services than men. Looking behind the average effect across countries, we find that this is driven by Botswana and the East African countries in our sample, with the effect being especially strong in Kenya (14.7 percent in 2006 and 16.3 percent in 2009). On the other hand, there is no gender difference in the use of informal financial services in Malawi, Namibia, South Africa, and Zambia. Many of the individual characteristics that explain the use of formal banking services also explain the use of informal financial services. The relationship between income and the use of informal financial services is positive in some but not all countries. In some countries owners of mobile phones are more likely to use informal financial services while in others it does not make a difference or they are less likely (Rwanda and Zambia). Maybe surprisingly, compared to individuals dependent on transfers, employed individuals are more likely to use informal finance in Botswana, Kenya, and Zambia as are self-employed in Kenya and Tanzania (2006), while individuals working in agriculture are only more likely to use informal financial services in Kenya. Rural people in Malawi, South Africa and Tanzania (2006) are more likely to use informal finance, while they are less likely to do so in Tanzania (2009). The relationship between education and the use of informal financial services is not consistent across countries. While in Botswana, Kenya, and Tanzania (2009), individuals with primary education are more likely to use informal financial services than individuals without any formal education, individuals with secondary or tertiary education are not more 24 likely to do so except for Botswana (and Uganda in the case of tertiary education). Older people are more likely to use informal finance only in Botswana, while they are less likely to do so in Rwanda and Tanzania (2009). Married individuals are more likely to use informal financial services in most, though not all, countries, with the exception of South Africa where they are less likely to do so. If individuals are the main earner in the household, they are more likely to use informal financial services in Tanzania (2009), but less likely in Malawi and Tanzania (2006). The main decision maker is more likely to use informal financial services only in Kenya (2009). There is no significant relationship between risk aversion and the likelihood of using informal financial services, while higher numeracy is positively associated with the use of informal financial services in Botswana, Malawi, and Tanzania (2006) and negatively in Uganda. The survey dummies in the pooled regression suggest that individuals in most countries are more likely to use to informal financial services than individuals in South Africa, while individuals in Namibia are less likely to do so and there is no significant difference in Zambia. Table 11 shows that, on average, women are less likely to be excluded from financial services, with significant cross-country variation behind the result from the pooled regression. Specifically, women in Botswana, Kenya, Tanzania, and Uganda, are less likely to be excluded from financial services (note that women in these same countries are more likely to use informal financial services), while women in Rwanda and South Africa are more likely to be excluded, although this relationship is significant only on the 10 percent level. There is no significant gender gap in financial exclusion in the other countries. The other individual characteristics that are significantly correlated with the use of formal and informal financial services are also significantly correlated with the likelihood of being excluded, though with the opposite sign. Higher income individuals are less likely to be excluded, and controlling for this income effect, owners of mobile phones are less likely to be 25 excluded. Self-employed are less likely to be excluded in Kenya, Namibia, Rwanda and Tanzania (2006), while agricultural workers are more likely to be excluded in Botswana and less likely in Kenya and Malawi. Individuals living in rural areas face a lower probability of exclusion in Malawi, while a higher probability in Tanzania (2009), with no geographic gap in other countries. Education is an important predictor of the likelihood of not being excluded, with the relationship between the likelihood of exclusion and educational attainment decreasing in a linear measure in all countries, except for South Africa. Older individuals are less likely to be excluded in many though not all countries, as are formally employed individuals. Married individuals are less likely to be excluded in most countries, while heads of household are no more or less likely to be excluded, with the exception of Kenya (2006) and Uganda, where the relationship is negative, and Malawi where the relationship is positive. Numeracy is negatively associated with the likelihood of exclusion in many though not all countries, while there seems to be an inconsistent relationship with risk aversion, negative in South Africa and positive in Zambia. The survey dummies in the pooled regression suggest that with the exception of Kenya (2006), where there is no significant difference, the probability of exclusion is higher in all countries compared to South Africa. Table 12 explores why the significant variation in the use of formal financial services between men and women, shown in Figure 2, turns insignificant once we control for other individual characteristics. Here we present the differences in individual characteristics between men and women for each survey and the economic effect this has on the different use of formal banking services by men and women. Specifically, for the pooled regression and each individual survey regression of Table 9, we multiply the coefficient estimate on each explanatory variable with the sample difference between male and female individuals. The biggest effect seems to come from income differences between male and female 26 individuals, explaining between 0.9 percentage points in Uganda and 4.5 percentage points in Tanzania (2006). Another big effect stems from the lower level of education of women, which can explain why women are less likely to use formal banking services. Adding the economic effect of lower education across primary, secondary and tertiary education levels in the pooled regression, we find a total effect of 4.5 percentage points. Another large effect comes from women being less likely to be the household head, accounting for a 3.2 percentage point difference in the pooled regression. The employment status is another important factor explaining gender differences in the use of formal banking services. Women are less likely to be formally employed than men, with the economic effect being 3.1 percentage points in the pooled regression. Finally, the ownership of mobile phones seems