Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/oxford-university-press/financial-literacy-and-intra-household-decision-making-evidence-from-N80YVNLMbw
References (3)
-
Alsop (2005)
Measuring empowerment: An analytic framework -
Allendorf (2007)
Do women’s land rights promote empowerment and child health in Nepal?World Development, 35
-
Abbott (2018)
Women, land and empowerment in RwandaJournal of International Development, 30
- Publisher
- Oxford University Press
- Copyright
- Copyright © 2021 Centre for the Study of African Economies
- ISSN
- 0963-8024
- eISSN
- 1464-3723
- DOI
- 10.1093/jae/ejaa007
- Publisher site
- See Article on Publisher Site
Abstract
Abstract Research has consistently shown that women’s involvement in household decision making positively affects household outcomes such as nutrition and education of children. Is financial literacy a determinant for women to participate in intra-household decision making? Using data on savings groups in Rwanda, we examine this relationship and show that women with higher financial literacy are more involved in financial and expenditure decisions. Instrumental variable estimations suggest a causal link. For this reason, we perform a decomposition analysis breaking down the gender gap in financial literacy into differences based on observed socio-demographic and psychological characteristics and differences in returns on these characteristics. Our results show high explanatory power by education, happiness, symptoms of depression and openness but also suggest that a substantial fraction can be explained by differences in returns. We argue that this results from a strong role of society and culture. 1. Introduction Strengthening women empowerment within the household is not only a desirable goal in itself but also has other positive welfare effects (Duflo, 2012). Stronger involvement of women in household decision making can have important effects on outcomes such as child mortality (Moursund & Kravdal, 2003, Thomas, 1990), nutrition, health and education of children (DFID, 2010, Duflo, 2003, Thomas, 1993). As one of the Sustainable Development Goals, the United Nations (UN) has therefore announced women empowerment and higher involvement in household decisions as an integral part of the 2030 agenda (UN, 2015). Liberating and enhancing women’s capacity to make choices within the household are crucial to women empowerment (Alsop et al., 2005, Manda & Mwakubo, 2014). Kabeer (1999) conceptualizes women empowerment and divides the ability to make choices into three moments in time. She frames the first moment as pre-condition or resource, the second moment as action or agency and the third one as achievement; whereby agency tends to be operationalized as decision making. Based on early household models (McElroy & Horney, 1981), resources often comprise material resources, such as income (Anderson & Eswaran, 2009, Bobonis, 2009, De Brauw et al., 2014) and land ownership (Allendorf, 2007, Doss, 2006). More recently, resources have also been defined more broadly as human capital (see Doss, 2013, Fiala & He, 2017, for reviews). This article contributes to the literature on the determinants of women’s agency at home by examining the effect of a specific type of human capital, that is, financial literacy, on a specific type of agency, that is, financial decision making within the household. Throughout this manuscript, financial literacy will refer to understanding of financial concepts, such as interest rate, risk diversification and inflation. Following the framework developed by Kabeer (1999), financial literacy should act as a resource of empowerment by increasing women’s ability and self-confidence to make financial decisions and ultimately enhance their involvement in intra-household decision making. Using household data of savings group members in Rwanda, we further aim to understand an important mechanism behind why researchers find that membership in groups that jointly perform financial tasks empowers women and increases their involvement in household decisions (Hashemi et al., 1996, Karlan et al., 2017): women may become more financially literate as part of their group membership and thus increase their decision-making power within the household. Rwanda is an ideal place to study savings groups because groups here are, in many ways, representative of savings groups in Africa (see Section 2 for details). We first run ordinary least squares (OLS) regressions to look at correlations between financial literacy and women’s involvement in households financial and expenditure decisions. We find that there is a strong and positive relationship. The cross-sectional design of this study, however, prevents us from making causal statements based on linear regressions. Although theory predicts a positive effect of financial literacy on women’s involvement in household financial decisions (Kabeer, 1999), causality may occur in both directions. To establish causality of this effect, we chose an instrumental variable (IV) regression approach. This is a common approach in both the literature on the determinants of women’s agency at home and the literature on financial literacy (Lusardi & Mitchell, 2011). For instance, Doss (2001) and Duflo & Udry (2004) use rainfall shocks to instrument for women’s agricultural income and find expenditure shifts towards education and food. Our identification strategy is based on financial literacy levels of other savings group members as an instrument for woman’s own financial literacy level. Higher financial literacy levels of peers provide women with more information on finances but are argued to not directly affect their decision-making power at home. OLS and IV estimates are comparable both in size and significance. This result and several robustness tests to validate our instrument allow us to conclude that the effect is causal and runs from financial literacy to increased involvement by women in household decision making. These results motivate further analysis on how to improve women’s financial literacy levels. We first show that, in line with the literature (Bucher-Koenen et al., 2017, Xu & Zia, 2012), women have lower financial literacy than men. So far, however, little discussion exists about the reasons behind this gender gap. That is why in a second step, we look at drivers behind this gender gap in financial literacy performing a detailed decomposition analysis. Unlike previous studies, we have information on deeply rooted personality traits in addition to more standard socio-demographic measures. The results show that 46% of the gap stems from gender differences in endowments, particularly women’s lower educational attainment (17%), lower openness for new ideas (12%), lower happiness (5%) and greater symptoms of depression (3%). A total of 54% of the gender gap stem from gender differences in returns on characteristics and thus remain unexplained by observed characteristics. We interpret this as a strong role of society and culture as has previously been shown by Filipiak & Walle (2015) for matrilineal societies in India and Grohmann et al. (2016) in Thailand. This article adds to the literature on the effect of financial literacy on financial decision making, as reviewed by Lusardi & Mitchell (2014). The research to date has tended to focus on financial outcomes rather than on decision-making processes. For example, IV analyses show that financial literacy improves retirement planning (Lusardi & Mitchell, 2007), wealth accumulation (van Rooij et al., 2012) and stock market participation (van Rooij et al., 2011) and reduces the amount of debt held (Gathergood, 2012, Lusardi & Tufano, 2015). In developing countries, the literature is less extensive and the majority experimentally evaluate financial literacy programmes (see Kaiser & Menkhoff, 2017, for a meta-analysis). In Indonesia and India, Cole et al. (2011) find only modest effects on account ownership for the poorest segment of the treated populations. Doi et al. (2014) and Sayinzoga et al. (2015) find significant impacts of financial literacy training on savings in the Philippines and Rwanda, respectively. A second set of studies has looked at the link between access to finances and household decision making. Using a randomized experiment, Ashraf et al. (2010) find that households are more likely to buy female-oriented durables when they get access to a commitment savings product. This implies women’s increased control over monetary decisions at home. Likewise, Hashemi et al. (1996) provide evidence that bank or committee memberships increase participation in household and purchasing decisions. Despite increasing evidence on the material resources on intra-household decision making, there are currently few studies which attempt to provide rigorous estimates of the impact of skills and human capital. Our research makes three main contributions. First, to our knowledge, this is the first study to empirically examine the link between financial literacy and intra-household decision making. Second, we use the framework of savings groups to understand whether financial literacy is one mechanism why group membership is often described to increase women’s decision-making power at home. Third, we disentangle the determinants of the gender gap in financial literacy taking into account a number of relevant personality traits. These results can aid in designing policies intended to increase financial literacy levels for women. In particular, the effect of education and personality, together with the large unexplained part of the decomposition analysis, suggests that women may benefit from tailored training that not only teaches financial concepts but also focuses on forming life skills, such as self-confidence and gender awareness. Following this introduction, the remainder of the manuscript is organized as follows: Section 2 describes the setting and Section 3 presents the data. Section 4 looks at women’s involvement in household decisions. In Section 5, we perform a decomposition analysis. Section 6 provides robustness and Section 7 concludes. 2. Setting 2.1. Country context and savings groups The Rwandan government and development organizations have made great efforts to promote women empowerment leading to a more balanced picture between men and women in comparison with other sub-Saharan African countries. Eventhough in Rwanda, women are well represented in Parliament and other leadership organs, the country is still a patriarchal society. Cultural norms persist and men are still the decision makers (Abbott et al., 2018). The majority of our female sample is informally employed in the agricultural sector with low probability to be economically included. In this context, women still lag behind in decision making both socially and economically and the question how to increase their household decision-making power is highly relevant. Our analysis relies on primary household data of savings groups members in the Southern Province of Rwanda, a rural area where the majority of people save in informal groups such as tontines or Village Savings and Loan Associations (VSLAs).1 Among 30 African economies, a recent VSLA global outreach report states Rwanda with 19,634 VSLAs in fourth place in absolute terms. Given its population, Rwanda has the second greatest density of VSLAs in Africa.2 Every 10th person listed as a VSLA member in Africa lives in Rwanda (CARE, 2017). About 54% of adults in Rwanda use informal savings groups to manage their savings (FinScope, 2016). In the rural South, this number is slightly higher, and savings groups provide an important tool of ensuring financial inclusion for the most vulnerable. A typical VSLA in Rwanda consists of 15 to 30 people and is gender mixed.3 According to the VSLA global outreach report, Rwanda’s VSLAs have on average 29 members of which 79% are women (CARE, 2017). A total of 77% in our sample are women and selected VSLAs have on average 28 members. Therefore, the sample is representative for Rwanda’s VSLAs because the characteristics represent the parent population in relevant ways. Members meet once a week to contribute to or borrow from a shared fund. Savings are often as little as one or two hundred Rwandan Francs (RWF) (less than 0.25 USD) per week. Eight to twelve months after the savings circle has started, each member will receive her share-out of the fund and her accumulated savings. It is likely that this regular meeting and contribution structure may increase understanding of financial concepts and that this, together with selection issues, means that our sample has higher interest in money management that likely goes beyond the financial literacy of other rural residents. The decomposition analysis benefits from this, as unobservable factors related to financial interest can to some extent be neglected. Given that members voluntarily select themselves into groups, it is possible that the composition of groups is related to wealth, education or other socio-demographic characteristics. A comparison of our sample to the Rwandan Housing and Population Census 2012 (NISR, 2012) shows that the sample is comparatively less educated and poorer than the overall Rwandan population. This is, however, not systematic between men and women. 