a significant factor in explaining the univariate gender gap in use of formal banking services, adding another two percentage points. Overall, the results in Table 12 suggest that the fact that women have lower income, are less formally educated, are less likely to be head of household and are less likely to be formally employed across the countries in our sample explains why they are less likely to use formal banking services. This suggests that it is not discrimination in the banking system or lower inherent demand by women that drives their lower use of banking services, but rather disadvantages in other areas. However, these results also suggest that some of the findings might be driven by the survey methodology of interviewing individuals rather than households; women might have indirect access to formal financial services through their formally employed husbands who function as household heads. In sensitivity tests not shown9, we also explored whether the relationship between education, income, household head and married, on the one hand, and the use of financial services, on the other hand, varied between men and women. Few of the interaction terms, 9 Results available upon request 27 however, enter significantly, and mostly with differing signs across surveys. Overall, there seems little evidence that education and marital status are differently related with use of financial services across genders. We also differentiated according to the financial service individuals are or are not using ­ credit, savings, insurance and transaction services. Here we follow the definition by Porteous (2007) that captures both formal and semi-formal financial service providers. We find that females in Malawi and Tanzania (2006) are more likely to use transaction services, while there is no significant difference at the 5 percent level in other countries. Females in Botswana, Kenya and Zambia are more likely to use savings services, while they are less likely to use them in Rwanda. There are no significant gender differences in credit and insurance services10. 4. Conclusions This paper assesses gender differences in the use of financial services by enterprises and households in Sub-Saharan Africa. We find little if any evidence for a gender gap, either for enterprises or households. Enterprises with female ownership participation in Sub- Saharan Africa use as much external financing as enterprises without female ownership participation and female individuals are as likely to use formal financial services as male individuals, once we control for an array of other characteristics. While this might seem surprising, our results suggest that one has to look beyond simple gender comparisons and explore the reasons why we find a lower financial market participation of women. In the case of enterprises, there is evidence of selection bias, i.e., female entrepreneurs have to overcome higher barriers in the first place, as evidenced by their higher tendency to innovate and higher legal burden in African countries compared to their male peers. Further, firms with female ownership participation tend to be of smaller size, and smaller firms have, on average, less 10 Results available upon request. 28 access to external finance. Additionally, we find some limited evidence that females owned business tend to enter more likely sectors that, on average, are characterized by more limited use of external finance (i.e. average number of companies with bank accounts). In the case of individuals, univariate comparisons show a lower formal financial sector participation rate of females as they score lower on many other dimensions related to the use of financial services, including income, education and formal employment, but also personal life factors such as not being head of household. These barriers that women face as individuals to access formal financial services might also explain the selection bias among female entrepreneurs that we found in the first part of the paper. Are women disadvantaged in access to financial services? Yes, but the reasons are not within the financial sector, they lie in other dimensions related to female participation in the modern market economy. Policies to expand access to financial services by women have to address these other dimensions if women are to reap the benefit of financial services as much as men. 29 References Aghion, Philippe, and Patrick Bolton. 1997. A Theory of Trickle-Down Growth and Development. The Review of Economic Studies 64, 151-72. Ai, Chunrong and Edward Norton. 2003. Interaction terms in logit and probit models. Economics Letters 80,123-129 Banerjee, Abhijit and Esther Duflo. 2008. Do Firms Want to Borrow More? Testing Credit Constraints Using a Directed Lending Program. MIT mimeo. Beck, Thorsten, Patrick Behr and Andreas Madestam. 2011. Sex and Credit: Is There a Gender Bias in Microfinance? Mimeo. Beck, Thorsten, Asl Demirgüç-Kunt and Ross Levine. 2007. Finance, Inequality, and the Poor. Journal of Economic Growth 12, 27-49. Beck, Thorsten, Asli Demirgüç-Kunt, and Vojislav Maksimovic. 2005. Financial and Legal Constraints to Firm Growth: Does Firm Size Matter? Journal of Finance 60, 137-177. Beck, Thorsten, Asli Demirgüç-Kunt, and Vojislav Maksimovic. 2008. Financing Patterns Around the World: Are Small Firms Different? Journal of Financial Economics 89,467-87. Bruhn, Miruam. 2009. Female-Owned Firms in Latin America: characteristics, performance, and obstacles to growth. World Bank Policy Research Working Paper 5122. Brush, Candida, Nancy Carter, Elizabeth Gatewood, Patricia Greene, Myra Hart, 2004. Gatekeepers of venture growth: The role and participation of women in the venture capital industry, mimeo Buvinic, Mayra and Marguerite Berger, 1990. Sex differences in access to a small enterprise development fund in Peru, World Development 18, 695-705. Cull, Robert and Kinnon Scott. 2010. Measuring Household Usage of Financial Services: Does it Matter How and Whom You Ask? World Bank Economic Review 24, 199-233. De Mel, Suresh, David McKenzie and Chris Woodruff, 2008. Returns to Capital in Microenterprises: Evidence from a Field Experiment, Quarterly Journal of Economics 123, 1329-72. Dupas, Pascaline and Jonathan Robinson. 2009. Savings Constraints and Microenterprise Development: Evidence from a Field Experiment in Kenya. NBER Working Paper 14693. Gajigo, Ousman and Mary Hallward-Driemeier. 2010. Constraints and Opportunities for New Enterpreneurs in Africa. World Bank Working Paper. Galor, Oded and Omer Moav. 2004. From Physical to Human Capital Accumulation: Inequality and the Process of Development. Review of Economic Studies 71, 1001-26. 30 Galor, Oded, and Joseph Zeira. 1993. Income Distribution and Macroeconomics. The Review of Economic Studies 60, 35-52. Hallward-Driemeier, Mary (ed.) 2011a. "Expanding Opportunities for Women Entrepreneurs in Africa." World Bank. Hallward-Driemeier, Mary (ed.) 2011. "Improving the Legal Investment Climate for Women in Africa". World Bank. Hallward-Driemeier, Mary and Reyes Aterido. 2010. "Whose Business is it Anyway". World Bank Working Paper. Honohan , Patrick and Thorsten Beck. 2007. Making Finance Work for Africa, World Bank, Washington, D.C. Honohan, Patrick. 2004. Financial Sector Policy and the Poor. Working Paper No. 43, World Bank, Washington, D.C. Johnson, Susan, 2004. Gender norms in financial markets: Evidence from Kenya. World Development 32, 1355-1374. Klapper, Leora, Luc Laeven, Raghuram Rajan, 2006. Entry regulation as a barrier to entrepreneurship. Journal of Financial Economics, 82, 591-629. Klapper, Leora and Simon Parker. 2010. Gender and the Business Environment for New Firm Creation, World Bank Research Observer. Lopez, Humberto and Luis Serven. 2009. Too Poor to Grow. World Bank Policy Research Working Paper 5012. Lusardi, Annamaria and Peter Tufano, 2009. Debt literacy, financial experiences, and overindebtedness. NBER Working Paper 14808. Muravyev, Alexander, Dorothea Schäfer and Oleksandr Talavera, 2007. Entrepreneurs' gender and financial constraints: Evidence from international data. Discussion Papers of DIW Berlin 706, German Institute for Economic Research. ` Naidoo, Sharda and Anne Hilton, 2006. Access to finance for women entrepreneurs in South Africa. World Bank, Washington, DC. Porteous, David 2007. Finance Service Access and Usage in Southern and East Africa: What do Finscope Surveys Tell Us? Finmark Trust South Africa. Richardson, Pat, Rhona Howarth and Gerry Finnegan, 2004. The challenges of growing small businesses: Insights from women entrepreneurs in Africa. ILO Geneva, SEED WP No. 47. World Bank. 2007. Finance for All? Policies and Pitfalls in Expanding Access. Washington, D.C. World Bank. 2010. "Women, Business and the Law". Washington, D.C. 31 Table 1. Panel A. Summary Statistics Enterprise Surveys Variable Variable description n. obs. share st.dev. min max Female At least 1 principal owner is female 35,135 0.36 0.48 0 1 Africa Firm is in an Africa country 35,135 0.30 0.46 0 1 Micro 1-5 employees 35,135 0.13 0.33 0 1 Small 6-10 employees 35,135 0.23 0.42 0 1 Medium 11-49 employees 35,135 0.38 0.49 0 1 Large 50 or more employees 35,135 0.27 0.44 0 1 public Government ownership 35,135 0.05 0.22 0 1 private, limited Private, limited ownership 35,135 0.55 0.50 0 1 sole proprietor Sole proprietor 35,135 0.30 0.46 0 1 partnership Partnership ownership 35,135 0.07 0.26 0 1 other Other ownership 35,135 0.02 0.14 0 1 Fin_Formal 1 if overdraft or checking/saving account or creditline or loan 35,135 0.55 0.50 0 1 Fin_inv % New investments paid w/ (priv or gvmt) banks or fin.inst. 18,807 17.54 33.04 0 100 Fin_wkcap % Working capital paid w/ (priv or gvmt) banks or fin.inst. 24,791 9.18 21.36 0 100 logEmployment Number of permanent workers -log 35,135 3.13 1.38 0 9.94 Innov_prod Firm improved products in last 3yrs 9,375 0.61 0.49 0 1 Innov_proc Firm improved production process in last 3yrs 9,385 0.57 0.49 0 1 lgCity Firm is in the capital or city w/ population of 1Mn or more 35,135 0.40 0.49 0 1 Age Age of firm -log 34,827 2.46 0.85 0 5.74 Exporter Firm exports directly at least 10% of total sales 35,086 0.13 0.34 0 1 Foreign at least 10% of firm is owned by foreign private sector 35,032 0.10 0.31 0 1 Index Women- 79 0.71 0.26 0.13 1 Business&Law1 1=equality; 0=no equality; 1 average: property, inheritance, law, work all industries; same night hrs; married women; tax liability; pregnant, nursing women Table 1.Panel B. Distibution of Companies with some Female Ownership All Countries Africa Male Female Total Male Female Total Sole Proprietors n. obs 7,714 2,861 10,575 4,648 1,447 6,095 % 72.95 27.05 100 76 24 100 Formal n. obs 22,602 12,533 35,135 7,724 2,984 10,708 % 64.33 35.67 100 72 28 100 Informal n. obs 1,388 678 2,066 % 67.18 32.82 100 Total n. obs 22,602 12,533 35,135 9,112 3,662 12,774 % 64.33 35.67 100 71.33 28.67 100 32 Table 2: Explaining Share of Investment and Working Capital Financed by External Financial Institution (Formal Companies) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Access to formal Finance Share Investment Financed Externally Share Work. capital Financed Externally Formal Sole Proprietor Formal Sole Proprietor Formal Sole Proprietor Female -0.026 -0.028 -0.091 -0.086 -0.928 -0.706 -1.373 -1.502 1.833 1.653 5.982 6.082 [0.033] [0.032] [0.055] [0.053] [1.959] [1.827] [4.562] [4.437] [1.232] [1.198] [3.342]* [3.309]* Africa -0.464 -0.456 -0.306 -0.293 -13.183 -14.533 -13.950 -13.636 -7.342 -7.572 -6.942 -6.467 [0.054]*** [0.054]*** [0.051]*** [0.051]*** [3.293]*** [3.148]*** [3.607]*** [3.588]*** [2.882]** [2.611]*** [2.352]*** [2.339]*** Employment 0.065 0.060 0.096 0.096 1.705 1.796 1.282 1.303 1.882 1.774 1.239 1.432 [0.007]*** [0.006]*** [0.020]*** [0.018]*** [0.357]*** [0.403]*** [1.104] [1.090] [0.471]*** [0.470]*** [1.210] [1.183] Female*Africa 0.074 0.076 0.066 0.067 4.761 4.777 -0.104 0.823 -0.407 -0.283 -4.638 -4.563 [0.045] [0.043]* [0.061] [0.060] [2.562]* [2.424]* [5.572] [5.390] [1.594] [1.524] [3.500] [3.503] Female*Employment 0.013 0.012 0.032 0.029 0.299 0.252 -0.142 -0.125 -0.404 -0.372 -2.429 -2.506 [0.005]** [0.005]** [0.016]** [0.016]* [0.445] [0.413] [1.542] [1.502] [0.436] [0.399] [1.337]* [1.308]* Afr*Employment 0.067 0.061 -0.015 -0.021 1.155 1.033 -0.021 -0.425 -0.383 -0.628 -0.374 -0.757 [0.017]*** [0.016]*** [0.027] [0.024] [1.432] [1.496] [1.255] [1.205] [0.861] [0.844] [1.190] [1.163] Female*Africa*Employment -0.009 -0.009 -0.010 -0.010 -0.508 -0.653 1.198 0.810 0.084 0.005 1.888 1.849 [0.015] [0.015] [0.020] [0.021] [0.926] [0.912] [2.188] [2.128] [0.602] [0.565] [1.414] [1.405] private-limited 0.093 0.095 6.096 5.555 1.701 1.698 [0.020]*** [0.021]*** [1.944]*** [1.778]*** [1.294] [1.174] sole-proprietors -0.029 -0.026 3.032 2.864 -1.244 -0.974 [0.029] [0.029] [1.820] [1.706]* [1.455] [1.348] partnership 0.020 0.024 4.979 4.605 -0.280 -0.056 [0.023] [0.021] [2.394]** [2.095]** [1.442] [1.335] legal-status_other 0.008 0.008 9.454 9.495 5.265 5.155 [0.036] [0.035] [3.706]** [3.690]** [1.772]*** [1.847]*** age 0.024 0.020 -0.118 1.167 0.538 0.398 [0.007]*** [0.007]*** [0.384] [0.572]** [0.305]* [0.316] lgcity 0.056 0.043 8.348 4.227 5.349 2.458 [0.027]** [0.026] [2.511]*** [1.908]** [1.466]*** [1.195]** exporter 0.070 0.031 2.623 -0.330 2.761 -0.606 [0.029]** [0.029] [1.649] [1.632] [1.359]** [1.217] foreign -0.036 -0.047 -4.026 1.804 -2.362 -1.097 [0.020]* [0.049] [1.332]*** [2.736] [0.874]*** [1.190] industry fe yes yes yes yes yes yes yes yes yes yes yes yes country controls 1 yes yes yes yes yes yes yes yes yes yes yes yes Constant 0.452 0.361 0.340 0.263 11.665 6.414 16.705 12.315 16.682 11.994 13.848 11.512 [0.085]*** [0.089]*** [0.066]*** [0.075]*** [4.153]*** [3.703]* [4.416]*** [4.252]*** [3.926]*** [3.579]*** [3.200]*** [3.188]*** Observations 