2.2. Sampling Sampling was done in two random stages. First, we stratified the sample by district and drew a total of 300 VSLAs from a complete list of all active VSLAs in southern Rwanda.4 Second, we randomly selected five individuals from each VSLA. This was done by first compiling a list of all active members of the visited VSLA. Using smart mobile devices, a random number generator then randomly selected five names from this list. Our sample is, hence, representative for VSLA members in Rwanda’s Southern Province. The target population is older than 18 years. Respondents also qualify as poor according to Rwanda’s poverty levels5 and have limited access to formal financial services provider. We designed the questionnaire specifically to answer questions regarding financial issues of the household. It contains questions on the household’s socio-demographic variables, household composition, intra-household decision making, financial services used and financial literacy. Each interview took about 45 minutes and was conducted in a privacy-secured setting without partner and other family members present. The final sample collected in 2015 includes 283 of the 300 selected VSLAs and 1405 respondents, 1081 women and 324 men. 17 VSLAs from the initial list of active groups either no longer existed or could not be reached. No VSLA refused to participate in the survey. 3. Descriptive statistics and variables 3.1. Socio-demographics Summary statistics are presented in Table 1 separated by gender.6 Respondents are on average 43 years old, married and poorly educated. Only 57% of women and 72% of men can spell a simple word in the local language correctly. Women are also more likely to be widowed than men. Looking at measures for personality such as happiness and the depression index7, women’s indices are below that of men. The majority of respondents report farming as their main occupation. The average household size is about five. What is interesting in this data is that the highest proportion of female savings groups members tend to belong to the lowest income quartile, whereas the highest proportion of their male counterparts belong to the upper income quartile.8 We also construct an asset index that is the first principal component of the respondents reported assets. This asset index indicates that women participate in VSLAs out of poorer households than men. Moreover, mobile phone ownership is less likely in households of female than male savings groups members. Table 1 Descriptive statistics . Women . Men . . . . Mean . S.D. . Obs. . Mean . S.D. . Obs. . Difference . |$p$|-Value . Outcome: decisions Income: self 0.41 0.49 1,081 0.29 0.45 324 -0.124 0.000 Income: both 0.46 0.50 1,081 0.63 0.48 324 0.176 0.000 Income: involved 0.87 0.33 1,081 0.92 0.27 324 0.051 0.009 Credit: self 0.41 0.49 1,081 0.27 0.44 324 -0.145 0.000 Credit: both 0.55 0.50 1,081 0.69 0.46 324 0.144 0.000 Credit: involved 0.96 0.19 1,081 0.96 0.19 324 -0.001 0.944 Investment: self 0.42 0.49 1,054 0.27 0.45 317 -0.153 0.000 Investment: both 0.51 0.50 1,054 0.66 0.47 317 0.149 0.000 Investment: involved 0.93 0.25 1,054 0.93 0.25 317 -0.004 0.846 Financial decisions—index 2.78 0.59 1,054 2.82 0.60 317 0.042 0.346 Food: self 0.58 0.49 1,080 0.25 0.43 321 -0.333 0.000 Food: both 0.36 0.48 1,080 0.50 0.50 321 0.143 0.000 Food: involved 0.93 0.25 1,080 0.74 0.44 321 -0.190 0.000 Own health: self 0.63 0.48 1,081 0.56 0.50 324 -0.065 0.049 Own health: both 0.31 0.46 1,081 0.37 0.48 324 0.054 0.094 Own health: involved 0.94 0.23 1,081 0.93 0.25 324 -0.011 0.488 Own clothes: self 0.61 0.49 1,081 0.56 0.50 324 -0.046 0.187 Own clothes: both 0.33 0.47 1,081 0.39 0.49 324 0.055 0.100 Own clothes: involved 0.94 0.23 1,081 0.95 0.21 324 0.009 0.537 Energy: self 0.50 0.50 1,079 0.33 0.47 323 -0.171 0.000 Energy: both 0.36 0.48 1,079 0.47 0.50 323 0.106 0.002 Energy: involved 0.86 0.35 1,079 0.80 0.40 323 -0.064 0.015 Adult’s exp. decisions—index (4) 3.68 0.78 1,078 3.42 0.97 321 -0.261 0.000 Child’s health: self 0.42 0.49 881 0.12 0.33 254 -0.296 0.000 Child’s health: both 0.54 0.50 881 0.81 0.40 254 0.271 0.000 Child’s health: involved 0.95 0.21 881 0.93 0.26 254 -0.024 0.178 Child’s clothes: self 0.45 0.50 880 0.13 0.34 254 -0.318 0.000 Child’s clothes: both 0.51 0.50 880 0.77 0.42 254 0.261 0.000 Child’s clothes: involved 0.96 0.20 880 0.90 0.30 254 -0.058 0.006 All exp. decisions—index (6) 5.61 0.99 877 5.29 1.24 254 -0.315 0.001 Explanatory variables Financial literacy—index 2.36 1.22 1,081 2.83 1.13 324 0.477 0.000 Age 43.90 13.39 1,081 40.51 13.94 324 -3.388 0.001 Spell 0.57 0.50 1,081 0.72 0.45 324 0.151 0.000 Secondary education and more 0.07 0.26 1,081 0.09 0.28 324 0.015 0.411 Single 0.06 0.24 1,081 0.18 0.38 324 0.113 0.000 Married 0.67 0.47 1,081 0.77 0.42 324 0.107 0.000 Divorced 0.07 0.25 1,081 0.03 0.16 324 -0.042 0.000 Widowed 0.20 0.40 1,081 0.02 0.15 324 -0.178 0.000 Happiness 2.76 0.68 1,081 2.91 0.62 324 0.144 0.000 Depression 0.38 0.49 1,081 0.27 0.45 324 -0.110 0.000 Household size 4.97 1.97 1,081 5.06 2.09 324 0.088 0.494 Children (0–5 years) 0.64 0.79 1,081 0.83 0.90 324 0.190 0.000 Children (6–12 years) 1.07 1.05 1,081 1.05 1.10 324 -0.021 0.752 Children (13–17 years) 0.68 0.92 1,081 0.59 0.93 324 -0.087 0.114 Expenditure (Q1) 0.28 0.45 1,081 0.18 0.38 324 -0.097 0.000 Expenditure (Q2) 0.24 0.43 1,081 0.28 0.45 324 0.037 0.186 Expenditure (Q3) 0.25 0.43 1,081 0.24 0.42 324 -0.015 0.590 Expenditure (Q4) 0.23 0.42 1,081 0.31 0.46 324 0.075 0.014 Assets—index -0.10 1.53 1,081 0.35 1.53 324 0.454 0.000 Owns mobile phone 0.42 0.49 1,081 0.66 0.47 324 0.241 0.000 Instruments Group average fin. lit. 2.47 0.74 1,081 2.46 0.72 324 -0.012 0.815 Distance to nearest school 0.01 0.01 1,052 0.01 0.01 313 0.001 0.012 . Women . Men . . . . Mean . S.D. . Obs. . Mean . S.D. . Obs. . Difference . |$p$|-Value . Outcome: decisions Income: self 0.41 0.49 1,081 0.29 0.45 324 -0.124 0.000 Income: both 0.46 0.50 1,081 0.63 0.48 324 0.176 0.000 Income: involved 0.87 0.33 1,081 0.92 0.27 324 0.051 0.009 Credit: self 0.41 0.49 1,081 0.27 0.44 324 -0.145 0.000 Credit: both 0.55 0.50 1,081 0.69 0.46 324 0.144 0.000 Credit: involved 0.96 0.19 1,081 0.96 0.19 324 -0.001 0.944 Investment: self 0.42 0.49 1,054 0.27 0.45 317 -0.153 0.000 Investment: both 0.51 0.50 1,054 0.66 0.47 317 0.149 0.000 Investment: involved 0.93 0.25 1,054 0.93 0.25 317 -0.004 0.846 Financial decisions—index 2.78 0.59 1,054 2.82 0.60 317 0.042 0.346 Food: self 0.58 0.49 1,080 0.25 0.43 321 -0.333 0.000 Food: both 0.36 0.48 1,080 0.50 0.50 321 0.143 0.000 Food: involved 0.93 0.25 1,080 0.74 0.44 321 -0.190 0.000 Own health: self 0.63 0.48 1,081 0.56 0.50 324 -0.065 0.049 Own health: both 0.31 0.46 1,081 0.37 0.48 324 0.054 0.094 Own health: involved 0.94 0.23 1,081 0.93 0.25 324 -0.011 0.488 Own clothes: self 0.61 0.49 1,081 0.56 0.50 324 -0.046 0.187 Own clothes: both 0.33 0.47 1,081 0.39 0.49 324 0.055 0.100 Own clothes: involved 0.94 0.23 1,081 0.95 0.21 324 0.009 0.537 Energy: self 0.50 0.50 1,079 0.33 0.47 323 -0.171 0.000 Energy: both 0.36 0.48 1,079 0.47 0.50 323 0.106 0.002 Energy: involved 0.86 0.35 1,079 0.80 0.40 323 -0.064 0.015 Adult’s exp. decisions—index (4) 3.68 0.78 1,078 3.42 0.97 321 -0.261 0.000 Child’s health: self 0.42 0.49 881 0.12 0.33 254 -0.296 0.000 Child’s health: both 0.54 0.50 881 0.81 0.40 254 0.271 0.000 Child’s health: involved 0.95 0.21 881 0.93 0.26 254 -0.024 0.178 Child’s clothes: self 0.45 0.50 880 0.13 0.34 254 -0.318 0.000 Child’s clothes: both 0.51 0.50 880 0.77 0.42 254 0.261 0.000 Child’s clothes: involved 0.96 0.20 880 0.90 0.30 254 -0.058 0.006 All exp. decisions—index (6) 5.61 0.99 877 5.29 1.24 254 -0.315 0.001 Explanatory variables Financial literacy—index 2.36 1.22 1,081 2.83 1.13 324 0.477 0.000 Age 43.90 13.39 1,081 40.51 13.94 324 -3.388 0.001 Spell 0.57 0.50 1,081 0.72 0.45 324 0.151 0.000 Secondary education and more 0.07 0.26 1,081 0.09 0.28 324 0.015 0.411 Single 0.06 0.24 1,081 0.18 0.38 324 0.113 0.000 Married 0.67 0.47 1,081 0.77 0.42 324 0.107 0.000 Divorced 0.07 0.25 1,081 0.03 0.16 324 -0.042 0.000 Widowed 0.20 0.40 1,081 0.02 0.15 324 -0.178 0.000 Happiness 2.76 0.68 1,081 2.91 0.62 324 0.144 0.000 Depression 0.38 0.49 1,081 0.27 0.45 324 -0.110 0.000 Household size 4.97 1.97 1,081 5.06 2.09 324 0.088 0.494 Children (0–5 years) 0.64 0.79 1,081 0.83 0.90 324 0.190 0.000 Children (6–12 years) 1.07 1.05 1,081 1.05 1.10 324 -0.021 0.752 Children (13–17 years) 0.68 0.92 1,081 0.59 0.93 324 -0.087 0.114 Expenditure (Q1) 0.28 0.45 1,081 0.18 0.38 324 -0.097 0.000 Expenditure (Q2) 0.24 0.43 1,081 0.28 0.45 324 0.037 0.186 Expenditure (Q3) 0.25 0.43 1,081 0.24 0.42 324 -0.015 0.590 Expenditure (Q4) 0.23 0.42 1,081 0.31 0.46 324 0.075 0.014 Assets—index -0.10 1.53 1,081 0.35 1.53 324 0.454 0.000 Owns mobile phone 0.42 0.49 1,081 0.66 0.47 324 0.241 0.000 Instruments Group average fin. lit. 2.47 0.74 1,081 2.46 0.72 324 -0.012 0.815 Distance to nearest school 0.01 0.01 1,052 0.01 0.01 313 0.001 0.012 Notes: S.D. stands for standard deviation and Obs. stands for observations. The last column reports the |$p$|-value of an OLS regression of the listed variable on the indicator for gender with robust standard errors clustered at the VSLA level. Open in new tab Table 1 Descriptive statistics . Women . Men . . . . Mean . S.D. . Obs. . Mean . S.D. . Obs. . Difference . |$p$|-Value . Outcome: decisions Income: self 0.41 0.49 1,081 0.29 0.45 324 -0.124 0.000 Income: both 0.46 0.50 1,081 0.63 0.48 324 0.176 0.000 Income: involved 0.87 0.33 1,081 0.92 0.27 324 0.051 0.009 Credit: self 0.41 0.49 1,081 0.27 0.44 324 -0.145 0.000 Credit: both 0.55 0.50 1,081 0.69 0.46 324 0.144 0.000 Credit: involved 0.96 0.19 1,081 0.96 0.19 324 -0.001 0.944 Investment: self 0.42 0.49 1,054 0.27 0.45 317 -0.153 0.000 Investment: both 0.51 0.50 1,054 0.66 0.47 317 0.149 0.000 Investment: involved 0.93 0.25 1,054 0.93 0.25 317 -0.004 0.846 Financial decisions—index 2.78 0.59 1,054 2.82 0.60 317 0.042 0.346 Food: self 0.58 0.49 1,080 0.25 0.43 321 -0.333 0.000 Food: both 0.36 0.48 1,080 0.50 0.50 321 0.143 0.000 Food: involved 0.93 0.25 1,080 0.74 0.44 321 -0.190 0.000 Own health: self 0.63 0.48 1,081 0.56 0.50 324 -0.065 0.049 Own health: both 0.31 0.46 1,081 0.37 0.48 324 0.054 0.094 Own health: involved 0.94 0.23 1,081 0.93 0.25 324 -0.011 0.488 Own clothes: self 0.61 0.49 1,081 0.56 0.50 324 -0.046 0.187 Own clothes: both 0.33 0.47 1,081 0.39 0.49 324 0.055 0.100 Own clothes: involved 0.94 0.23 1,081 0.95 0.21 324 0.009 0.537 Energy: self 0.50 0.50 1,079 0.33 0.47 323 -0.171 0.000 Energy: both 0.36 0.48 1,079 0.47 0.50 323 0.106 0.002 Energy: involved 0.86 0.35 1,079 0.80 0.40 323 -0.064 0.015 Adult’s exp. decisions—index (4) 3.68 0.78 1,078 3.42 0.97 321 -0.261 0.000 Child’s health: self 0.42 0.49 881 0.12 0.33 254 -0.296 0.000 Child’s health: both 0.54 0.50 881 0.81 0.40 254 0.271 0.000 Child’s health: involved 0.95 0.21 881 0.93 0.26 254 -0.024 0.178 Child’s clothes: self 0.45 0.50 880 0.13 0.34 254 -0.318 0.000 Child’s clothes: both 0.51 0.50 880 0.77 0.42 254 0.261 0.000 Child’s clothes: involved 0.96 0.20 880 0.90 0.30 254 -0.058 0.006 All exp. decisions—index (6) 5.61 0.99 877 5.29 1.24 254 -0.315 0.001 Explanatory variables Financial literacy—index 2.36 1.22 1,081 2.83 1.13 324 0.477 0.000 Age 43.90 13.39 1,081 40.51 13.94 324 -3.388 0.001 Spell 0.57 0.50 1,081 0.72 0.45 324 0.151 0.000 Secondary education and more 0.07 0.26 1,081 0.09 0.28 324 0.015 0.411 Single 0.06 0.24 1,081 0.18 0.38 324 0.113 0.000 Married 0.67 0.47 1,081 0.77 0.42 324 0.107 0.000 Divorced 0.07 0.25 1,081 0.03 0.16 324 -0.042 0.000 Widowed 0.20 0.40 1,081 0.02 0.15 324 -0.178 0.000 Happiness 2.76 0.68 1,081 2.91 0.62 324 0.144 0.000 Depression 0.38 0.49 1,081 0.27 0.45 324 -0.110 0.000 Household size 4.97 1.97 1,081 5.06 2.09 324 0.088 0.494 Children (0–5 years) 0.64 0.79 1,081 0.83 0.90 324 0.190 0.000 Children (6–12 years) 1.07 1.05 1,081 1.05 1.10 324 -0.021 0.752 Children (13–17 years) 0.68 0.92 1,081 0.59 0.93 324 -0.087 0.114 Expenditure (Q1) 0.28 0.45 1,081 0.18 0.38 324 -0.097 0.000 Expenditure (Q2) 0.24 0.43 1,081 0.28 0.45 324 0.037 0.186 Expenditure (Q3) 0.25 0.43 1,081 0.24 0.42 324 -0.015 0.590 Expenditure (Q4) 0.23 0.42 1,081 0.31 0.46 324 0.075 0.014 Assets—index -0.10 1.53 1,081 0.35 1.53 324 0.454 0.000 Owns mobile phone 0.42 0.49 1,081 0.66 0.47 324 0.241 0.000 Instruments Group average fin. lit. 2.47 0.74 1,081 2.46 0.72 324 -0.012 0.815 Distance to nearest school 0.01 0.01 1,052 0.01 0.01 313 0.001 0.012 . Women . Men . . . . Mean . S.D. . Obs. . Mean . S.D. . Obs. . Difference . |$p$|-Value . Outcome: decisions Income: self 0.41 0.49 1,081 0.29 0.45 324 -0.124 0.000 Income: both 0.46 0.50 1,081 0.63 0.48 324 0.176 0.000 Income: involved 0.87 0.33 1,081 0.92 0.27 324 0.051 0.009 Credit: self 0.41 0.49 1,081 0.27 0.44 324 -0.145 0.000 Credit: both 0.55 0.50 1,081 0.69 0.46 324 0.144 0.000 Credit: involved 0.96 0.19 1,081 0.96 0.19 324 -0.001 0.944 Investment: self 0.42 0.49 1,054 0.27 0.45 317 -0.153 0.000 Investment: both 0.51 0.50 1,054 0.66 0.47 317 0.149 0.000 Investment: involved 0.93 0.25 1,054 0.93 0.25 317 -0.004 0.846 Financial decisions—index 2.78 0.59 1,054 2.82 0.60 317 0.042 0.346 Food: self 0.58 0.49 1,080 0.25 0.43 321 -0.333 0.000 Food: both 0.36 0.48 1,080 0.50 0.50 321 0.143 0.000 Food: involved 0.93 0.25 1,080 0.74 0.44 321 -0.190 0.000 Own health: self 0.63 0.48 1,081 0.56 0.50 324 -0.065 0.049 Own health: both 0.31 0.46 1,081 0.37 0.48 324 0.054 0.094 Own health: involved 0.94 0.23 1,081 0.93 0.25 324 -0.011 0.488 Own clothes: self 0.61 0.49 1,081 0.56 0.50 324 -0.046 0.187 Own clothes: both 0.33 0.47 1,081 0.39 0.49 324 0.055 0.100 Own clothes: involved 0.94 0.23 1,081 0.95 0.21 324 0.009 0.537 Energy: self 0.50 0.50 1,079 0.33 0.47 323 -0.171 0.000 Energy: both 0.36 0.48 1,079 0.47 0.50 323 0.106 0.002 Energy: involved 0.86 0.35 1,079 0.80 0.40 323 -0.064 0.015 Adult’s exp. decisions—index (4) 3.68 0.78 1,078 3.42 0.97 321 -0.261 0.000 Child’s health: self 0.42 0.49 881 0.12 0.33 254 -0.296 0.000 Child’s health: both 0.54 0.50 881 0.81 0.40 254 0.271 0.000 Child’s health: involved 0.95 0.21 881 0.93 0.26 254 -0.024 0.178 Child’s clothes: self 0.45 0.50 880 0.13 0.34 254 -0.318 0.000 Child’s clothes: both 0.51 0.50 880 0.77 0.42 254 0.261 0.000 Child’s clothes: involved 0.96 0.20 880 0.90 0.30 254 -0.058 0.006 All exp. decisions—index (6) 5.61 0.99 877 5.29 1.24 254 -0.315 0.001 Explanatory variables Financial literacy—index 2.36 1.22 1,081 2.83 1.13 324 0.477 0.000 Age 43.90 13.39 1,081 40.51 13.94 324 -3.388 0.001 Spell 0.57 0.50 1,081 0.72 0.45 324 0.151 0.000 Secondary education and more 0.07 0.26 1,081 0.09 0.28 324 0.015 0.411 Single 0.06 0.24 1,081 0.18 0.38 324 0.113 0.000 Married 0.67 0.47 1,081 0.77 0.42 324 0.107 0.000 Divorced 0.07 0.25 1,081 0.03 0.16 324 -0.042 0.000 Widowed 0.20 0.40 1,081 0.02 0.15 324 -0.178 0.000 Happiness 2.76 0.68 1,081 2.91 0.62 324 0.144 0.000 Depression 0.38 0.49 1,081 0.27 0.45 324 -0.110 0.000 Household size 4.97 1.97 1,081 5.06 2.09 324 0.088 0.494 Children (0–5 years) 0.64 0.79 1,081 0.83 0.90 324 0.190 0.000 Children (6–12 years) 1.07 1.05 1,081 1.05 1.10 324 -0.021 0.752 Children (13–17 years) 0.68 0.92 1,081 0.59 0.93 324 -0.087 0.114 Expenditure (Q1) 0.28 0.45 1,081 0.18 0.38 324 -0.097 0.000 Expenditure (Q2) 0.24 0.43 1,081 0.28 0.45 324 0.037 0.186 Expenditure (Q3) 0.25 0.43 1,081 0.24 0.42 324 -0.015 0.590 Expenditure (Q4) 0.23 0.42 1,081 0.31 0.46 324 0.075 0.014 Assets—index -0.10 1.53 1,081 0.35 1.53 324 0.454 0.000 Owns mobile phone 0.42 0.49 1,081 0.66 0.47 324 0.241 0.000 Instruments Group average fin. lit. 2.47 0.74 1,081 2.46 0.72 324 -0.012 0.815 Distance to nearest school 0.01 0.01 1,052 0.01 0.01 313 0.001 0.012 Notes: S.D. stands for standard deviation and Obs. stands for observations. The last column reports the |$p$|-value of an OLS regression of the listed variable on the indicator for gender with robust standard errors clustered at the VSLA level. Open in new tab 3.2. Intra-household decision making The first part of this article focuses on financial decision making as outcomes of interest. Outcome variables are defined as who within the household decides on income, credit, investment and expenditure decisions: the respondent, their spouse or the two together. Expenditure decisions are further divided into energy and food expenses of the household, women’s own health and clothing expenses and children’s health and clothing expenses. These indicators are similar to those included in Demographic and Health Surveys and have previously been used by Allendorf (2007) and Connelly et al. (2010) to study intra-household decision making. More recently, researchers have drawn attention to the fact that these questions oversimplify the complexity of intra-household decision making. In particular, questions should not only ask about involvement but also existing structures and roles that household members take (Bernard et al., 2020). However, for the purpose of this study, these questions suffice as we are interested in the simple action of financial decision making. We, therefore, believe that, although these questions may not capture all aspects of household financial decision making, there is still valuable information in the answers to these questions. Our aim is to test whether one reason why women are more involved in household decision making is because they have better financial literacy. Table 1 provides descriptive statistics. The majority of both sexes indicate to jointly decide on financial matters. In comparison with men, women are more likely to report that they either make the decision themselves or that their husbands make the decision alone. On the contrary, men are more likely than women to report that both partners make the decision together. Patterns are consistent for all types of financial decisions. 3.3. Financial literacy gender gap We measure financial literacy using an adjusted version of the Lusardi & Mitchell (2011) questions, which were developed further by Cole et al. (2011). This approach focuses on numeracy skills for calculating financial trade-offs. Questions are the following: |$\bullet$| Suppose you borrow RWF 10,000 from a moneylender at an interest rate of two% per month, with no repayment for three months. After three months, do you owe less than RWF 10,200, exactly RWF 10,200 or more than RWF 10,200? |$\bullet$| If you have RWF 10,000 in a savings account earning 1% interest per annum, and prices for goods and services rise 2% over a 1-year period, can you buy more than, less than or the same amount of goods in one year as you could today, with the money in the account? |$\bullet$| Is it riskier to plant multiple crops or one crop? |$\bullet$| Suppose you need to borrow RWF 50,000. Two people offer you a loan. One loan requires you to pay back RWF 60,000 in 1 month. The second loan requires you to pay back in one month RWF 50,000 plus 15% interest. Which loan represents a better deal for you? All questions were multiple choice: two questions with two possible answers and two questions with three possible answers. Respondents also had the option to answer ‘I don’t know’ or to refuse to answer. We generate an index to measure financial literacy in which respondents are given one point for each correct answer that she gives. The aggregated financial literacy index in our main regressions ranges from zero to four. In comparison to studies in countries with a similar level of development, respondents in our Rwandan sample are slightly more financially literate, for example, more literate than the Indian sample used in Cole et al. (2011). The proportion of correct answers is highest for the question on risk-diversification. Since 74% of respondents stated the agricultural sector as their main source of income, this might be obvious as the question is framed in a manner requiring agricultural knowledge. In contrast, knowledge in basic numeracy is low. Table 2 shows the distribution of financial literacy questions divided by gender. On average, women are less likely than men to provide correct answers. Only 45% of female respondents and 61% of male respondents correctly answered the borrowing decision. A total of 57% of men showed basic understanding of interest and inflation. In contrast, only 45% of women correctly dealt with these economic concepts. While 34% of men correctly answered all four questions, only 22% of women did so. In addition, women are more likely to indicate that they do not know the correct answer. As many as 26% of women indicated that they do not know the answer to the first compound interest question, whereas the proportion of men is much lower (15%). A total of 35% of women gave at least one ‘don’t know’ response to one of four financial literacy questions, the proportion of men doing so is about 19%. Overall, the financial literacy level is significantly lower for women than for men, irrespective of how financial literacy is measured. Table 2 Distribution of financial literacy questions divided by gender . Sample . Women . Men . . . . Mean . Obs. . Mean . Obs. . Mean . Obs. . Difference . |$p$|-Value . Compound interest Less than RWF 10,200 0.07 1,405 0.08 1,081 0.06 324 -0.012 0.468 Exactly RWF 10,200 0.05 1,405 0.06 1,081 0.02 324 -0.035 0.003 More than RWF 10,200 (correct) 0.64 1,405 0.61 1,081 0.77 324 0.161 0.000 Don’t know 0.23 1,405 0.26 1,081 0.15 324 -0.114 0.000 Inflation Less (correct) 0.59 1,405 0.56 1,081 0.67 324 0.104 0.001 Same 0.04 1,405 0.04 1,081 0.04 324 0.003 0.778 More 0.18 1,405 0.18 1,081 0.19 324 0.016 0.545 Don’t know 0.20 1,405 0.22 1,081 0.10 324 -0.123 0.000 Risk diversification One crop 0.24 1,405 0.25 1,081 0.20 324 -0.054 0.059 Multiple crops (correct) 0.75 1,405 0.73 1,081 0.79 324 0.054 0.056 Don’t know 0.01 1,405 0.01 1,081 0.01 324 -0.001 0.932 Borrowing decision RWF 60,000 0.32 1,405 0.33 1,081 0.28 324 -0.048 0.117 RWF 50,000 + 15% (correct) 0.49 1,405 0.45 1,081 0.61 324 0.158 0.000 Don’t know 0.19 1,405 0.22 1,081 0.11 324 -0.109 0.000 Cross-question consistency Wrong: interest and inflation 0.52 1,405 0.55 1,081 0.43 324 -0.126 0.000 Correct: interest and inflation 0.48 1,405 0.45 1,081 0.57 324 0.126 0.000 General indicators All questions correct 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 Zero correct answers 0.06 1,405 0.07 1,081 0.03 324 -0.035 0.003 At least one don’t know 0.31 1,405 0.35 1,081 0.19 324 -0.162 0.000 All don’t know 0.00 1,405 0.01 1,081 0.00 324 -0.006 0.014 Aggregated indices FL index 2.47 1,405 2.36 1,081 2.83 324 0.477 0.000 FL factor score 0.00 1,405 -0.11 1,081 0.36 324 0.474 0.000 FL dummy 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 . Sample . Women . Men . . . . Mean . Obs. . Mean . Obs. . Mean . Obs. . Difference . |$p$|-Value . Compound interest Less than RWF 10,200 0.07 1,405 0.08 1,081 0.06 324 -0.012 0.468 Exactly RWF 10,200 0.05 1,405 0.06 1,081 0.02 324 -0.035 0.003 More than RWF 10,200 (correct) 0.64 1,405 0.61 1,081 0.77 324 0.161 0.000 Don’t know 0.23 1,405 0.26 1,081 0.15 324 -0.114 0.000 Inflation Less (correct) 0.59 1,405 0.56 1,081 0.67 324 0.104 0.001 Same 0.04 1,405 0.04 1,081 0.04 324 0.003 0.778 More 0.18 1,405 0.18 1,081 0.19 324 0.016 0.545 Don’t know 0.20 1,405 0.22 1,081 0.10 324 -0.123 0.000 Risk diversification One crop 0.24 1,405 0.25 1,081 0.20 324 -0.054 0.059 Multiple crops (correct) 0.75 1,405 0.73 1,081 0.79 324 0.054 0.056 Don’t know 0.01 1,405 0.01 1,081 0.01 324 -0.001 0.932 Borrowing decision RWF 60,000 0.32 1,405 0.33 1,081 0.28 324 -0.048 0.117 RWF 50,000 + 15% (correct) 0.49 1,405 0.45 1,081 0.61 324 0.158 0.000 Don’t know 0.19 1,405 0.22 1,081 0.11 324 -0.109 0.000 Cross-question consistency Wrong: interest and inflation 0.52 1,405 0.55 1,081 0.43 324 -0.126 0.000 Correct: interest and inflation 0.48 1,405 0.45 1,081 0.57 324 0.126 0.000 General indicators All questions correct 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 Zero correct answers 0.06 1,405 0.07 1,081 0.03 324 -0.035 0.003 At least one don’t know 0.31 1,405 0.35 1,081 0.19 324 -0.162 0.000 All don’t know 0.00 1,405 0.01 1,081 0.00 324 -0.006 0.014 Aggregated indices FL index 2.47 1,405 2.36 1,081 2.83 324 0.477 0.000 FL factor score 0.00 1,405 -0.11 1,081 0.36 324 0.474 0.000 FL dummy 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 Notes: Obs. stands for observations. The last column reports the |$p$|-value of an OLS regression of the listed variable on the indicator for gender with robust standard errors clustered at the VSLA level. Open in new tab Table 2 Distribution of financial literacy questions divided by gender . Sample . Women . Men . . . . Mean . Obs. . Mean . Obs. . Mean . Obs. . Difference . |$p$|-Value . Compound interest Less than RWF 10,200 