35135 34684 10575 10505 18807 18604 4469 4446 24791 24615 9242 9203 R-squared 0.24 0.26 0.25 0.26 0.05 0.06 0.08 0.09 0.09 0.11 0.08 0.09 * significant at 10%; ** significant at 5%; *** significant at 1% Notes: cols 1-4 dprobit; cols 5-12 OLS; omitted category public; se clustered at the country level 1 controls include GDP per capita, GDP growth previous 3 years and inflation 33 Table 3: Explaining Access to Formal Financing (Africa Only) (1) (2) (3) (4) (5) (6) (7) (8) Formal Firms Informal All Formal Sole Proprietor All Informal Sole Proprietor Female 0.034 0.032 -0.028 -0.024 0.041 0.007 0.030 -0.003 [0.030] [0.028] [0.022] [0.023] [0.018]** [0.016] [0.017] [0.026] Employment 0.110 0.095 0.073 0.067 [0.015]*** [0.015]*** [0.017]*** [0.015]*** Fem*Employment 0.005 0.005 0.022 0.020 [0.012] [0.011] [0.010]** [0.011]* private-limited 0.031 0.035 0.192 [0.041] [0.044] [0.038]*** sole-proprietors -0.109 -0.102 0.153 -0.112 [0.054]* [0.055]* [0.030]*** [0.066] partnership -0.039 -0.030 0.214 -0.028 [0.045] [0.043] [0.028]*** [0.058] legal-status_other -0.026 -0.020 [0.113] [0.108] age 0.028 0.021 0.036 0.030 [0.004]*** [0.006]*** [0.007]*** [0.012]* lgcity 0.017 0.014 -0.073 -0.058 [0.023] [0.023] [0.027]* [0.022]* exporter 0.129 0.069 0.166 -0.066 [0.023]*** [0.032]** [0.087] [0.067] foreign -0.030 0.002 -0.105 -0.030 [0.026] [0.023] [0.056] [0.057] industry fe yes yes yes yes yes yes yes yes country controls 1 yes yes yes yes yes yes yes yes Constant 0.061 -0.008 0.052 0.006 -0.011 0.073 0.109 0.085 [0.103] [0.096] [0.116] [0.119] [0.088] [0.189] [0.083] [0.148] Observations 10708 10670 6095 6081 1962 604 1575 483 R-squared 0.25 0.26 0.11 0.11 0.03 0.07 0.02 0.05 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Notes: se clustered at the country level 34 Table 4: Explaining Share of Investment Financed by External Financial Institutions (Africa Only) (1) (2) (3) (4) (5) (6) (7) (8) Formal Informal All formal Sole Proprietor All Informal Sole Proprietor Female 3.045 2.785 -0.866 -0.665 -0.555 -2.806 -1.345 -2.040 [1.742]* [1.699] [3.328] [3.189] [1.852] [2.420] [1.448] [1.455] Employment 2.347 2.022 1.373 1.319 [0.867]** [0.844]** [0.566]** [0.578]** Fem*Employment -0.187 -0.087 0.896 0.789 [0.818] [0.802] [1.581] [1.514] private-limited 1.238 1.214 14.987 5.173 [3.954] [4.125] [8.245]* [5.477] sole-proprietors -4.637 -4.338 5.730 -5.861 [4.320] [4.462] [4.476] [2.833] partnership -2.223 -2.061 9.727 [3.651] [3.818] [4.463]** legal-status_other -7.730 -7.020 [4.666] [4.794] age -0.363 0.010 2.809 1.824 [0.534] [0.620] [1.198]* [0.927] lgcity -0.085 1.829 -10.564 -0.241 [1.969] [1.596] [3.233]** [1.975] exporter 4.147 -2.281 -3.249 -0.486 [2.799] [1.858] [2.650] [0.817] foreign 1.225 3.706 3.280 -1.952 [2.201] [2.970] [5.958] [1.587] industry fe yes yes yes yes yes yes yes yes country controls 1 yes yes yes yes yes yes yes yes Constant 9.715 9.758 7.128 6.276 -7.171 32.481 1.175 8.959 [9.021] [9.237] [8.752] [8.516] [4.532] [11.654]** [0.845] [3.714]* Observations 4749 4733 2426 2421 604 211 484 173 R-squared 0.07 0.07 0.03 0.04 0.04 0.23 0.04 0.09 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Notes: se clustered at the country level 1 controls include GDP per capita, GDP growth previous 3 years and inflation 35 Table 5: Explaining Share of Working Capital Financed by External Financial Institutions (Africa Only) (1) (2) (3) (4) (5) (6) (7) (8) Formal Informal All formal Sole Proprietor All Informal Sole Proprietor Female 1.503 1.423 1.221 1.272 1.185 0.800 0.560 -0.115 [0.976] [0.947] [0.884] [0.921] [0.982] [1.451] [0.854] [0.383] Employment 1.442 1.267 1.049 1.023 [0.510]*** [0.467]** [0.339]*** [0.318]*** Fem*Employment -0.295 -0.281 -0.450 -0.480 [0.389] [0.385] [0.417] [0.440] private-limited -0.075 -0.165 1.328 [1.932] [2.032] [0.902] sole-proprietors -3.015 -3.090 1.684 -1.088 [2.070] [2.174] [0.621]** [1.093] partnership -2.662 -2.685 4.591 4.670 [1.991] [2.071] [1.519]*** [2.296] legal-status_other -2.692 -2.693 [2.973] [3.011] age 0.065 0.125 0.932 0.467 [0.213] [0.163] [0.402]* [0.170]* lgcity 0.812 0.396 -3.930 -3.114 [1.068] [0.741] [1.723]* [2.263] exporter 2.520 -0.162 0.136 0.108 [1.044]** [0.663] [0.444] [0.886] foreign -0.669 -0.461 -0.597 0.113 [0.928] [0.575] [0.310] [0.270] industry fe yes yes yes yes yes yes yes yes country controls 1 yes yes yes yes yes yes yes yes Constant 6.799 6.373 2.612 2.270 -1.071 4.064 0.977 4.160 [3.018]** [3.271]* [2.598] [2.589] [2.228] [3.722] [1.391] [1.871]* Observations 10696 10658 6091 6077 1953 604 1570 483 R-squared 0.06 0.06 0.03 0.03 0.02 0.24 0.02 0.14 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Notes: se clustered at the country level 1 controls include GDP per capita, GDP growth previous 3 years and inflation 36 Table 6. Distribution Legal Ownership status Male Female Total public obs 969 887 1,856 % 4.29 7.08 5.28 private, limited obs 11,994 7,469 19,463 % 53.07 59.59 55.39 sole proprietor obs 7,714 2,861 10,575 % 34.13 22.83 30.10 partnership obs 1,516 1,022 2,538 % 6.71 8.15 7.22 other obs 409 294 703 % 1.81 2.35 2.00 Total obs 22,602 12,533 35,135 % 100 100 100 37 Table 7. Panel A: Characteristics of Formal Enterprises (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Dep var Log-Employment Product Innovation Process Innovation All Sole Proprietor All Sole Proprietor All Sole Proprietor Female -0.083 -0.078 -0.110 -0.139 0.060 0.058 0.070 0.080 0.036 0.034 0.046 0.050 [0.031]*** [0.038]** [0.048]** [0.059]** [0.019]*** [0.018]*** [0.031]** [0.028]*** [0.015]** [0.016]** [0.027]* [0.025]* Africa -0.146 -0.198 -0.270 -0.371 -0.108 -0.114 -0.095 -0.088 -0.192 -0.198 -0.203 -0.206 [0.096] [0.109]* [0.115]** [0.133]*** [0.038]*** [0.042]** [0.056] [0.061] [0.036]*** [0.038]*** [0.051]*** [0.051]*** Female*Africa -0.020 -0.050 -0.007 0.008 -0.006 -0.005 -0.088 -0.094 0.015 0.017 -0.081 -0.082 [0.049] [0.053] [0.075] [0.085] [0.032] [0.030] [0.041]** [0.042]** [0.029] [0.028] [0.038]** [0.037]** age 0.324 0.157 0.022 0.001 0.016 0.003 [0.022]*** [0.017]*** [0.008]** [0.012] [0.007]** [0.015] lgcity -0.004 0.090 0.029 0.070 0.019 0.049 [0.057] [0.054] [0.021] [0.031]** [0.021] [0.031] exporter 0.791 1.040 0.133 0.224 0.133 0.139 [0.052]*** [0.106]*** [0.035]*** [0.058]*** [0.029]*** [0.039]*** foreign 0.577 0.597 0.021 0.067 0.021 0.001 [0.055]*** [0.127]*** [0.022] [0.077] [0.020] [0.094] private-limited -0.737 -0.940 0.003 -0.019 -0.048 -0.070 [0.083]*** [0.093]*** [0.033] [0.036] [0.023]** [0.025]*** sole-proprietors -1.519 -1.923 -0.130 -0.188 -0.175 -0.230 [0.100]*** [0.100]*** [0.042]*** [0.053]*** [0.029]*** [0.037]*** partnership -1.031 -1.330 -0.077 -0.125 -0.155 -0.203 [0.099]*** [0.104]*** [0.065] [0.074] [0.041]*** [0.047]*** legal-status_other -0.827 -1.011 -0.005 -0.022 -0.086 -0.101 [0.126]*** [0.144]*** [0.057] [0.061] [0.035]** [0.038]** industry fe yes yes yes yes yes yes yes yes yes yes yes yes country controls 1 yes yes yes yes yes yes yes yes yes yes yes yes Constant 3.244 4.561 2.000 2.578 0.501 0.644 0.522 0.584 0.579 0.696 0.578 0.628 [0.144]*** [0.144]*** [0.173]*** [0.170]*** [0.062]*** [0.058]*** [0.090]*** [0.099]*** [0.070]*** [0.053]*** [0.054]*** [0.059]*** Observations 34684 35135 10505 10575 9294 9392 2964 2984 9303 9402 2966 2986 R-squared 0.35 0.25 0.22 0.13 0.06 0.05 0.05 0.03 0.07 0.06 0.05 0.05 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% omitted category public; se clustered at the country level; 1 controls include GDP per capita, GDP growth previous 3 years and inflation 38 Table 7. Panel B: Industry Specific Access to Finance of Female Formal Enterprises Probit (1) (2) (3) (4) (5) (6) Dep. Var.: Female All formal Sole Proprietor Fin_Formal_s -0.634 -1.491 [0.180]*** [0.318]*** Fin_inv_s -0.006 -0.013 [0.004] [0.009] Fin_wkcap_s 0.002 -0.004 [0.005] [0.014] country fe yes yes yes yes yes yes Constant -0.764 -0.992 -1.100 -0.517 -1.038 -1.197 [0.090]*** [0.059]*** [0.036]*** [0.148]*** [0.118]*** [0.083]*** Observations 36877 36877 36877 11363 11363 11363 Pseudo-Rsq 0.04 0.04 0.04 0.05 0.05 0.05 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% se clustered at the country level 39 Table 8. Summary Statistics Household Surveys Variable Variable description n. obs. mean st.dev. min max Banking uses now banking services 43908 0.23 0.42 0 1 Informal uses now unregistered financial services 43908 0.27 0.45 0 1 Excluded not banked; not formal or informal financial institutions 43908 0.57 0.49 0 1 Female 1 if respondent female 43908 0.53 0.50 0 1 Married 1 if married 43905 0.59 0.49 0 1 HH_head 1 if household head 43897 0.46 0.50 0 1 No-education less than primary 43240 0.29 0.45 0 1 Primary primary complete (and) less than secondary complete 43240 0.47 0.50 0 1 Secondary secondary or vocational training complete (and) less than tertiary complete 43240 0.20 0.40 0 1 Tertiary tertiary complete or more 43240 0.03 0.16 0 1 Numeracy does not know what inflation is 35590 0.41 0.49 0 1 Owns_mobile owns/uses pre-paid or contract cell phone 43883 0.33 0.47 0 1 Age_log age (years) -log 43736 3.51 0.40 2.77 4.65 Employed (main) source of income is from a wage (company or individual) 43908 0.18 0.39 0 1 Self_employed (main) source of income is from own business 43908 0.17 0.38 0 1 Agriculture 40008 (main) source of income is from selling agricultural, livestock or fishing products 0.40 0.49 0 1 Riskaverse disagree that 'to get ahead on life one need to take some risks' 13257 0.24 0.43 0 1 Rural lives in a rural area 43908 0.65 0.48 0 1 Earner 1 if household main earner 25962 0.69 0.46 0 1 Decision_mkr makes financial decisions (self or with spouse) 25859 0.74 0.44 0 1 log_income log individual monthly income -LCU (ALL in USD) 31151 2.93 3.47 -9.10 9.37 region number identifying different regions 43908 8.0 11.1 1 55 weight weights 42708 3189.2 4806.8 0.026 122826 40 Table 9. Use of Banking Services (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) ALL Botswana Kenya06 Kenya09 Malawi Namibia Rwanda SouthAfrica Tanzania06 Tanzania09 Uganda Zambia Female 0.017 0.029 0.039 -0.015 0.026 -0.015 -0.002 -0.122 -0.004 -0.003 -0.018 -0.001 [0.018] [0.041] [0.042] [0.030] [0.018] [0.040] [0.022] [0.051]** [0.010] [0.008] [0.025] [0.003] Rural 0.001 0.174 -0.142 -0.069 -0.057 -0.045 -0.016 -0.066 -0.024 -0.003 -0.004 0.004 [0.020] [0.084]** [0.038]*** [0.029]** [0.022]*** [0.049] [0.020] [0.038]* [0.013]* [0.009] [0.029] [0.003] Primary 0.186 0.250 0.020 0.152 -0.013 0.095 0.112 0.205 0.043 0.025 0.086 0.010 [0.026]*** [0.061]*** [0.036] [0.034]*** [0.018] [0.068] [0.021]*** [0.059]*** [0.025]* [0.010]** [0.028]*** [0.007] Secondary 0.437 0.370 0.333 0.338 0.144 0.232 0.421 0.312 0.105 0.174 0.308 0.118 [0.026]*** [0.067]*** [0.046]*** [0.041]*** [0.032]*** [0.071]*** [0.098]*** [0.055]*** [0.026]*** [0.037]*** [0.066]*** [0.052]** Tertiary 0.655 0.521 0.704 0.563 0.614 0.846 0.257 0.641 0.682 0.399 0.572 [0.022]*** [0.079]*** [0.060]*** [0.069]*** [0.102]*** [0.064]*** [0.027]*** [0.206]*** [0.102]*** [0.171]** [0.207]*** Age 0.256 0.198 0.226 0.200 0.051 0.105 0.072 0.255 0.024 0.044 -0.017 0.021 [0.030]*** [0.083]** [0.046]*** [0.040]*** [0.022]** [0.070] [0.025]*** [0.075]*** [0.012]** [0.011]*** [0.038] [0.011]* Income 0.077 0.146 0.036 0.046 0.033 0.026 0.031 0.010 0.036 0.004 [0.010]*** [0.017]*** [0.005]*** [0.009]*** [0.014]** [0.008]*** [0.006]*** [0.003]*** [0.008]*** [0.001]*** Employed 0.404 0.115 0.185 0.130 0.134 0.374 0.125 0.153 0.024 0.095 0.010 0.052 [0.019]*** [0.055]** [0.046]*** [0.038]*** [0.034]*** [0.049]*** [0.076]* [0.054]*** [0.016] [0.035]*** [0.044] [0.026]** Self_employed 0.227 -0.018 0.120 0.074 -0.005 0.189 0.191 0.042 -0.009 -0.013 -0.051 0.008 [0.027]*** [0.070] [0.057]** [0.039]* [0.027] [0.061]*** [0.081]** [0.087] [0.011] [0.010] [0.034] [0.007] Agriculture -0.202 0.185 0.068 0.051 -0.029 0.038 -0.021 -0.045 -0.092 0.001 [0.075]*** [0.079]** [0.037]* [0.022]** [0.097] [0.033] [0.012]* [0.013]*** [0.035]*** [0.004] Owns_mobile 0.321 0.198 0.308 0.297 0.183 0.306 0.085 0.110 0.041 0.123 0.293 0.026 [0.021]*** [0.044]*** [0.039]*** [0.024]*** [0.020]*** [0.046]*** [0.053] [0.056]** [0.016]** [0.018]*** [0.033]*** [0.015]* Married 0.099 0.125 0.031 -0.035 -0.018 0.149 0.021 -0.032 0.005 0.009 -0.006 0.003 [0.020]*** [0.063]** [0.037] [0.030] [0.020] [0.051]*** [0.021] [0.051] [0.010] [0.008] [0.023] [0.003] HH_head 0.081 -0.013 -0.014 0.026 -0.005 0.040 0.005 -0.027 -0.004 0.002 0.042 -0.000 [0.021]*** [0.061] [0.036] [0.036] [0.020] [0.050] [0.022] [0.058] [0.010] [0.009] [0.025]* [0.005] Earner 0.113 0.014 0.030 -0.916 -0.022 -0.001 [0.058]* [0.031] [0.046] [0.097]*** [0.021] [0.005] Decision_taker 0.112 0.027 0.053 0.011 -0.004 [0.033]*** [0.019] [0.027]* [0.010] [0.007] Riskaverse -0.087 0.027 0.017 0.003 -0.004 [0.045]* [0.043] [0.047] [0.027] [0.003] Numeracy 0.061 0.101 0.141 0.128 0.032 0.017 0.020 0.121 0.006 [0.045] [0.026]*** [0.037]*** [0.059]** [0.021] [0.010]* [0.008]** [0.036]*** [0.004] Botswana -0.196 [0.041]*** Kenya06 -0.290 [0.015]*** Kenya09 -0.326 [0.022]*** Malawi -0.306 [0.012]*** Namibia -0.009 [0.042] Rwanda -0.257 [0.020]*** Tanzania06 -0.304 [0.011]*** Tanzania09 -0.302 [0.011]*** Uganda -0.292 [0.014]*** Zambia -0.410 [0.020]*** regional f.e. yes yes yes yes yes yes yes yes yes yes yes yes Observations 43044 974 4237 6589 4712 1065 1894 3345 3097 5864 1587 3152 pseudo-r2 0.42 0.37 0.38 0.46 0.36 0.47 0.33 0.27 0.31 0.41 0.41 0.58 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Uganda06: household income; Kenya09: expenses; pooled regresions log-income (USD) 41 Table 10. Use of Informal Services (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) ALL Botswana Kenya06 Kenya09 Malawi Namibia Rwanda SouthAfrica Tanzania06 Tanzania09 Uganda Zambia Female 0.037 0.115 0.147 0.163 -0.012 -0.000 -0.084 