0.07 1,405 0.08 1,081 0.06 324 -0.012 0.468 Exactly RWF 10,200 0.05 1,405 0.06 1,081 0.02 324 -0.035 0.003 More than RWF 10,200 (correct) 0.64 1,405 0.61 1,081 0.77 324 0.161 0.000 Don’t know 0.23 1,405 0.26 1,081 0.15 324 -0.114 0.000 Inflation Less (correct) 0.59 1,405 0.56 1,081 0.67 324 0.104 0.001 Same 0.04 1,405 0.04 1,081 0.04 324 0.003 0.778 More 0.18 1,405 0.18 1,081 0.19 324 0.016 0.545 Don’t know 0.20 1,405 0.22 1,081 0.10 324 -0.123 0.000 Risk diversification One crop 0.24 1,405 0.25 1,081 0.20 324 -0.054 0.059 Multiple crops (correct) 0.75 1,405 0.73 1,081 0.79 324 0.054 0.056 Don’t know 0.01 1,405 0.01 1,081 0.01 324 -0.001 0.932 Borrowing decision RWF 60,000 0.32 1,405 0.33 1,081 0.28 324 -0.048 0.117 RWF 50,000 + 15% (correct) 0.49 1,405 0.45 1,081 0.61 324 0.158 0.000 Don’t know 0.19 1,405 0.22 1,081 0.11 324 -0.109 0.000 Cross-question consistency Wrong: interest and inflation 0.52 1,405 0.55 1,081 0.43 324 -0.126 0.000 Correct: interest and inflation 0.48 1,405 0.45 1,081 0.57 324 0.126 0.000 General indicators All questions correct 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 Zero correct answers 0.06 1,405 0.07 1,081 0.03 324 -0.035 0.003 At least one don’t know 0.31 1,405 0.35 1,081 0.19 324 -0.162 0.000 All don’t know 0.00 1,405 0.01 1,081 0.00 324 -0.006 0.014 Aggregated indices FL index 2.47 1,405 2.36 1,081 2.83 324 0.477 0.000 FL factor score 0.00 1,405 -0.11 1,081 0.36 324 0.474 0.000 FL dummy 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 . Sample . Women . Men . . . . Mean . Obs. . Mean . Obs. . Mean . Obs. . Difference . |$p$|-Value . Compound interest Less than RWF 10,200 0.07 1,405 0.08 1,081 0.06 324 -0.012 0.468 Exactly RWF 10,200 0.05 1,405 0.06 1,081 0.02 324 -0.035 0.003 More than RWF 10,200 (correct) 0.64 1,405 0.61 1,081 0.77 324 0.161 0.000 Don’t know 0.23 1,405 0.26 1,081 0.15 324 -0.114 0.000 Inflation Less (correct) 0.59 1,405 0.56 1,081 0.67 324 0.104 0.001 Same 0.04 1,405 0.04 1,081 0.04 324 0.003 0.778 More 0.18 1,405 0.18 1,081 0.19 324 0.016 0.545 Don’t know 0.20 1,405 0.22 1,081 0.10 324 -0.123 0.000 Risk diversification One crop 0.24 1,405 0.25 1,081 0.20 324 -0.054 0.059 Multiple crops (correct) 0.75 1,405 0.73 1,081 0.79 324 0.054 0.056 Don’t know 0.01 1,405 0.01 1,081 0.01 324 -0.001 0.932 Borrowing decision RWF 60,000 0.32 1,405 0.33 1,081 0.28 324 -0.048 0.117 RWF 50,000 + 15% (correct) 0.49 1,405 0.45 1,081 0.61 324 0.158 0.000 Don’t know 0.19 1,405 0.22 1,081 0.11 324 -0.109 0.000 Cross-question consistency Wrong: interest and inflation 0.52 1,405 0.55 1,081 0.43 324 -0.126 0.000 Correct: interest and inflation 0.48 1,405 0.45 1,081 0.57 324 0.126 0.000 General indicators All questions correct 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 Zero correct answers 0.06 1,405 0.07 1,081 0.03 324 -0.035 0.003 At least one don’t know 0.31 1,405 0.35 1,081 0.19 324 -0.162 0.000 All don’t know 0.00 1,405 0.01 1,081 0.00 324 -0.006 0.014 Aggregated indices FL index 2.47 1,405 2.36 1,081 2.83 324 0.477 0.000 FL factor score 0.00 1,405 -0.11 1,081 0.36 324 0.474 0.000 FL dummy 0.24 1,405 0.22 1,081 0.34 324 0.123 0.000 Notes: Obs. stands for observations. The last column reports the |$p$|-value of an OLS regression of the listed variable on the indicator for gender with robust standard errors clustered at the VSLA level. Open in new tab Our findings confirm results found in other studies on financial literacy and gender, where women are more likely to say that they do not know the answer and perform worse than men (Bucher-Koenen et al., 2017). This is true even for the most educated women (Mahdavi & Horton, 2014). So far, only very little evidence exists on the reasons behind this gender gap. Grohmann et al. (2016) argue that the gender gap is caused by culture and that financial literacy is similar between sexes in Thailand because Thai women are traditionally in charge of financial matters. Likewise, Filipiak & Walle (2015) find that women in matrilineal societies in India have better financial literacy than women living in patrilineal societies. Hsu (2016) attributes women’s lower financial literacy to specialization of tasks within the household. 4. Financial literacy and decision making 4.1. OLS analysis To examine the link between financial literacy and intra-household financial decision making, we first estimate a simple linear probability model. We regress the financial literacy index, |$FL$|, described above on three binary outcomes and three indices, |$DM$|. Binary outcome variables equal one if a woman participates in income, credit and/or investment decisions at home. The first index aggregates these indicators to a ‘Financial decisions—index’. The ‘Adult’s expenditure decisions—index’ is defined by women’s participation in food, own health, own clothes and energy decisions. The third index adds two decision-making categories related to children’s health and clothes to an ‘All expenditure decisions—index’.9 Indices comprise the sum of decisions women are involved in. The estimation equation is given as follows:10 $$\begin{align}& DM=\alpha+\beta FL+\gamma X+u,\end{align}$$(1) where |$X$| denotes a set of control variables that have been shown to be correlated with someone’s financial literacy level, such as age, whether the person can read and write and marital status with being single acting as the excluded category (Lusardi & Mitchell, 2014). We further control for the number of household members and the number of children in different age groups11. Four expenditure quartile dummies proxy for income because it is commonly hard to measure in developing countries. The lowest expenditure quartile is excluded from the regression. An asset index using the first principal of a principal component analysis additionally controls for household wealth. |$u$| is the equation specific error term and standard errors are clustered at the VSLA level. Results are shown in Table 3. In columns one to three, we display outcome variables that are unity if a woman is involved in a financial decision. This means that she reports that she either makes the decision alone or that she and her husband jointly decide. We focus on this outcome variable as it is woman’s involvement in financial decision making that is argued to have positive effects on household welfare, rather than women taking decisions alone (Duflo, 2012). Nevertheless, we also look at sole decision making in the Appendix.12 In columns four to six, we report three decisions indices: one for finances, one for adult’s expenditures and one for all expenditures. Results for individual expenditure decisions are shown in the Appendix.13 Table 3 OLS results of financial literacy on decision making . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.024|$^{***}$| 0.014|$^{***}$| 0.020|$^{***}$| 0.056|$^{***}$| 0.073|$^{***}$| 0.089|$^{***}$| (0.009) (0.005) (0.007) (0.017) (0.024) (0.034) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.038|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.021) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000|$^{*}$| (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.027 0.004 0.015 0.030 0.066 0.113 (0.025) (0.016) (0.018) (0.048) (0.061) (0.091) Married -0.040 -0.017 -0.114|$^{***}$| -0.178|$^{**}$| 0.073 -0.105 (0.041) (0.027) (0.034) (0.079) (0.144) (0.226) Divorced 0.102|$^{**}$| -0.003 -0.073|$^{**}$| 0.014 0.342|$^{**}$| 0.201 (0.042) (0.032) (0.036) (0.087) (0.147) (0.242) Widowed 0.063 0.006 -0.082|$^{**}$| -0.020 0.308|$^{**}$| 0.175 (0.046) (0.027) (0.035) (0.083) (0.142) (0.223) Household size -0.025|$^{***}$| -0.007 -0.016|$^{**}$| -0.047|$^{***}$| -0.024 -0.051 (0.009) (0.005) (0.006) (0.015) (0.017) (0.038) Children (0-5 years) 0.029|$^{*}$| 0.012 0.020 0.059|$^{**}$| 0.041 (0.017) (0.010) (0.014) (0.029) (0.046) Children (6-12 years) 0.018 -0.004 0.023|$^{**}$| 0.039 0.064 (0.012) (0.008) (0.010) (0.024) (0.040) Children (13-17 years) 0.012 0.005 0.011 0.028 0.042 (0.013) (0.006) (0.009) (0.020) (0.037) Expenditure (Q2) 0.044|$^{*}$| 0.006 0.020 0.062 0.077 0.111 (0.025) (0.017) (0.021) (0.051) (0.071) (0.106) Expenditure (Q3) 0.014 -0.007 0.001 0.001 0.027 0.003 (0.031) (0.018) (0.023) (0.058) (0.072) (0.112) Expenditure (Q4) 0.042 0.005 -0.016 0.030 0.037 0.107 (0.030) (0.018) (0.025) (0.056) (0.072) (0.105) Assets—index -0.007 0.001 0.004 -0.003 -0.019 -0.028 (0.008) (0.004) (0.006) (0.014) (0.017) (0.024) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.069 0.027 0.045 0.066 0.087 0.062 . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.024|$^{***}$| 0.014|$^{***}$| 0.020|$^{***}$| 0.056|$^{***}$| 0.073|$^{***}$| 0.089|$^{***}$| (0.009) (0.005) (0.007) (0.017) (0.024) (0.034) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.038|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.021) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000|$^{*}$| (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.027 0.004 0.015 0.030 0.066 0.113 (0.025) (0.016) (0.018) (0.048) (0.061) (0.091) Married -0.040 -0.017 -0.114|$^{***}$| -0.178|$^{**}$| 0.073 -0.105 (0.041) (0.027) (0.034) (0.079) (0.144) (0.226) Divorced 0.102|$^{**}$| -0.003 -0.073|$^{**}$| 0.014 0.342|$^{**}$| 0.201 (0.042) (0.032) (0.036) (0.087) (0.147) (0.242) Widowed 0.063 0.006 -0.082|$^{**}$| -0.020 0.308|$^{**}$| 0.175 (0.046) (0.027) (0.035) (0.083) (0.142) (0.223) Household size -0.025|$^{***}$| -0.007 -0.016|$^{**}$| -0.047|$^{***}$| -0.024 -0.051 (0.009) (0.005) (0.006) (0.015) (0.017) (0.038) Children (0-5 years) 0.029|$^{*}$| 0.012 0.020 0.059|$^{**}$| 0.041 (0.017) (0.010) (0.014) (0.029) (0.046) Children (6-12 years) 0.018 -0.004 0.023|$^{**}$| 0.039 0.064 (0.012) (0.008) (0.010) (0.024) (0.040) Children (13-17 years) 0.012 0.005 0.011 0.028 0.042 (0.013) (0.006) (0.009) (0.020) (0.037) Expenditure (Q2) 0.044|$^{*}$| 0.006 0.020 0.062 0.077 0.111 (0.025) (0.017) (0.021) (0.051) (0.071) (0.106) Expenditure (Q3) 0.014 -0.007 0.001 0.001 0.027 0.003 (0.031) (0.018) (0.023) (0.058) (0.072) (0.112) Expenditure (Q4) 0.042 0.005 -0.016 0.030 0.037 0.107 (0.030) (0.018) (0.025) (0.056) (0.072) (0.105) Assets—index -0.007 0.001 0.004 -0.003 -0.019 -0.028 (0.008) (0.004) (0.006) (0.014) (0.017) (0.024) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.069 0.027 0.045 0.066 0.087 0.062 Notes: The table reports coefficients of multivariate regressions with standard errors clustered at VSLA level in brackets. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. Outcome variables in columns 1–3 are indicator variables equal to one if a woman has decision-making power in that category. The index in column 4 counts the number of financial decisions (columns 1–3) women are involved in. The index in column 5 counts four expenditure decisions that are unrelated to children. The index in column 6 includes two additional questions related to children’s health and clothes. Open in new tab Table 3 OLS results of financial literacy on decision making . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.024|$^{***}$| 0.014|$^{***}$| 0.020|$^{***}$| 0.056|$^{***}$| 0.073|$^{***}$| 0.089|$^{***}$| (0.009) (0.005) (0.007) (0.017) (0.024) (0.034) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.038|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.021) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000|$^{*}$| (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.027 0.004 0.015 0.030 0.066 0.113 (0.025) (0.016) (0.018) (0.048) (0.061) (0.091) Married -0.040 -0.017 -0.114|$^{***}$| -0.178|$^{**}$| 0.073 -0.105 (0.041) (0.027) (0.034) (0.079) (0.144) (0.226) Divorced 0.102|$^{**}$| -0.003 -0.073|$^{**}$| 0.014 0.342|$^{**}$| 0.201 (0.042) (0.032) (0.036) (0.087) (0.147) (0.242) Widowed 0.063 0.006 -0.082|$^{**}$| -0.020 0.308|$^{**}$| 0.175 (0.046) (0.027) (0.035) (0.083) (0.142) (0.223) Household size -0.025|$^{***}$| -0.007 -0.016|$^{**}$| -0.047|$^{***}$| -0.024 -0.051 (0.009) (0.005) (0.006) (0.015) (0.017) (0.038) Children (0-5 years) 0.029|$^{*}$| 0.012 0.020 0.059|$^{**}$| 0.041 (0.017) (0.010) (0.014) (0.029) (0.046) Children (6-12 years) 0.018 -0.004 0.023|$^{**}$| 0.039 0.064 (0.012) (0.008) (0.010) (0.024) (0.040) Children (13-17 years) 0.012 0.005 0.011 0.028 0.042 (0.013) (0.006) (0.009) (0.020) (0.037) Expenditure (Q2) 0.044|$^{*}$| 0.006 0.020 0.062 0.077 0.111 (0.025) (0.017) (0.021) (0.051) (0.071) (0.106) Expenditure (Q3) 0.014 -0.007 0.001 0.001 0.027 0.003 (0.031) (0.018) (0.023) (0.058) (0.072) (0.112) Expenditure (Q4) 0.042 0.005 -0.016 0.030 0.037 0.107 (0.030) (0.018) (0.025) (0.056) (0.072) (0.105) Assets—index -0.007 0.001 0.004 -0.003 -0.019 -0.028 (0.008) (0.004) (0.006) (0.014) (0.017) (0.024) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.069 0.027 0.045 0.066 0.087 0.062 . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.024|$^{***}$| 0.014|$^{***}$| 0.020|$^{***}$| 0.056|$^{***}$| 0.073|$^{***}$| 0.089|$^{***}$| (0.009) (0.005) (0.007) (0.017) (0.024) (0.034) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.038|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.021) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000|$^{*}$| (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.027 0.004 0.015 0.030 0.066 0.113 (0.025) (0.016) (0.018) (0.048) (0.061) (0.091) Married -0.040 -0.017 -0.114|$^{***}$| -0.178|$^{**}$| 0.073 -0.105 (0.041) (0.027) (0.034) (0.079) (0.144) (0.226) Divorced 0.102|$^{**}$| -0.003 -0.073|$^{**}$| 0.014 0.342|$^{**}$| 0.201 (0.042) (0.032) (0.036) (0.087) (0.147) (0.242) Widowed 0.063 0.006 -0.082|$^{**}$| -0.020 0.308|$^{**}$| 0.175 (0.046) (0.027) (0.035) (0.083) (0.142) (0.223) Household size -0.025|$^{***}$| -0.007 -0.016|$^{**}$| -0.047|$^{***}$| -0.024 -0.051 (0.009) (0.005) (0.006) (0.015) (0.017) (0.038) Children (0-5 years) 0.029|$^{*}$| 0.012 0.020 0.059|$^{**}$| 0.041 (0.017) (0.010) (0.014) (0.029) (0.046) Children (6-12 years) 0.018 -0.004 0.023|$^{**}$| 0.039 0.064 (0.012) (0.008) (0.010) (0.024) (0.040) Children (13-17 years) 0.012 0.005 0.011 0.028 0.042 (0.013) (0.006) (0.009) (0.020) (0.037) Expenditure (Q2) 0.044|$^{*}$| 0.006 0.020 0.062 0.077 0.111 (0.025) (0.017) (0.021) (0.051) (0.071) (0.106) Expenditure (Q3) 0.014 -0.007 0.001 0.001 0.027 0.003 (0.031) (0.018) (0.023) (0.058) (0.072) (0.112) Expenditure (Q4) 0.042 0.005 -0.016 0.030 0.037 0.107 (0.030) (0.018) (0.025) (0.056) (0.072) (0.105) Assets—index -0.007 0.001 0.004 -0.003 -0.019 -0.028 (0.008) (0.004) (0.006) (0.014) (0.017) (0.024) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.069 0.027 0.045 0.066 0.087 0.062 Notes: The table reports coefficients of multivariate regressions with standard errors clustered at VSLA level in brackets. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. Outcome variables in columns 1–3 are indicator variables equal to one if a woman has decision-making power in that category. The index in column 4 counts the number of financial decisions (columns 1–3) women are involved in. The index in column 5 counts four expenditure decisions that are unrelated to children. The index in column 6 includes two additional questions related to children’s health and clothes. Open in new tab Results in Table 3 show that there is a clear and strong correlation between financial literacy and women’s involvement in household decision making. This relationship is positive and economically significant for individual financial decisions and all indices. It is significant despite the large number of control variables being considered. Apart from financial literacy, other expected patterns can be observed. Older women tend to be more involved in intra-household decision making, albeit this relationship is non-linear. Women who are married, divorced or widowed are less likely to be involved in household decisions than women who are single. There is also a negative relationship between a woman having decision-making power within the household and the size of that household in which she lives. Interestingly, women’s decision-making power is not significantly associated with income and wealth. The most striking result to emerge from the data is, however, the correlation between financial literacy and women’s involvement in household decision making. The effect of being able to correctly answer one more financial literacy question is, for example, larger than the effect of being one year older. In the next section, we employ IV regressions to establish whether this finding is causal. 4.2. IV analysis The cross-sectional design of this study poses potential endogeneity problems regarding the link between financial literacy and intra-household decision making due to omitted variable bias or reverse causality. For example, unobservable personal attributes could drive financial literacy and intra-household decision making at the same time. Similarly, it is possible that reverse causality is at play and that decision makers use their greater agency to learn about financial matters. Of course, better financial literacy might then further enhance involvement in household decisions. To support our argumentation for a positive causal relationship running from financial literacy to more involvement in intra-household decision making by women; we, first, draw on theoretical backing. In line with the concept by Kabeer (1999), financial literacy can be thought of as a resource that affects decision making. Second, we employ a two stage least square (2SLS) approach—an approach that is commonly used to resolve issues of endogeneity regarding the effect of financial literacy on financial behaviour. Examples are the financial situation of respondents’ siblings (van Rooij et al., 2011) or the amount of respondents’ education dedicated to economics as an instrument for their own financial literacy (van Rooij et al., 2012). We collected a number of potential instruments such as whether parents taught their children how to budget, the proportion of people in a district who report the nearest bank to be less than 30 minutes away, the proportion of people who report the nearest market to be less than 30 minutes away and the quality of public transport. Yet, none of these potential instruments pass the standard tests for weak instruments. Instead, our identification strategy is based on the VSLA’s average financial literacy index excluding the person who is examined. This instrument is highly correlated with the financial literacy of that person as group members are likely to benefit from each others financial knowledge. The regression of the instrument on a woman’s financial literacy index is the first stage. It uses the exogenous instrument to predict the endogenous variable: $$\begin{align}& FL=\eta+\rho Z+\mu X+v\end{align}$$(2) where |$FL$| is financial literacy and |$Z$| is the average financial literacy of other group members. The control variables, |$X$|, are the same as for previous OLS regressions. Results of that first stage are shown in Table A9 in the Appendix. In the second stage, the predicted values of the first stage |$\widehat{FL}$| are used as regressors to replace the endogenous variable: $$\begin{align}& DM=\alpha_{IV}+\beta_{IV} \widehat{FL}+\gamma_{IV} X+u_{IV}.\end{align}$$(3) The IV regression results of the second stage, as shown in Table 4, indicate similar patterns as simple OLS regression analyses in Table 3.14 Financial literacy has a significantly positive effect on women’s involvement in intra-household financial decision making. This holds for the three financial decisions and for the three indices. Interestingly, the size of the coefficient is similar between OLS and IV models, which is unusual in the financial literacy literature. Our identification strategy assumes that after adding all demographic, household and wealth controls, the average financial literacy of other group members has no direct impact on a woman’s decision-making power at home. This also applies when the individual woman influences her peers. She may affect financial literacy of the group but should not directly influence decision making in other group members’ households. We discuss and examine potential threats to this identification strategy here and in the robustness section. First, it is possible that women in groups with higher financial literacy are also more likely to influence financial decision making within the household directly without improving financial literacy of the individual woman. Although this cannot be tested, it is unlikely for theoretical reasons and because evidence suggests otherwise. Following conventional intra-household decision-making theory, each household member’s contribution to the household determines decision-making power at home (McElroy & Horney, 1981). Hence, we argue that intra-household decision making is a private process determined by the members of that household. Despite theory suggesting that the exclusion restriction holds, we further validate our IV identification strategy. One concern may arise if financial literacy levels vary with the location people live in. This is especially worrisome when some savings groups are on average more financially literate than others because they live in more progressive areas where women are also more involved in household decisions. Mapping the study groups, however, mitigates this concern because the variation in VSLA average financial literacy is not systematic between rural and more urban areas (see Figure A1). Table A11 in the Appendix further shows no significant correlations between group financial literacy levels and distances to urban spots such as markets or health centres. Another concern to identification is homophily. It is imaginable that people who are similar and have similar financial literacy levels choose to form a group together. However, the structure of VSLAs in Rwanda makes this very unlikely. In most cases, each village has only one savings group that is set up by a village agent (VA).15 In our data, the average distance to the nearest savings group is about 1.6 km linear distance. Southern Rwanda is very mountainous; the roads are often in bad conditions and so travel takes a long time. Therefore, it would be extremely difficult to travel even to the next village for weekly group meetings. Moreover, groups with high financial literacy and groups with low financial literacy may differ in other ways, which then influences their financial decision making. In Table A12, we run |$t$|-tests between groups with financial literacy above the mean and groups with financial literacy below the mean and see that there are almost no significant differences in observables. We also show that the coefficient on the IV is not sensitive to the inclusion of covariates (see Table A13). This indicates little effect of unobserved variable bias, which would further threaten the exclusion restriction. Another potential challenge occurs when the group is only more financially literate because the instrumented individual is more financially literate. We mitigate this concern by instrumenting financial literacy of the woman with the average financial literacy of all group members that are served by the same VA. The VA is someone who teaches the group about the VSLA concept, including saving and borrowing mechanisms, and recruits its members. She operates across villages. As a consequence, members of these groups do not necessarily know each other but are taught about financial literacy by the same person. Estimates in the second panel of Table A14 confirm main results in the first panel, which supports our claim that the group level influences the individual level. The exclusion restriction may also be violated if group members not only learn about finances but also about intra-household decision making or exert some sort of social pressure on each other. To examine the robustness of our results to this possibility, the third panel in Table A14 takes the group financial literacy average for male members only and uses this as an instrument for woman’s own financial literacy level. Estimates remain positive, albeit only statistically significant for investment decisions at home. This might be caused by a small sample because male and female members were randomly selected to enter our sample and men are typically underrepresented in these groups. Another way to mitigate this concern is to identify a sample for whom learning about finances is difficult. The small number of elderly, who are equal or above the 90th age percentile, provides such a sample because the ability to learn new things might decrease with age. In panel four, we see a change in significance, size and sign. Further evidence against group members learning about intra-household decision making from each other comes from splitting the sample into groups that were formed before or after spring 2012, which is the median date of group formation. The first stage is shown in Table A15 and the second stage in Table A16. According to meta-study on financial education by Kaiser & Menkhoff (2017), financial literacy is easier and faster to change than financial decision making. We, therefore, hypothesize that if learning about financial decision making from each other is the channel through which our instrument works, we should see a stronger relationship between financial literacy and financial decision making in the second stage for older groups. This is not the case. Another concern could be other group specific characteristics, apart from the location, that influence both financial literacy and decision-making processes within the household. Table A17 in the Appendix, therefore, shows IV results with group fixed effects. These findings establish similar relationships between financial literacy and intra-household decision making, albeit only statistically significant for two out of three indices. As a consequence, we rule out unobserved variables or selection into groups that are better financially educated as drivers of our results. We believe that the only channel through which the instrument affects intra-household decision making is via the financial literacy of the individual. We, therefore, argue that there is a causal relationship and that higher financial literacy strengthens women’s intra-household decision making. Table 4 IV results of financial literacy on decision making . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.065|$^{**}$| 0.038|$^{**}$| 0.075|$^{***}$| 0.172|$^{***}$| 0.286|$^{***}$| 0.400|$^{***}$| (0.033) (0.018) (0.029) (0.065) (0.093) (0.126) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.037|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.023) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.002 -0.010 -0.019 -0.042 -0.061 -0.076 (0.029) (0.017) (0.026) (0.058) (0.082) (0.119) Married -0.033 -0.013 -0.101|$^{***}$| -0.151|$^{*}$| 0.094 -0.075 (0.041) (0.027) (0.035) (0.080) (0.144) (0.236) Divorced 0.114|$^{***}$| 0.004 -0.053 0.055 0.387|$^{**}$| 0.267 (0.044) (0.031) (0.036) (0.090) (0.151) (0.257) Widowed 0.075 0.013 -0.062|$^{*}$| 0.023 0.354|$^{**}$| 0.258 (0.048) (0.027) (0.035) (0.088) (0.149) (0.243) Household size -0.025|$^{***}$| -0.007 -0.015|$^{**}$| -0.046|$^{***}$| -0.030|$^{*}$| -0.058 (0.009) (0.005) (0.006) (0.014) (0.017) (0.040) Children (0–5 years) 0.027 0.011 0.017 0.053|$^{*}$| 0.032 (0.018) (0.010) (0.014) (0.029) (0.053) Children (6–12 years) 0.016 -0.005 0.020|$^{*}$| 0.031 0.055 (0.012) (0.008) (0.010) (0.025) (0.044) Children (13–17 years) 0.011 0.004 0.009 0.023 0.037 (0.013) (0.006) (0.009) (0.020) (0.043) Expenditure (Q2) 0.038 0.002 0.011 0.043 0.044 0.061 (0.025) (0.017) (0.022) (0.053) (0.068) (0.106) Expenditure (Q3) -0.002 -0.016 -0.020 -0.043 -0.053 -0.106 (0.032) (0.020) (0.028) (0.063) (0.075) (0.120) Expenditure (Q4) 0.031 -0.001 -0.030 0.000 -0.018 0.026 (0.032) (0.019) (0.026) (0.059) (0.071) (0.107) Assets—index -0.007 0.000 0.003 -0.004 -0.022 -0.027 (0.008) (0.004) (0.006) (0.014) (0.017) (0.025) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.051 0.007 -0.018 0.018 -0.008 -0.065 |$F$|-statistic 58.262 58.262 54.482 54.482 60.569 52.894 . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.065|$^{**}$| 0.038|$^{**}$| 0.075|$^{***}$| 0.172|$^{***}$| 0.286|$^{***}$| 0.400|$^{***}$| (0.033) (0.018) (0.029) (0.065) (0.093) (0.126) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.037|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.023) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.002 -0.010 -0.019 -0.042 -0.061 -0.076 (0.029) (0.017) (0.026) (0.058) (0.082) (0.119) Married -0.033 -0.013 -0.101|$^{***}$| -0.151|$^{*}$| 0.094 -0.075 (0.041) (0.027) (0.035) (0.080) (0.144) (0.236) Divorced 0.114|$^{***}$| 0.004 -0.053 0.055 0.387|$^{**}$| 0.267 (0.044) (0.031) (0.036) (0.090) (0.151) (0.257) Widowed 0.075 0.013 -0.062|$^{*}$| 0.023 0.354|$^{**}$| 0.258 (0.048) (0.027) (0.035) (0.088) (0.149) (0.243) Household size -0.025|$^{***}$| -0.007 -0.015|$^{**}$| -0.046|$^{***}$| -0.030|$^{*}$| -0.058 (0.009) (0.005) (0.006) (0.014) (0.017) (0.040) Children (0–5 years) 0.027 0.011 0.017 0.053|$^{*}$| 0.032 (0.018) (0.010) (0.014) (0.029) (0.053) Children (6–12 years) 0.016 -0.005 0.020|$^{*}$| 0.031 0.055 (0.012) (0.008) (0.010) (0.025) (0.044) Children (13–17 years) 0.011 0.004 0.009 0.023 0.037 (0.013) (0.006) (0.009) (0.020) (0.043) Expenditure (Q2) 0.038 0.002 0.011 0.043 0.044 0.061 (0.025) (0.017) (0.022) (0.053) (0.068) (0.106) Expenditure (Q3) -0.002 -0.016 -0.020 -0.043 -0.053 -0.106 (0.032) (0.020) (0.028) (0.063) (0.075) (0.120) Expenditure (Q4) 0.031 -0.001 -0.030 0.000 -0.018 0.026 (0.032) (0.019) (0.026) (0.059) (0.071) (0.107) Assets—index -0.007 0.000 0.003 -0.004 -0.022 -0.027 (0.008) (0.004) (0.006) (0.014) (0.017) (0.025) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.051 0.007 -0.018 0.018 -0.008 -0.065 |$F$|-statistic 58.262 58.262 54.482 54.482 60.569 52.894 Notes: The table reports coefficients of IV regressions with standard errors clustered at VSLA level in brackets. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. Outcome variables in columns 1–3 are indicator variables equal to one if a woman has decision-making power in that category. The index in column 4 counts the number of financial decisions (columns 1–3) women are involved in. The index in column 5 counts four expenditure decisions that are unrelated to children. The index in column 6 includes two additional questions related to children’s health and clothes. The instrument used is the average group index of financial literacy excluding the individual considered. Open in new tab Table 4 IV results of financial literacy on decision making . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.065|$^{**}$| 0.038|$^{**}$| 0.075|$^{***}$| 0.172|$^{***}$| 0.286|$^{***}$| 0.400|$^{***}$| (0.033) (0.018) (0.029) (0.065) (0.093) (0.126) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.037|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.023) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.002 -0.010 -0.019 -0.042 -0.061 -0.076 (0.029) (0.017) (0.026) (0.058) (0.082) (0.119) Married -0.033 -0.013 -0.101|$^{***}$| -0.151|$^{*}$| 0.094 -0.075 (0.041) (0.027) (0.035) (0.080) (0.144) (0.236) Divorced 0.114|$^{***}$| 0.004 -0.053 0.055 0.387|$^{**}$| 0.267 (0.044) (0.031) (0.036) (0.090) (0.151) (0.257) Widowed 0.075 0.013 -0.062|$^{*}$| 0.023 0.354|$^{**}$| 0.258 (0.048) (0.027) (0.035) (0.088) (0.149) (0.243) Household size -0.025|$^{***}$| -0.007 -0.015|$^{**}$| -0.046|$^{***}$| -0.030|$^{*}$| -0.058 (0.009) (0.005) (0.006) (0.014) (0.017) (0.040) Children (0–5 years) 0.027 0.011 0.017 0.053|$^{*}$| 0.032 (0.018) (0.010) (0.014) (0.029) (0.053) Children (6–12 years) 0.016 -0.005 0.020|$^{*}$| 0.031 0.055 (0.012) (0.008) (0.010) (0.025) (0.044) Children (13–17 years) 0.011 0.004 0.009 0.023 0.037 (0.013) (0.006) (0.009) (0.020) (0.043) Expenditure (Q2) 0.038 0.002 0.011 0.043 0.044 0.061 (0.025) (0.017) (0.022) (0.053) (0.068) (0.106) Expenditure (Q3) -0.002 -0.016 -0.020 -0.043 -0.053 -0.106 (0.032) (0.020) (0.028) (0.063) (0.075) (0.120) Expenditure (Q4) 0.031 -0.001 -0.030 0.000 -0.018 0.026 (0.032) (0.019) (0.026) (0.059) (0.071) (0.107) Assets—index -0.007 0.000 0.003 -0.004 -0.022 -0.027 (0.008) (0.004) (0.006) (0.014) (0.017) (0.025) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.051 0.007 -0.018 0.018 -0.008 -0.065 |$F$|-statistic 58.262 58.262 54.482 54.482 60.569 52.894 . (1) . (2) . (3) . (4) . (5) . (6) . . Income: involved . Credit: involved . Investment: involved . Financial decisions—index . Adult’s exp. decisions—index . All exp. decisions—index . . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Financial literacy 0.065|$^{**}$| 0.038|$^{**}$| 0.075|$^{***}$| 0.172|$^{***}$| 0.286|$^{***}$| 0.400|$^{***}$| (0.033) (0.018) (0.029) (0.065) (0.093) (0.126) Controls Age 0.015|$^{**}$| 0.010|$^{**}$| 0.015|$^{***}$| 0.037|$^{***}$| 0.057|$^{***}$| 0.046|$^{**}$| (0.006) (0.004) (0.005) (0.013) (0.013) (0.023) Age|$^2$| -0.000|$^{*}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.000|$^{**}$| -0.001|$^{***}$| -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.002 -0.010 -0.019 -0.042 -0.061 -0.076 (0.029) (0.017) (0.026) (0.058) (0.082) (0.119) Married -0.033 -0.013 -0.101|$^{***}$| -0.151|$^{*}$| 0.094 -0.075 (0.041) (0.027) (0.035) (0.080) (0.144) (0.236) Divorced 0.114|$^{***}$| 0.004 -0.053 0.055 0.387|$^{**}$| 0.267 (0.044) (0.031) (0.036) (0.090) (0.151) (0.257) Widowed 0.075 0.013 -0.062|$^{*}$| 0.023 0.354|$^{**}$| 0.258 (0.048) (0.027) (0.035) (0.088) (0.149) (0.243) Household size -0.025|$^{***}$| -0.007 -0.015|$^{**}$| -0.046|$^{***}$| -0.030|$^{*}$| -0.058 (0.009) (0.005) (0.006) (0.014) (0.017) (0.040) Children (0–5 years) 0.027 0.011 0.017 0.053|$^{*}$| 0.032 (0.018) (0.010) (0.014) (0.029) (0.053) Children (6–12 years) 0.016 -0.005 0.020|$^{*}$| 0.031 0.055 (0.012) (0.008) (0.010) (0.025) (0.044) Children (13–17 years) 0.011 0.004 0.009 0.023 0.037 (0.013) (0.006) (0.009) (0.020) (0.043) Expenditure (Q2) 0.038 0.002 0.011 0.043 0.044 0.061 (0.025) (0.017) (0.022) (0.053) (0.068) (0.106) Expenditure (Q3) -0.002 -0.016 -0.020 -0.043 -0.053 -0.106 (0.032) (0.020) (0.028) (0.063) (0.075) (0.120) Expenditure (Q4) 0.031 -0.001 -0.030 0.000 -0.018 0.026 (0.032) (0.019) (0.026) (0.059) (0.071) (0.107) Assets—index -0.007 0.000 0.003 -0.004 -0.022 -0.027 (0.008) (0.004) (0.006) (0.014) (0.017) (0.025) Observations 1081 1081 1054 1054 1078 877 |$R^{2}$| 0.051 0.007 -0.018 0.018 -0.008 -0.065 |$F$|-statistic 58.262 58.262 54.482 54.482 60.569 52.894 Notes: The table reports coefficients of IV regressions with standard errors clustered at VSLA level in brackets. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. Outcome variables in columns 1–3 are indicator variables equal to one if a woman has decision-making power in that category. The index in column 4 counts the number of financial decisions (columns 1–3) women are involved in. The index in column 5 counts four expenditure decisions that are unrelated to children. The index in column 6 includes two additional questions related to children’s health and clothes. The instrument used is the average group index of financial literacy excluding the individual considered. Open in new tab 5. Analysing the financial literacy gender gap 5.1. Empirical strategy Previous sections show the following: (a) that there is a significant difference in financial literacy between men and women and (b) that financial literacy is an important aspect for women’s involvement in intra-household decision making. As a tangible consequence, this section investigates why men outperform women on financial literacy and so aims to inform policy makers on how to improve women’s financial literacy. This is done in two steps. First, we run a simple multivariate regression with the financial literacy index as dependent variable to explore the heterogeneity along potential covariates. Second, we use the multivariate decomposition technique popularized by Blinder (1973) and Oaxaca (1973) to study mean outcome differences in financial literacy between men and women. The decomposition tests two explanatory approaches: (i) one that explains differences based on observed characteristics (‘the endowment effect’) and (ii) another that explains differences in returns on these characteristics (‘the coefficient effect’). Differences in financial literacy may exist due to gender differences in endowments; for example, when women are less educated than men. What would be the average financial literacy of women if they would be just as educated as men? Would this counterfactual financial literacy level of women be improved? Or would women still face lower returns to education and thus score lower in financial literacy tests, most likely due to societal or environmental factors. Previous evidence shows that marital status, age, education and income can only partially explain the difference in financial literacy between men and women (Bucher-Koenen et al., 2017, Fonseca et al., 2012). That is why we examine whether differences in financial literacy hold when we apply men’s coefficients to women’s endowments.16 These findings are important to inform policymakers who aim to increase women’s financial literacy by highlighting the relative contribution of personal characteristics (the endowment effect or explained variables) and the cultural and societal context the person lives in (the coefficient effect or unexplained variables). A general formulation of the twofold decomposition technique is provided by Yun (2004). He proposes to decompose differences not only in sample means but rather in first moments and so to extend the linear Blinder–Oaxaca decomposition to non-linear models. Accordingly, the level of financial literacy, |$Y$|, can be explained by a given set of observable characteristics, |$X$|, and coefficients, |$\beta$|: $$\begin{align*}& Y=F(X\beta),\end{align*}$$ where the mapping function, |$F(.)