0.012 0.127 0.063 0.058 0.003 [0.006]*** [0.034]*** [0.041]*** [0.029]*** [0.018] [0.000] [0.038]** [0.012] [0.026]*** [0.019]*** [0.027]** [0.002] Rural 0.009 -0.072 0.010 0.002 0.191 0.000 0.008 0.043 0.079 -0.051 0.018 0.001 [0.007] [0.065] [0.037] [0.031] [0.017]*** [0.001] [0.039] [0.015]*** [0.036]** [0.025]** [0.036] [0.003] Primary 0.016 0.183 0.098 0.075 -0.009 -0.001 -0.034 -0.011 -0.009 0.074 0.029 0.004 [0.008]** [0.051]*** [0.042]** [0.030]** [0.017] [0.002] [0.030] [0.012] [0.037] [0.021]*** [0.027] [0.003] Secondary 0.015 0.178 -0.135 -0.040 -0.048 -0.001 -0.156 -0.035 -0.015 -0.004 -0.005 -0.004 [0.009]* [0.063]*** [0.051]*** [0.038] [0.025]* [0.001] [0.047]*** [0.013]*** [0.039] [0.037] [0.048] [0.003] Tertiary 0.043 0.267 -0.063 -0.221 -0.139 -0.000 -0.031 -0.045 -0.193 0.296 -0.004 [0.023]* [0.105]** [0.093] [0.066]*** [0.037]*** [0.000] [0.005]*** [0.106] [0.029]*** [0.151]** [0.004] Age 0.045 0.354 -0.059 -0.044 0.007 -0.000 -0.083 0.005 -0.042 -0.071 -0.000 -0.000 [0.009]*** [0.063]*** [0.048] [0.035] [0.021] [0.001] [0.043]* [0.012] [0.032] [0.026]*** [0.034] [0.004] Income 0.024 0.031 0.006 0.000 0.006 -0.000 0.037 0.006 0.011 0.001 [0.009]*** [0.013]** [0.003]** [0.000] [0.010] [0.001] [0.011]*** [0.004]* [0.008] [0.000]** Owns_mobile 0.033 0.143 0.009 0.074 -0.023 0.000 -0.112 -0.017 0.056 -0.004 0.056 -0.006 [0.008]*** [0.038]*** [0.046] [0.027]*** [0.018] [0.000] [0.043]*** [0.010] [0.033]* [0.022] [0.029]* [0.003]** Employed 0.097 0.236 0.183 0.142 -0.021 -0.000 -0.073 0.016 0.068 -0.097 -0.089 0.030 [0.011]*** [0.050]*** [0.052]*** [0.033]*** [0.027] [0.000] [0.061] [0.015] [0.041]* [0.044]** [0.049]* [0.014]** Self_employed 0.050 0.053 0.216 0.136 -0.007 0.046 -0.016 0.152 0.054 -0.088 0.008 [0.012]*** [0.066] [0.051]*** [0.034]*** [0.025] [0.081] [0.007]** [0.040]*** [0.034] [0.050]* [0.006] Agriculture -0.007 0.131 0.118 0.030 0.016 0.000 0.032 -0.004 0.002 [0.077] [0.054]** [0.029]*** [0.019] [0.073] [0.035] [0.029] [0.052] [0.004] Married 0.049 0.104 0.068 0.063 -0.022 0.003 -0.029 -0.016 0.048 0.024 0.099 0.001 [0.008]*** [0.050]** [0.039]* [0.028]** [0.018] [0.003] [0.034] [0.008]** [0.026]* [0.019] [0.025]*** [0.002] HH_head 0.013 -0.048 0.090 0.003 -0.025 0.000 -0.016 -0.009 -0.011 0.002 0.071 -0.013 [0.007]* [0.047] [0.047]* [0.034] [0.020] [0.000] [0.040] [0.013] [0.027] [0.019] [0.028]** [0.005]** Earner 0.081 -0.068 -0.331 0.099 0.005 [0.049] [0.037]* [0.152]** [0.028]*** [0.003] Decision_taker 0.108 -0.034 -0.007 -0.018 -0.002 [0.030]*** [0.021] [0.037] [0.029] [0.019] Riskaverse 0.044 -0.000 0.020 -0.010 -0.003 [0.040] [0.000] [0.015] [0.031] [0.002] Numeracy 0.072 -0.004 0.063 0.000 -0.014 0.053 -0.017 -0.089 -0.002 [0.039]* [0.027] [0.034]* [0.000] [0.035] [0.025]** [0.020] [0.031]*** [0.003] Botswana 0.282 [0.036]*** Kenya06 0.611 [0.049]*** Kenya09 0.397 [0.044]*** Malawi 0.221 [0.046]*** Namibia -0.077 [0.011]*** Rwanda 0.359 [0.052]*** Tanzania06 0.344 [0.051]*** Tanzania09 0.388 [0.052]*** Uganda 0.242 [0.046]*** Zambia -0.009 [0.012] regional f.e. yes yes yes yes yes yes yes yes yes yes yes yes Observations 43044 974 4377 6589 4712 228 1894 3345 3097 5864 1587 2824 pseudo-r2 0.36 0.29 0.13 0.09 0.07 0.3 0.05 0.19 0.1 0.08 0.06 0.23 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Uganda06: household income; Kenya09: expenses; pooled regresions log-income (USD) 42 Table 11. Excluded from Financial Services (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) ALL Botswana Kenya06 Kenya09 Malawi Namibia Rwanda SouthAfrica Tanzania06 Tanzania09 Uganda Zambia Female -0.073 -0.129 -0.106 -0.087 0.013 0.015 0.072 0.056 -0.117 -0.075 -0.084 0.000 [0.019]*** [0.040]*** [0.037]***[0.023]*** [0.024] [0.040] [0.042]* [0.033]* [0.030]*** [0.023]*** [0.038]** [0.009] Rural 0.009 -0.069 0.047 0.024 -0.158 0.033 0.025 -0.004 -0.048 0.074 -0.000 -0.012 [0.021] [0.078] [0.031] [0.025] [0.028]*** [0.050] [0.045] [0.024] [0.040] [0.030]** [0.048] [0.010] Primary -0.137 -0.238 -0.037 -0.064 0.033 -0.112 -0.081 -0.139 0.015 -0.118 -0.079 -0.027 [0.024]*** [0.052]*** [0.035] [0.022]*** [0.022] [0.068]* [0.036]** [0.042]*** [0.042] [0.025]*** [0.037]** [0.014]* Secondary -0.338 -0.295 -0.076 -0.124 -0.136 -0.251 -0.217 -0.183 -0.083 -0.250 -0.326 -0.190 [0.025]*** [0.061]*** [0.041]* [0.028]***[0.036]***[0.071]***[0.071]*** [0.040]*** [0.044]* [0.044]*** [0.061]***[0.034]*** Tertiary -0.505 -0.399 -0.306 -0.182 -0.467 -0.481 -0.138 -0.587 -0.481 -0.351 -0.705 [0.020]*** [0.038]*** [0.018]***[0.039]***[0.041]*** [0.044]*** [0.019]*** [0.055]*** [0.069]*** [0.133]***[0.117]*** Age -0.259 -0.283 -0.043 -0.093 -0.072 -0.130 0.015 -0.195 0.001 -0.011 0.017 -0.049 [0.028]*** [0.079]*** [0.041] [0.027]***[0.028]*** [0.069]* [0.050] [0.055]*** [0.036] [0.031] [0.051] [0.014]*** Income -0.052 -0.087 -0.026 -0.044 -0.060 -0.016 -0.084 -0.025 -0.053 -0.009 [0.009]*** [0.012]***[0.004]***[0.009]***[0.018]*** [0.005]*** [0.013]*** [0.005]*** [0.012]***[0.001]*** Employed -0.405 -0.190 -0.208 -0.099 -0.108 -0.380 -0.055 -0.141 -0.096 -0.231 0.094 -0.149 [0.018]*** [0.053]*** [0.039]***[0.024]***[0.037]***[0.049]*** [0.086] [0.037]*** [0.044]** [0.059]*** [0.075] [0.032]*** Self_employed -0.202 -0.007 -0.218 -0.080 -0.013 -0.209 -0.254 -0.015 -0.108 -0.041 0.117 -0.023 [0.024]*** [0.072] [0.036]***[0.027]*** [0.033] [0.058]***[0.078]*** [0.056] [0.043]** [0.039] [0.066]* [0.016] Agriculture 0.201 -0.139 -0.084 -0.045 -0.003 -0.026 0.056 0.047 0.074 0.005 [0.096]** [0.039]***[0.021]*** [0.025]* [0.095] [0.087] [0.037] [0.036] [0.065] [0.012] Owns_mobile -0.290 -0.190 -0.162 -0.171 -0.168 -0.303 -0.100 -0.052 -0.118 -0.221 -0.311 -0.063 [0.020]*** [0.042]*** [0.036]***[0.025]***[0.022]***[0.046]*** [0.070] [0.039] [0.036]*** [0.027]*** [0.039]***[0.020]*** Married -0.154 -0.178 -0.129 -0.040 0.042 -0.145 -0.095 0.042 -0.057 -0.069 -0.097 -0.021 [0.019]*** [0.053]*** [0.034]*** [0.023]* [0.024]* [0.052]*** [0.039]** [0.039] [0.030]* [0.024]*** [0.034]*** [0.009]** HH_head -0.061 -0.018 -0.083 -0.004 0.045 -0.046 -0.034 0.007 0.021 -0.016 -0.111 0.012 [0.021]*** [0.059] [0.041]** [0.027] [0.027]* [0.050] [0.045] [0.039] [0.031] [0.025] [0.039]*** [0.016] Earner -0.081 0.037 -0.026 0.606 -0.099 -0.008 [0.056] [0.042] [0.046] [0.071]*** [0.041]** [0.016] Decision_taker -0.080 -0.009 -0.007 -0.019 0.015 [0.028]*** [0.027] [0.047] [0.033] [0.023] Riskaverse 0.027 -0.029 -0.061 0.002 0.016 [0.046] [0.043] [0.024]** [0.040] [0.009]* Numeracy -0.140 -0.054 -0.211 -0.166 -0.029 -0.098 -0.021 -0.048 -0.016 [0.044]*** [0.021]***[0.040]***[0.057]*** [0.042] [0.029]*** [0.024] [0.043] [0.010] Bostwana 0.191 [0.048]*** Kenya06 0.024 [0.051] Kenya09 0.120 [0.044]*** Malawi 0.315 [0.033]*** Namibia 0.158 [0.045]*** Rwanda 0.098 [0.047]** Tanzania06 0.263 [0.034]*** Tanzania09 0.176 [0.040]*** Uganda 0.297 [0.034]*** Zambia 0.548 [0.023]*** regional f.e. yes yes yes yes yes yes yes yes yes yes yes yes Observations 43044 974 4377 6589 4712 1065 1894 3345 3097 5864 1587 3152 pseudor2 0.39 0.35 0.2 0.25 0.11 0.48 0.11 0.31 0.16 0.19 0.18 0.5 Robust standard errors in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Uganda06: household income; Kenya09: expenses; pooled regresions log-income (USD) 43 Table 12. Economic Effect of Gender Gaps in the Use of Formal Finance ALL Botswana Kenya06 Kenya09 Malawi Namibia Gap*bet Gap*bet Gap tstat Gap*beta Gap tstat Gap*beta Gap tstat Gap*beta Gap tstat Gap*beta Gap tstat a Gap tstat a Rural 0.01 -1.60* 0.0000 -0.01 0.28 -0.0013 0.04 -2.61*** -0.0052 0.01 -1.07 -0.0008 0.01 -0.55 -0.0003 0.01 -0.37 -0.0005 Primary -0.03 6.10*** -0.0055 0.07 -2.58*** 0.0178 -0.06 4.24*** -0.0012 0.00 -0.03 0.0001 0.00 0.12 0.0000 0.02 -0.67 0.0018 Secondary -0.05 13.96*** -0.0236 -0.01 0.45 -0.0044 -0.08 6.19*** -0.0261 -0.08 7.71*** -0.0273 -0.05 4.89*** -0.0067 -0.05 1.76** -0.0115 Tertiary -0.01 7.78*** -0.0077 -0.04 3.05*** -0.0213 -0.02 5.09*** -0.0163 -0.02 4.64*** -0.0092 0.00 0.2 -0.0005 -0.01 0.68 Age -0.05 13.33*** -0.0132 0.01 -0.32 0.0014 -0.05 4.28*** -0.0120 -0.06 5.59*** -0.0118 -0.07 5.91*** -0.0034 0.02 -0.99 0.0022 Income -0.53 13.44*** -0.32 1.50* -0.0248 -0.2 7.45*** -0.0294 -0.39 4.81*** -0.0142 -0.69 3.07*** -0.0316 Employed -0.08 20.93*** -0.0311 -0.12 4.43*** -0.0141 -0.12 9.19*** -0.0219 -0.12 11.94***-0.0160 -0.04 3.95*** -0.0053 -0.09 3.37*** -0.0354 Self_employed 0.00 -0.14 0.0001 0.04 -2.12** -0.0006 0.02 -1.71** 0.0025 0.01 -0.53 0.0004 0.00 -0.34 0.0000 0.01 -0.83 0.0028 Agriculture -0.02 4.07*** -0.02 1.17 0.0032 -0.02 1.38* -0.0037 -0.03 2.29** -0.0018 -0.01 0.54 -0.0004 -0.01 1.19 0.0004 Owns_mobile -0.06 13.69*** -0.0197 0.04 -1.27 0.0070 -0.1 7.07*** -0.0293 -0.12 9.93*** -0.0365 -0.03 2.47*** -0.0060 -0.07 2.63*** -0.0225 HH_head -0.4 90.59*** -0.0321 -0.16 5.84*** 0.0021 -0.39 28.76*** 0.0054 -0.48 43.84*** -0.0125 -0.55 46.07*** 0.0027 -0.07 2.39*** -0.0027 Married -0.05 10.12*** -0.0047 -0.01 0.4 -0.0012 -0.06 3.73*** -0.0017 -0.06 4.95*** 0.0021 -0.1 8.39*** 0.0019 -0.03 1.21 -0.0045 Decision_tak -0.05 8.54*** -0.04 3.18*** -0.0040 -0.09 7.87*** -0.0024 Earner -0.19 33.36*** -0.12 4.03*** -0.0131 -0.06 9.73*** -0.0009 0.05 -1.95** 0.0014 Riskaverse 0.03 -3.70*** 0.08 -3.47*** -0.0070 -0.02 0.8 -0.0006 Numeracy -0.08 15.71*** -0.07 2.74*** -0.0046 -0.15 12.76***-0.0156 -0.02 2.00** -0.0023 -0.01 0.61 -0.0018 Rwanda SouthAfrica Tanzania06 Tanzania09 Uganda Zambia Gap*bet Gap*bet Gap tstat Gap*beta Gap tstat Gap*beta Gap tstat Gap*beta Gap tstat Gap*beta Gap tstat a Gap tstat a Rural 0.00 -0.07 0.0000 0.00 -0.31 -0.0003 0.04 -2.96*** -0.0010 -0.08 8.21*** 0.0002 -0.07 4.16*** 0.0003 0.00 -0.17 0.0000 Primary -0.05 2.22** -0.0058 0.03 -1.62* 0.0053 -0.06 4.38*** -0.0027 -0.06 5.07*** -0.0014 -0.04 1.91** -0.0033 0.00 0.29 0.0000 Secondary -0.02 2.13** -0.0101 -0.04 2.57*** -0.0124 -0.03 2.59*** -0.0034 -0.02 2.69*** -0.0030 -0.05 3.71*** -0.0144 -0.08 5.47*** -0.0089 Tertiary -0.01 2.14** -0.0097 0.00 0.05 -0.0001 -0.01 2.41*** -0.0048 -0.01 4.12*** -0.0053 -0.01 2.54*** -0.0054 -0.01 3.71*** -0.0073 Age -0.03 1.55* -0.0021 0.01 -0.42 0.0014 -0.08 6.85*** -0.0019 -0.09 10.28***-0.0039 -0.03 2.00** 0.0005 -0.1 8.39*** -0.0021 Income -0.46 6.24*** -0.0153 -0.23 1.92** -0.0059 -1.43 12.44*** -0.0445 -0.66 9.27*** -0.0066 -0.25 3.58*** -0.0091 -1.93 8.73*** -0.0077 Employed -0.06 4.72*** -0.0081 -0.11 7.18*** -0.0169 -0.07 7.51*** -0.0018 -0.02 4.88*** -0.0021 -0.05 4.02*** -0.0005 -0.07 5.81*** -0.0038 Self_employed -0.02 1.31* -0.0037 -0.02 2.30** -0.0008 -0.03 2.70*** 0.0003 -0.02 2.23** 0.0003 0.03 -1.93** -0.0016 0.03 -2.25** 0.0002 Agriculture 0.07 -3.50*** 0.0028 -0.05 4.10*** 0.0011 -0.05 4.74*** 0.0024 -0.03 1.49* 0.0025 -0.08 6.06*** -0.0001 Owns_mobile 0 0.2 -0.0002 0.03 -2.43*** 0.0038 -0.08 7.58*** -0.0035 -0.08 7.72*** -0.0095 -0.07 4.42*** -0.0213 -0.04 2.80*** -0.0009 HH_head -0.44 20.80*** -0.0022 -0.28 18.56*** 0.0077 -0.21 15.78*** 0.0008 -0.49 48.90*** -0.0010 -0.48 29.88*** -0.0201 -0.42 29.71*** 0.0000 Married -0.13 5.73*** -0.0028 -0.09 5.62*** 0.0028 -0.1 7.41*** -0.0005 0.03 -2.86*** 0.0003 -0.03 1.65** 0.0002 -0.04 2.48*** -0.0001 Decision_taker 0.01 -0.37 0.0004 -0.04 3.39*** -0.0005 -0.04 3.55*** 0.0001 Earner -0.11 7.77*** 0.0975 -0.07 9.62*** 0.0015 -0.4 27.72*** 0.0004 Riskaverse 0.04 -3.17*** 0.0007 0.00 0.05 0.0000 0.03 -2.55*** -0.0001 Numeracy -0.06 3.06*** -0.0019 -0.13 9.41*** -0.0023 -0.12 11.29***-0.0024 -0.07 5.57*** -0.0089 -0.07 5.65*** -0.0004 note: bold if beta siginificant at 10% (table 9) 44 Table A1. Sample for Enterprise Analysis country obs country obs Albania2007 294 Kyrgyz Republic2009 230 Albania2009 172 Lao PDR2009 360 Angola2006 423 Latvia2009 256 Argentina2006 1,025 Lithuania2009 269 Armenia2009 371 Macedonia, FYR2009 360 Azerbaijan2009 374 Mali2007 610 Belarus2008 262 Mauritania2006 237 Bhutan2009 250 Mexico2006 1,415 Bolivia2006 598 Moldova2009 356 Botswana2006 342 Mongolia2009 360 Brazil2009 1,185 Mozambique2007 479 Bulgaria2007 1,002 Namibia2006 322 Bulgaria2009 282 Nepal2009 368 Burkina Faso2006 133 Nicaragua2006 469 Burundi2006 268 Nigeria2007 2,309 Cameroon2006 153 Panama2006 581 Cape Verde2006 96 Paraguay2006 597 Chile2006 994 Peru2006 625 Colombia2006 993 Poland2009 422 Congo, Dem. Rep.2006 339 Romania2009 501 Croatia2007 615 Russian Federation2009 969 Croatia2009 157 Rwanda2006 209 Czech Republic2009 238 Senegal2007 505 Ecuador2006 647 Serbia2009 381 El Salvador2006 676 Slovak Republic2009 258 Estonia2009 254 Slovenia2009 270 Gambia, The2006 174 SouthAfrica2007 925 Georgia2008 363 Swaziland2006 305 Ghana2007 494 Tajikistan2008 351 Guatemala2006 511 Tanzania2006 413 Guinea-Bissau2006 159 Turkey2008 1,134 Honduras2006 430 Uganda2006 563 Hungary2009 281 Ukraine2008 806 Kazakhstan2009 533 Uruguay2006 582 Kenya2007 654 Zambia2007 596 Total 35,135 45 Table A2. Summary Statistics Household Surveys by Country Bostwana 2004 Kenya 2006 Kenya 2009 Malawi 2008 Namibia 2004 Rwanda 2008 Variable Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Banking 1200 0.43 0.50 0 1 4418 0.19 0.39 0 1 6598 0.25 0.44 0 1 4993 0.19 0.39 0 1 1200 0.53 0.50 0 1 2000 0.14 0.35 0 1 Informal 1200 0.31 0.46 0 1 4418 0.50 0.50 0 1 6598 0.50 0.50 0 1 4993 0.25 0.43 0 1 1200 0.01 0.08 0 1 2000 0.26 0.44 0 1 Excluded 1200 0.46 0.50 0 1 4418 0.39 0.49 0 1 6598 0.36 0.48 0 1 4993 0.55 0.50 0 1 1200 0.46 0.50 0 1 2000 0.53 0.50 0 1 Transactions 1200 0.58 0.49 0 1 4418 0.32 0.47 0 1 6598 0.49 0.50 0 1 4993 0.40 0.49 0 1 1200 0.58 0.49 0 1 2000 0.21 0.40 0 1 Savings 1200 0.51 0.50 0 1 4418 0.39 0.49 0 1 6598 0.53 0.50 0 1 4993 0.36 0.48 0 1 1200 0.52 0.50 0 1 2000 0.38 0.49 0 1 Credit 1200 0.20 0.40 0 1 4418 0.22 0.41 0 1 6598 0.22 0.41 0 1 4993 0.19 0.39 0 1 1200 0.10 0.30 0 1 2000 0.24 0.43 0 1 Insurance 1200 0.29 0.45 0 1 4418 0.07 0.25 0 1 6598 0.09 0.29 0 1 4993 0.05 0.22 0 1 1200 0.18 0.38 0 1 2000 0.81 0.39 0 1 Female 1200 0.51 0.50 0 1 4418 0.56 0.50 0 1 6598 0.59 0.49 0 1 4993 0.52 0.50 0 1 1200 0.50 0.50 0 1 2000 0.64 0.48 0 1 Married 1197 0.21 0.40 0 1 4418 0.61 0.49 0 1 6598 0.60 0.49 0 1 4993 0.74 0.44 0 1 1200 0.25 0.43 0 1 2000 0.55 0.50 0 1 HH_head 1199 0.41 0.49 0 1 4418 0.38 0.48 0 1 6598 0.49 0.50 0 1 4993 0.50 0.50 0 1 1200 0.38 0.48 0 1 2000 0.51 0.50 0 1 No-education 1200 0.27 0.44 0 1 4402 0.43 0.50 0 1 6598 0.43 0.50 0 1 4993 0.18 0.39 0 1 1200 0.14 0.34 0 1 2000 0.42 0.49 0 1 Primary 1200 0.36 0.48 0 1 4402 0.36 0.48 0 1 6598 0.32 0.46 0 1 4993 0.55 0.50 0 1 1200 0.44 0.50 0 1 2000 0.50 0.50 0 1 Secondary 1200 0.32 0.47 0 1 4402 0.23 0.42 0 1 6598 0.23 0.42 0 1 4993 0.13 0.34 0 1 1200 0.40 0.49 0 1 2000 0.06 0.24 0 1 Tertiary 1200 0.06 0.23 0 1 4402 0.02 0.15 0 1 6598 0.02 0.14 0 1 4993 0.02 0.15 0 1 1200 0.03 0.16 0 1 2000 0.01 0.12 0 1 Numeracy 1200 0.34 0.48 0 1 0 6598 0.40 0.49 0 1 4993 0.09 0.29 0 1 1200 0.20 0.40 0 1 2000 0.78 0.41 0 1 Owns_mobile 1200 0.38 0.48 0 1 4393 0.27 0.45 0 1 6598 0.47 0.50 0 1 4993 0.33 0.47 0 1 1200 0.37 0.48 0 1 2000 0.07 0.25 0 1 Age_log 1200 3.48 0.39 2.89 4.47 4418 3.51 0.41 2.77 4.50 6597 3.57 0.42 2.77 4.65 4993 3.51 0.40 2.89 4.58 1198 3.45 0.37 2.77 4.48 2000 3.55 0.40 2.89 4.51 Employed 1200 0.37 0.48 0 1 4418 0.24 0.43 0 1 6598 0.22 0.42 0 1 4993 0.15 0.36 0 1 1200 0.40 0.49 0 1 2000 0.10 0.30 0 1 Self_emplo~d 1200 0.10 0.29 0 1 4418 0.20 0.40 0 1 6598 0.20 0.40 0 1 4993 0.15 0.35 0 1 1200 0.11 0.31 0 1 2000 0.11 0.31 0 1 Agriculture 1200 0.06 0.23 0 1 4418 0.35 0.48 0 1 6598 0.34 0.47 0 1 4993 0.46 0.50 0 1 1200 0.05 0.21 0 1 2000 0.72 0.45 0 1 Riskaverse 1200 0.21 0.41 0 1 0 0 0 1200 0.35 0.48 0 1 0 Rural 1200 0.67 0.47 0 1 4418 0.68 0.47 0 1 6598 0.71 0.45 0 1 4993 0.81 0.39 0 1 1200 0.55 0.50 0 1 2000 0.75 0.44 0 1 Earner 1154 0.40 0.49 0 1 0 0 4968 0.95 0.23 0 1 1200 0.24 0.43 0 1 2000 0.36 0.48 0 1 Decision_mkr 0 0 6598 0.73 0.44 0 1 4993 0.80 0.40 0 1 0 2000 0.77 0.42 0 1 log_income 1023 4.70 3.44 -0.69 10.13 0 0 4737 7.85 2.82 -0.69 14.22 1091 4.63 3.71 -0.69 10.82 1894 8.60 1.57 -0.69 14.91 log_incomehh 737 6.86 2.31 -0.69 11.00 0 0 0 816 6.84 2.98 -0.69 12.21 0 log_expenses 0 0 6590 8.90 1.09 3.00 13.61 0 0 0 region 1200 3.41 1.08 1 4 4418 5 2 1 8 6598 4.30 2.12 1 8 4993 2.31 0.76 1 3 1200 7.27 3.56 1 13 2000 3.31 1.38 1 5 weight 1200 1.00 0.33 0.56 5.09 4418 4127 2778 33.294 25576 6598 1.00 0.80 0.03 9.02 4993 1212.33 726.71 197.68 3844.6 0 2000 1846.9 1235.7 100.7 9377.4 46 SouthAfrica 2008 Tanzania 2006 Tanzania 2009 Uganda 2006 Zambia 2005 All Variable Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max n. obs. mean st.dev. min max Banking 3900 0.70 0.46 0 1 4962 0.10 0.31 0 1 7680 0.11 0.32 0 1 2959 0.16 0.37 0 1 3998 0.14 0.35 0 1 43908 0.23 0.42 0 1 Informal 3900 0.08 0.28 0 1 4962 0.24 0.43 0 1 7680 0.27 0.44 0 1 2959 0.21 0.41 0 1 3998 0.03 0.17 0 1 43908 0.27 0.45 0 1 Excluded 3900 0.19 0.39 0 1 4962 0.66 0.47 0 1 7680 0.57 0.49 0 1 2959 0.62 0.48 0 1 3998 0.83 0.38 0 1 43908 0.57 0.49 0 1 Transactions 3900 0.75 0.43 0 1 4962 0.26 0.44 0 1 7680 0.15 0.36 0 1 2959 0.39 0.49 0 1 3998 0.26 0.44 0 1 43908 0.36 0.48 0 1 Savings 3900 0.41 0.49 0 1 4962 0.54 0.50 0 1 7680 0.68 0.46 0 1 2959 0.72 0.45 0 1 3998 0.23 0.42 0 1 43908 0.49 0.50 0 1 Credit 3900 0.32 0.47 0 1 4962 0.20 0.40 0 1 7680 0.17 0.37 0 1 2959 0.29 0.45 0 1 3998 0.05 0.21 0 1 43908 0.20 0.40 0 1 Insurance 3900 0.52 0.50 0 1 4962 0.06 0.24 0 1 7680 0.06 0.24 0 1 0 3998 0.06 0.25 0 1 40949 0.16 0.36 0 1 Female 3900 0.50 0.50 0 1 4962 0.47 0.50 0 1 7680 0.53 0.50 0 1 2959 0.52 0.50 0 1 3998 0.50 0.50 0 1 43908 0.53 0.50 0 1 Married 3900 0.38 0.49 0 1 4962 0.59 0.49 0 1 7680 0.73 0.44 0 1 2959 0.55 0.50 0 1 3998 0.51 0.50 0 1 43905 0.59 0.49 0 1 HH_head 3900 0.48 0.50 0 1 4962 0.34 0.47 0 1 7672 0.54 0.50 0 1 2957 0.55 0.50 0 1 3998 0.42 0.49 0 1 43897 0.46 0.50 0 1 No-education 3900 0.06 0.24 0 1 4798 0.23 0.42 0 1 7666 0.27 0.44 0 1 2485 0.44 0.50 0 1 3998 0.26 0.44 0 1 43240 0.29 0.45 0 1 Primary 3900 0.46 0.50 0 1 4798 0.51 0.50 0 1 7666 0.64 0.48 0 1 2485 0.43 0.50 0 1 3998 0.47 0.50 0 1 43240 0.47 0.50 0 1 Secondary 3900 0.37 0.48 0 1 4798 0.25 0.44 0 1 7666 0.09 0.28 0 1 2485 0.11 0.31 0 1 3998 0.26 0.44 0 1 43240 0.20 0.40 0 1 Tertiary 3900 0.11 0.31 0 1 4798 0.01 0.11 0 1 7666 0.01 0.08 0 1 2485 0.02 0.13 0 1 3998 0.01 0.11 0 1 43240 0.03 0.16 0 1 Numeracy 0 4962 0.50 0.50 0 1 7680 0.69 0.46 0 1 2959 0.16 0.36 0 1 3998 0.22 0.41 0 1 35590 0.41 0.49 0 1 Owns_mobile 3900 0.73 0.44 0 1 4962 0.19 0.39 0 1 7680 0.26 0.44 0 1 2959 0.28 0.45 0 1 3998 0.20 0.40 0 1 43883 0.33 0.47 0 1 Age_log 3900 3.56 0.41 2.77 4.60 4959 3.48 0.41 2.77 4.60 7680 3.52 0.38 2.77 4.60 2801 3.49 0.37 2.89 4.55 3990 3.34 0.37 3 4 43736 3.51 0.40 2.77 4.65 Employed 3900 0.37 0.48 0 1 4962 0.14 0.34 0 1 7680 0.04 0.20 0 1 2959 0.15 0.35 0 1 3998 0.19 0.40 0 1 43908 0.18 0.39 0 1 Self_emplo~d 3900 0.07 0.26 0 1 4962 0.20 0.40 0 1 7680 0.18 0.38 0 1 2959 0.28 0.45 0 1 3998 0.17 0.38 0 1 43908 0.17 0.38 0 1 Agriculture 0 4962 0.28 0.45 0 1 7680 0.60 0.49 0 1 2959 0.46 0.50 0 1 3998 0.26 0.44 0 1 40008 0.40 0.49 0 1 Riskaverse 3900 0.24 0.43 0 1 0 0 2959 0.21 0.41 0 1 3998 0.25 0.43 0 1 13257 0.24 0.43 0 1 Rural 3900 0.24 0.43 0 1 4962 0.45 0.50 0 1 7680 0.77 0.42 0 1 2959 0.71 0.45 0 1 3998 0.68 0.47 0 1 43908 0.65 0.48 0 1 Earner 0 4962 0.62 0.48 0 1 7680 0.89 0.31 0 1 0 3998 0.45 0.50 0 1 25962 0.69 0.46 0 1 Decision_mkr 0 4588 0.78 0.41 0 1 7680 0.69 0.46 0 1 0 0 25859 0.74 0.44 0 1 log_income 3345 5.80 3.42 -0.69 11.04 3427 9.08 3.43 -0.69 12.43 5886 9.94 2.76 -0.69 14.22 0 3158 8.67 6.27 -1 16 log_incomehh 3307 7.14 2.42 -0.69 11.04 2604 10.10 2.58 -0.69 12.43 5230 11.13 1.14 -0.69 14.22 1956 10.10 1.56 5.12 19.62 2292 12.07 3.97 -1 16 31151 2.93 3.47 -9.10 9.37 log_expenses 0 0 0 0 0 region 3900 4.88 2.54 1 9 4962 23.3 20.5 1 55 7680 13.8 12.8 1 55 2959 3.3 1.3 1 5 3998 5 3 1 9 43908 8.0 11.1 1 55 weight 3900 8201.1 8439.0 21.9 59288 4962 4259.6 5333.7 149.1 31201 7680 5457.8 5254.8 130.0 122826 2959 4452.2 2685.9 332 25309 3998 1 1 0.16 3.09 42708 3189.2 4806.8 0.026 122826 47 Figure 1: Women, Business and Law index across the world. 1.00 0.80 0.60 0.40 0.20 ` Sub-Saharan Africa Other Sample size: 79 countries Figure 2. Gender Gap in the Use of Finacial Services Panel A: Banking Panel B: Informal 60 80 70 Male 50 Male 60 Female 40 Female 50 30 40 30 20 20 10 10 0 0 Banking Informal Excluded Panel C: Excluded 100 country Gap tstat Gap tstat Gap tstat 90 Male Bostwana -2 0.89 4 -1.31 -4 1.33 80 Female Kenya06 -9 7.73 10 -6.55 0.00 0.38 70 60 Kenya09 -11 10.99 13 -10.7 -2 1.17 50 Malawi -3 2.21 -1 0.89 4 -2.84 40 Namibia -7 2.48 -1 2.11 7 -2.48 30 20 Rwanda -5 2.6 -4 1.76 8 -3.77 10 SouthAfrica -4 3.06 3 -3.02 -1 0.4 0 Tanzania06 -8 9.3 5 -4.65 3 -2.03 Tanzania09 -4 5.3 3 -3.09 1 -1.06 Uganda -10 7.33 1 -0.16 8 -4.05 Zambia -6 5.37 0 -1.24 6 -4.83 ALL -7 17.31 6 -14.1 1 -2.23 48