$|, can but not need to be linear as long as it is once differentiable (Yun, 2004). We estimate a linear probability model in the main specification and non-linear models as robustness checks. The difference in financial literacy, |$Y$|, at the first moment between men, |$A$|, and women, |$B$|, can be summarized in the following equation: $$\begin{eqnarray}\overline{Y}_A-\overline{Y}_B&=&[\overline{F(X_A\beta_A)}-\overline{F(X_B\beta_B)}]\end{eqnarray}$$(4) $$\begin{eqnarray}&=&[\underbrace{\overline{F(X_A\beta_A)}-\overline{F(X_B\beta_A)}}_{\textrm{endowment effect}}]+[\underbrace{\overline{F(X_B\beta_A)}-\overline{F(X_B\beta_B)}}_{\textrm{coefficient effect}}]\end{eqnarray}$$(5) The first part describes the overall endowment effect, whereby the latter indicates overall differences in coefficients. Estimating the relative contribution of each variable, |$i$|, to the total gender gap can yield a more detailed picture. Yun (2004) proposes to calculate weights to the endowments and coefficients effects as follows: $$\begin{align}& \overline{Y}_A-\overline{Y}_B=\sum_{i=1}^{i=K}W_{\Delta X}^i[\overline{F(X_A\beta_A)}-\overline{F(X_B\beta_A)}]+\sum_{i=1}^{i=K}W_{\Delta \beta}^i[\overline{F(X_B\beta_A)}-\overline{F(X_B\beta_B)}],\end{align}$$(6) where $$\begin{align*}& W_{\Delta X}^i=\frac{(\overline{X}_A^i-\overline{X}_B^i)\beta_A^i f(\overline{X}_A\beta_A)}{(\overline{X}_A-\overline{X}_B)\beta_A f(\overline{X}_A\beta_A)}=\frac{(\overline{X}_A^i-\overline{X}_B^i)\beta_A^i}{(\overline{X}_A-\overline{X}_B)\beta_A}\end{align*}$$ $$\begin{align*}& W_{\Delta \beta}^i=\frac{\overline{X}_B^i(\beta_A^i-\beta_B^i)f(\overline{X}_B\beta_B)}{\overline{X}_B(\beta_A-\beta_B)f(\overline{X}_B\beta_B)}=\frac{\overline{X}_B^i(\beta_A^i-\beta_B^i)}{\overline{X}_B(\beta_A-\beta_B)}\end{align*}$$ Weights add up exactly to one (100%) and can simply be calculated using the average values of characteristics and their coefficients (Yun, 2004).17 One caveat of detailed decomposition techniques is linked to categorical regressors. Usually, in a regression framework, one of the categories is chosen to be the base category. It is set to zero and all comparisons will be made relative to that category. Oaxaca & Ransom (1999), however, show that the results of the detailed decomposition are not invariant to the choice of the (omitted) base category. We, therefore, follow the solution by Yun (2008) and normalize the effects for a set of indicator variables representing one categorical regressor in the model. 5.2. Regression results Table 5 shows results of multivariate regression analyses. The outcome variable is the financial literacy index and the main variable of interest is the female dummy. For ease of interpretation, explanatory variables are collected into groups and separately introduced into the regression analysis. Results are in line with Lusardi & Mitchell (2014). Women have significantly lower financial literacy than men. As for other control variables, age is humped shaped. Financial literacy first increases with age and then falls for the elderly. This effect turns statistically insignificant when adding household composition variables but its direction remains robust. The number of children in the household may thus be an alternative measure for being middle aged and absorbs the effect of age. For all specifications, our results point to a strong and significantly positive relationship between the ability to write and being financially literate. In contrast, the marital status is insignificant. Happiness as one measure for well-being is significantly positive associated with financial literacy. Similarly, albeit not statistically significant, the depression scale is negatively correlated with financial literacy; meaning that people who are less depressed are more likely to be financially literate. The relationship between the economic status of the household and financial literacy is positive. Even though the asset index is statistically insignificant, we find that those with higher incomes are more financially literate. Having children at school-age also increases the probability of being financially literate. Further, exposure to financial concepts may vary by type of occupation and as such drives differences in financial literacy. Consistent with this theory, we observe that civil servants have higher financial literacy than those without any form of employment. Similar to Aterido et al. (2013), we interpret mobile phone ownership as a proxy for being more open to new ideas. Even if we control for household assets, the effect of mobile phone ownership on financial literacy is positive and statistically significant. Importantly, the coefficient on the female dummy remains significant and about the same size as we introduce a large set of additional control variables. This finding already indicates that the gender gap in financial literacy is not only driven by confounding factors, but that other non-observables also drive this gender gap. OLS regression coefficients in this section cannot be interpreted as causal. As a consequence and because variables that are potentially endogenous in these regressions are not significantly related to financial literacy, we will focus on only exogenous variables in later analyses. Table 5 OLS results on financial literacy . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Sociodemographics Female -0.301|$^{***}$| -0.289|$^{***}$| -0.276|$^{***}$| -0.280|$^{***}$| -0.255|$^{***}$| (0.070) (0.071) (0.071) (0.071) (0.072) Age 0.030|$^{**}$| 0.024|$^{*}$| 0.014 0.014 0.014 (0.014) (0.014) (0.017) (0.017) (0.017) Age|$^2$| -0.000|$^{***}$| -0.000|$^{**}$| -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.628|$^{***}$| 0.577|$^{***}$| 0.583|$^{***}$| 0.577|$^{***}$| 0.534|$^{***}$| (0.073) (0.075) (0.075) (0.075) (0.077) Married 0.055 -0.001 -0.113 -0.128 -0.102 (0.121) (0.122) (0.131) (0.133) (0.133) Divorced -0.037 -0.038 -0.131 -0.118 -0.107 (0.167) (0.167) (0.172) (0.169) (0.169) Widowed -0.162 -0.148 -0.223 -0.235 -0.228 (0.156) (0.157) (0.170) (0.172) (0.174) Happiness 0.167|$^{***}$| 0.143|$^{***}$| 0.152|$^{***}$| 0.147|$^{***}$| 0.147|$^{***}$| (0.054) (0.055) (0.055) (0.054) (0.053) Depression -0.110 -0.091 -0.093 -0.077 -0.078 (0.069) (0.069) (0.069) (0.069) (0.069) Income Expenditure (Q2) 0.142 0.130 0.118 0.111 (0.094) (0.094) (0.094) (0.094) Expenditure (Q3) 0.317|$^{***}$| 0.301|$^{***}$| 0.275|$^{***}$| 0.254|$^{***}$| (0.094) (0.095) (0.096) (0.097) Expenditure (Q4) 0.208|$^{**}$| 0.196|$^{*}$| 0.165 0.135 (0.100) (0.100) (0.101) (0.101) Assets Assets—index 0.036 0.037 0.031 0.007 (0.023) (0.023) (0.023) (0.024) Household Household size -0.018 -0.020 -0.018 (0.025) (0.025) (0.025) Children (0–5 years) 0.071 0.076|$^{*}$| 0.077|$^{*}$| (0.043) (0.044) (0.044) Children (6–12 years) 0.081|$^{**}$| 0.088|$^{**}$| 0.089|$^{**}$| (0.038) (0.038) (0.038) Children (13–17 years) 0.048 0.050 0.045 (0.043) (0.043) (0.042) Employment Farmer (independent) 0.269 0.291 (0.367) (0.361) Other independent occupation 0.389 0.398 (0.367) (0.361) Working for someone else -0.081 -0.065 (0.379) (0.373) Civil servant 0.901|$^{*}$| 0.915|$^{*}$| (0.496) (0.491) Openness Owns mobile phone 0.193|$^{***}$| (0.071) Observations 1405 1405 1405 1405 1405 |$R^2$| 0.160 0.171 0.176 0.183 0.187 . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Sociodemographics Female -0.301|$^{***}$| -0.289|$^{***}$| -0.276|$^{***}$| -0.280|$^{***}$| -0.255|$^{***}$| (0.070) (0.071) (0.071) (0.071) (0.072) Age 0.030|$^{**}$| 0.024|$^{*}$| 0.014 0.014 0.014 (0.014) (0.014) (0.017) (0.017) (0.017) Age|$^2$| -0.000|$^{***}$| -0.000|$^{**}$| -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.628|$^{***}$| 0.577|$^{***}$| 0.583|$^{***}$| 0.577|$^{***}$| 0.534|$^{***}$| (0.073) (0.075) (0.075) (0.075) (0.077) Married 0.055 -0.001 -0.113 -0.128 -0.102 (0.121) (0.122) (0.131) (0.133) (0.133) Divorced -0.037 -0.038 -0.131 -0.118 -0.107 (0.167) (0.167) (0.172) (0.169) (0.169) Widowed -0.162 -0.148 -0.223 -0.235 -0.228 (0.156) (0.157) (0.170) (0.172) (0.174) Happiness 0.167|$^{***}$| 0.143|$^{***}$| 0.152|$^{***}$| 0.147|$^{***}$| 0.147|$^{***}$| (0.054) (0.055) (0.055) (0.054) (0.053) Depression -0.110 -0.091 -0.093 -0.077 -0.078 (0.069) (0.069) (0.069) (0.069) (0.069) Income Expenditure (Q2) 0.142 0.130 0.118 0.111 (0.094) (0.094) (0.094) (0.094) Expenditure (Q3) 0.317|$^{***}$| 0.301|$^{***}$| 0.275|$^{***}$| 0.254|$^{***}$| (0.094) (0.095) (0.096) (0.097) Expenditure (Q4) 0.208|$^{**}$| 0.196|$^{*}$| 0.165 0.135 (0.100) (0.100) (0.101) (0.101) Assets Assets—index 0.036 0.037 0.031 0.007 (0.023) (0.023) (0.023) (0.024) Household Household size -0.018 -0.020 -0.018 (0.025) (0.025) (0.025) Children (0–5 years) 0.071 0.076|$^{*}$| 0.077|$^{*}$| (0.043) (0.044) (0.044) Children (6–12 years) 0.081|$^{**}$| 0.088|$^{**}$| 0.089|$^{**}$| (0.038) (0.038) (0.038) Children (13–17 years) 0.048 0.050 0.045 (0.043) (0.043) (0.042) Employment Farmer (independent) 0.269 0.291 (0.367) (0.361) Other independent occupation 0.389 0.398 (0.367) (0.361) Working for someone else -0.081 -0.065 (0.379) (0.373) Civil servant 0.901|$^{*}$| 0.915|$^{*}$| (0.496) (0.491) Openness Owns mobile phone 0.193|$^{***}$| (0.071) Observations 1405 1405 1405 1405 1405 |$R^2$| 0.160 0.171 0.176 0.183 0.187 Notes: The table reports coefficients of multivariate regressions with standard errors clustered at VSLA level in brackets. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. The outcome variable is the financial literacy—index, which is generated by giving one point for each financial literacy question answered correctly. Happiness and depression are scores on a scale designed to measure mental well-being. The omitted employment category is one if a person has no formal employment. Open in new tab Table 5 OLS results on financial literacy . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Sociodemographics Female -0.301|$^{***}$| -0.289|$^{***}$| -0.276|$^{***}$| -0.280|$^{***}$| -0.255|$^{***}$| (0.070) (0.071) (0.071) (0.071) (0.072) Age 0.030|$^{**}$| 0.024|$^{*}$| 0.014 0.014 0.014 (0.014) (0.014) (0.017) (0.017) (0.017) Age|$^2$| -0.000|$^{***}$| -0.000|$^{**}$| -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.628|$^{***}$| 0.577|$^{***}$| 0.583|$^{***}$| 0.577|$^{***}$| 0.534|$^{***}$| (0.073) (0.075) (0.075) (0.075) (0.077) Married 0.055 -0.001 -0.113 -0.128 -0.102 (0.121) (0.122) (0.131) (0.133) (0.133) Divorced -0.037 -0.038 -0.131 -0.118 -0.107 (0.167) (0.167) (0.172) (0.169) (0.169) Widowed -0.162 -0.148 -0.223 -0.235 -0.228 (0.156) (0.157) (0.170) (0.172) (0.174) Happiness 0.167|$^{***}$| 0.143|$^{***}$| 0.152|$^{***}$| 0.147|$^{***}$| 0.147|$^{***}$| (0.054) (0.055) (0.055) (0.054) (0.053) Depression -0.110 -0.091 -0.093 -0.077 -0.078 (0.069) (0.069) (0.069) (0.069) (0.069) Income Expenditure (Q2) 0.142 0.130 0.118 0.111 (0.094) (0.094) (0.094) (0.094) Expenditure (Q3) 0.317|$^{***}$| 0.301|$^{***}$| 0.275|$^{***}$| 0.254|$^{***}$| (0.094) (0.095) (0.096) (0.097) Expenditure (Q4) 0.208|$^{**}$| 0.196|$^{*}$| 0.165 0.135 (0.100) (0.100) (0.101) (0.101) Assets Assets—index 0.036 0.037 0.031 0.007 (0.023) (0.023) (0.023) (0.024) Household Household size -0.018 -0.020 -0.018 (0.025) (0.025) (0.025) Children (0–5 years) 0.071 0.076|$^{*}$| 0.077|$^{*}$| (0.043) (0.044) (0.044) Children (6–12 years) 0.081|$^{**}$| 0.088|$^{**}$| 0.089|$^{**}$| (0.038) (0.038) (0.038) Children (13–17 years) 0.048 0.050 0.045 (0.043) (0.043) (0.042) Employment Farmer (independent) 0.269 0.291 (0.367) (0.361) Other independent occupation 0.389 0.398 (0.367) (0.361) Working for someone else -0.081 -0.065 (0.379) (0.373) Civil servant 0.901|$^{*}$| 0.915|$^{*}$| (0.496) (0.491) Openness Owns mobile phone 0.193|$^{***}$| (0.071) Observations 1405 1405 1405 1405 1405 |$R^2$| 0.160 0.171 0.176 0.183 0.187 . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . |$\beta$| / SE . Sociodemographics Female -0.301|$^{***}$| -0.289|$^{***}$| -0.276|$^{***}$| -0.280|$^{***}$| -0.255|$^{***}$| (0.070) (0.071) (0.071) (0.071) (0.072) Age 0.030|$^{**}$| 0.024|$^{*}$| 0.014 0.014 0.014 (0.014) (0.014) (0.017) (0.017) (0.017) Age|$^2$| -0.000|$^{***}$| -0.000|$^{**}$| -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Spell 0.628|$^{***}$| 0.577|$^{***}$| 0.583|$^{***}$| 0.577|$^{***}$| 0.534|$^{***}$| (0.073) (0.075) (0.075) (0.075) (0.077) Married 0.055 -0.001 -0.113 -0.128 -0.102 (0.121) (0.122) (0.131) (0.133) (0.133) Divorced -0.037 -0.038 -0.131 -0.118 -0.107 (0.167) (0.167) (0.172) (0.169) (0.169) Widowed -0.162 -0.148 -0.223 -0.235 -0.228 (0.156) (0.157) (0.170) (0.172) (0.174) Happiness 0.167|$^{***}$| 0.143|$^{***}$| 0.152|$^{***}$| 0.147|$^{***}$| 0.147|$^{***}$| (0.054) (0.055) (0.055) (0.054) (0.053) Depression -0.110 -0.091 -0.093 -0.077 -0.078 (0.069) (0.069) (0.069) (0.069) (0.069) Income Expenditure (Q2) 0.142 0.130 0.118 0.111 (0.094) (0.094) (0.094) (0.094) Expenditure (Q3) 0.317|$^{***}$| 0.301|$^{***}$| 0.275|$^{***}$| 0.254|$^{***}$| (0.094) (0.095) (0.096) (0.097) Expenditure (Q4) 0.208|$^{**}$| 0.196|$^{*}$| 0.165 0.135 (0.100) (0.100) (0.101) (0.101) Assets Assets—index 0.036 0.037 0.031 0.007 (0.023) (0.023) (0.023) (0.024) Household Household size -0.018 -0.020 -0.018 (0.025) (0.025) (0.025) Children (0–5 years) 0.071 0.076|$^{*}$| 0.077|$^{*}$| (0.043) (0.044) (0.044) Children (6–12 years) 0.081|$^{**}$| 0.088|$^{**}$| 0.089|$^{**}$| (0.038) (0.038) (0.038) Children (13–17 years) 0.048 0.050 0.045 (0.043) (0.043) (0.042) Employment Farmer (independent) 0.269 0.291 (0.367) (0.361) Other independent occupation 0.389 0.398 (0.367) (0.361) Working for someone else -0.081 -0.065 (0.379) (0.373) Civil servant 0.901|$^{*}$| 0.915|$^{*}$| (0.496) (0.491) Openness Owns mobile phone 0.193|$^{***}$| (0.071) Observations 1405 1405 1405 1405 1405 |$R^2$| 0.160 0.171 0.176 0.183 0.187 Notes: The table reports coefficients of multivariate regressions with standard errors clustered at VSLA level in brackets. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. The outcome variable is the financial literacy—index, which is generated by giving one point for each financial literacy question answered correctly. Happiness and depression are scores on a scale designed to measure mental well-being. The omitted employment category is one if a person has no formal employment. Open in new tab 5.3. Decomposition results Decomposition results are shown in Table 6. Both analyses estimate a linear probability model with the financial literacy index as outcome variable. The left-hand side of the table does not contain a measure of wealth, whereas the right-hand side does in form of the asset index. The table reports the coefficient estimates along with percentage shares. Standard errors are cluster-adjusted at the VSLA level in order to account for intra-group correlation. Overall, the mean of the financial literacy index is 2.833 for men and 2.356 for women. This yields a gender gap of 0.477. The increase of 0.218 indicates that 46% of the gap stems from gender differences in endowments. The remaining 54% of the financial literacy gender gap can be attributed to gender differences in returns on these endowments. The second and third panels of Table 6 show results of the detailed decomposition. We see that spelling as a proxy for educational attainment contributes about 17% to the gender gap in financial literacy. Furthermore, happiness as a measure of individual well-being also eliminates the gap in financial literacy by 5%. Though statistically insignificant, improved symptoms of depression would also result in reduced gender differences in financial literacy. Further, mobile phone ownership can significantly reduce the gender gap by about 12%. The second analysis only differs to the first one by controlling for wealth. We can see that this specification yields similar results and that mobile phone ownership keeps its explanatory power. This suggests that mobile phone ownership is not only another wealth measure but also indicates something we interpret as openness to new ideas. Aterido et al. (2013) use a similar line of argumentation in order to explain the lower usage of formal banking services by women in sub-Saharan Africa. On the bottom line, this decomposition analysis shows that 46% of the financial literacy gender gap can be attributed to endowment effects. 20% of this can particularly be linked to personality traits such as openness (12%), happiness (5%) and depression (3%). The finding shows that a large part of the gender gap has its roots in social environments. We argue that the remaining coefficient effect also captures some of these cultural and societal circumstances in women’s lives—a point that is common in the literature on gender gaps in general. Scholars have argued that gender differences are broadly consistent with gender stereotypes across cultures (American Psychological Association, 1994, Costa et al., 2001, Nolen-Hoeksema, 1987, Thayer et al., 2003). Eagly (2013) explains that perceived differences between men and women might result from adoption of gender roles, which predetermine appropriate conduct for each gender. Table 6 Full decomposition results of financial literacy . Without assets . . With assets . . Coefficient . Share . . Coefficient . Share . Overall Male 2.833|$^{***}$| Male 2.833|$^{***}$| Female 2.356|$^{***}$| Female 2.356|$^{***}$| Difference 0.477|$^{***}$| 100.000 Difference 0.477|$^{***}$| 100,000 Endowment eff. 0.218|$^{***}$| 45.705 Endowment eff. 0.219|$^{***}$| 45.791 Coefficient eff. 0.259|$^{***}$| 54.295 Coefficient eff. 0.259|$^{***}$| 54.209 Endowment eff. Age -0.055 -11.438 Age -0.052 -10.922 Age|$^2$| 0.065 13.557 Age|$^2$| 0.063 13.244 Single 0.006 1.190 Single 0.006 1.274 Married 0.007 1.430 Married 0.006 1.195 Divorced 0.000 0.047 Divorced 0.000 0.020 Widowed 0.019 4.055 Widowed 0.019 3.912 Spell 0.083|$^{***}$| 17.490 Spell 0.083|$^{***}$| 17.379 Children (6–12) -0.002 -0.351 Children (6–12) -0.002 -0.349 Happy 0.023|$^{**}$| 4.809 Happy 0.022|$^{**}$| 4.675 Depression 0.012 2.503 Depression 0.012 2.440 Mobile phone 0.059|$^{***}$| 12.412 Mobile phone 0.053|$^{***}$| 11.161 Asset index 0.008 1.763 Coefficient eff. Age 1.196 250.603 Age 0.795 166.545 Age|$^2$| -0.551 -115.425 Age|$^2$| -0.400 -83.930 Single -0.025 -5.271 Single -0.022 -4.625 Married -0.034 -7.121 Married -0.055 -11.555 Divorced 0.016 3.274 Divorced 0.017 3.660 Widowed -0.006 -1.227 Widowed -0.006 -1.233 Spell 0.038 7.977 Spell -0.001 -0.308 Children (6–12) 0.031 6.455 Children (6–12) 0.026 5.552 Happy -0.447 -93.592 Happy -0.558 -116.863 Depression -0.035 -7.375 Depression -0.027 -5.581 Mobile phone 0.030 6.274 Mobile phone -0.091 -19.142 Asset index 0.035|$^{**}$| 7.341 Constant 0.046 9.723 Constant 0.546 114.349 . Without assets . . With assets . . Coefficient . Share . . Coefficient . Share . Overall Male 2.833|$^{***}$| Male 2.833|$^{***}$| Female 2.356|$^{***}$| Female 2.356|$^{***}$| Difference 0.477|$^{***}$| 100.000 Difference 0.477|$^{***}$| 100,000 Endowment eff. 0.218|$^{***}$| 45.705 Endowment eff. 0.219|$^{***}$| 45.791 Coefficient eff. 0.259|$^{***}$| 54.295 Coefficient eff. 0.259|$^{***}$| 54.209 Endowment eff. Age -0.055 -11.438 Age -0.052 -10.922 Age|$^2$| 0.065 13.557 Age|$^2$| 0.063 13.244 Single 0.006 1.190 Single 0.006 1.274 Married 0.007 1.430 Married 0.006 1.195 Divorced 0.000 0.047 Divorced 0.000 0.020 Widowed 0.019 4.055 Widowed 0.019 3.912 Spell 0.083|$^{***}$| 17.490 Spell 0.083|$^{***}$| 17.379 Children (6–12) -0.002 -0.351 Children (6–12) -0.002 -0.349 Happy 0.023|$^{**}$| 4.809 Happy 0.022|$^{**}$| 4.675 Depression 0.012 2.503 Depression 0.012 2.440 Mobile phone 0.059|$^{***}$| 12.412 Mobile phone 0.053|$^{***}$| 11.161 Asset index 0.008 1.763 Coefficient eff. Age 1.196 250.603 Age 0.795 166.545 Age|$^2$| -0.551 -115.425 Age|$^2$| -0.400 -83.930 Single -0.025 -5.271 Single -0.022 -4.625 Married -0.034 -7.121 Married -0.055 -11.555 Divorced 0.016 3.274 Divorced 0.017 3.660 Widowed -0.006 -1.227 Widowed -0.006 -1.233 Spell 0.038 7.977 Spell -0.001 -0.308 Children (6–12) 0.031 6.455 Children (6–12) 0.026 5.552 Happy -0.447 -93.592 Happy -0.558 -116.863 Depression -0.035 -7.375 Depression -0.027 -5.581 Mobile phone 0.030 6.274 Mobile phone -0.091 -19.142 Asset index 0.035|$^{**}$| 7.341 Constant 0.046 9.723 Constant 0.546 114.349 Notes: The table reports Blinder–Oaxaca decomposition results with standard errors clustered at VSLA level. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. Due to rounding shares may not add up. Open in new tab Table 6 Full decomposition results of financial literacy . Without assets . . With assets . . Coefficient . Share . . Coefficient . Share . Overall Male 2.833|$^{***}$| Male 2.833|$^{***}$| Female 2.356|$^{***}$| Female 2.356|$^{***}$| Difference 0.477|$^{***}$| 100.000 Difference 0.477|$^{***}$| 100,000 Endowment eff. 0.218|$^{***}$| 45.705 Endowment eff. 0.219|$^{***}$| 45.791 Coefficient eff. 0.259|$^{***}$| 54.295 Coefficient eff. 0.259|$^{***}$| 54.209 Endowment eff. Age -0.055 -11.438 Age -0.052 -10.922 Age|$^2$| 0.065 13.557 Age|$^2$| 0.063 13.244 Single 0.006 1.190 Single 0.006 1.274 Married 0.007 1.430 Married 0.006 1.195 Divorced 0.000 0.047 Divorced 0.000 0.020 Widowed 0.019 4.055 Widowed 0.019 3.912 Spell 0.083|$^{***}$| 17.490 Spell 0.083|$^{***}$| 17.379 Children (6–12) -0.002 -0.351 Children (6–12) -0.002 -0.349 Happy 0.023|$^{**}$| 4.809 Happy 0.022|$^{**}$| 4.675 Depression 0.012 2.503 Depression 0.012 2.440 Mobile phone 0.059|$^{***}$| 12.412 Mobile phone 0.053|$^{***}$| 11.161 Asset index 0.008 1.763 Coefficient eff. Age 1.196 250.603 Age 0.795 166.545 Age|$^2$| -0.551 -115.425 Age|$^2$| -0.400 -83.930 Single -0.025 -5.271 Single -0.022 -4.625 Married -0.034 -7.121 Married -0.055 -11.555 Divorced 0.016 3.274 Divorced 0.017 3.660 Widowed -0.006 -1.227 Widowed -0.006 -1.233 Spell 0.038 7.977 Spell -0.001 -0.308 Children (6–12) 0.031 6.455 Children (6–12) 0.026 5.552 Happy -0.447 -93.592 Happy -0.558 -116.863 Depression -0.035 -7.375 Depression -0.027 -5.581 Mobile phone 0.030 6.274 Mobile phone -0.091 -19.142 Asset index 0.035|$^{**}$| 7.341 Constant 0.046 9.723 Constant 0.546 114.349 . Without assets . . With assets . . Coefficient . Share . . Coefficient . Share . Overall Male 2.833|$^{***}$| Male 2.833|$^{***}$| Female 2.356|$^{***}$| Female 2.356|$^{***}$| Difference 0.477|$^{***}$| 100.000 Difference 0.477|$^{***}$| 100,000 Endowment eff. 0.218|$^{***}$| 45.705 Endowment eff. 0.219|$^{***}$| 45.791 Coefficient eff. 0.259|$^{***}$| 54.295 Coefficient eff. 0.259|$^{***}$| 54.209 Endowment eff. Age -0.055 -11.438 Age -0.052 -10.922 Age|$^2$| 0.065 13.557 Age|$^2$| 0.063 13.244 Single 0.006 1.190 Single 0.006 1.274 Married 0.007 1.430 Married 0.006 1.195 Divorced 0.000 0.047 Divorced 0.000 0.020 Widowed 0.019 4.055 Widowed 0.019 3.912 Spell 0.083|$^{***}$| 17.490 Spell 0.083|$^{***}$| 17.379 Children (6–12) -0.002 -0.351 Children (6–12) -0.002 -0.349 Happy 0.023|$^{**}$| 4.809 Happy 0.022|$^{**}$| 4.675 Depression 0.012 2.503 Depression 0.012 2.440 Mobile phone 0.059|$^{***}$| 12.412 Mobile phone 0.053|$^{***}$| 11.161 Asset index 0.008 1.763 Coefficient eff. Age 1.196 250.603 Age 0.795 166.545 Age|$^2$| -0.551 -115.425 Age|$^2$| -0.400 -83.930 Single -0.025 -5.271 Single -0.022 -4.625 Married -0.034 -7.121 Married -0.055 -11.555 Divorced 0.016 3.274 Divorced 0.017 3.660 Widowed -0.006 -1.227 Widowed -0.006 -1.233 Spell 0.038 7.977 Spell -0.001 -0.308 Children (6–12) 0.031 6.455 Children (6–12) 0.026 5.552 Happy -0.447 -93.592 Happy -0.558 -116.863 Depression -0.035 -7.375 Depression -0.027 -5.581 Mobile phone 0.030 6.274 Mobile phone -0.091 -19.142 Asset index 0.035|$^{**}$| 7.341 Constant 0.046 9.723 Constant 0.546 114.349 Notes: The table reports Blinder–Oaxaca decomposition results with standard errors clustered at VSLA level. *|$p < 0.10$|, **|$p < 0.05$| and ***|$p < 0.01$| denote statistical significance. Due to rounding shares may not add up. Open in new tab 6. Robustness In this robustness section, we commit our results to a number of checks. Results are presented in the online Appendix. First, we compare alternative financial literacy measures and confirm a positive relationship between financial literacy and household decision making. Similar to van Rooij et al. (2011), a financial literacy factor is derived using an iterated principle factor analysis, followed by the Bartlett method (Bartlett, 1937). The estimated factor score of the first factor acts as a proxy for financial literacy. We also use a dummy that is one if the respondent answered all financial literacy questions correctly. In comparison, Table A18 shows a positive and mostly significant relationship. Using probit instead of OLS regressions, we test whether the link between financial literacy and financial decision making is robust to changes in estimation strategy (see Table A19). The marginal effects are slightly smaller but very similar to the coefficients of the linear probability model in Table 3. In further robustness tests, the linear probability model is our preferred specification because it can easily be extended to IV estimation. Further, we control for education instead of using a dummy indicating whether the person is able to spell a simple word in Kinyarwanda. Several of the questions used to measure financial literacy require the respondent to calculate percentages and to understand the principle of compound interest—both of which require more than basic literacy. That is why we control for educational attainment as robustness analysis, omitting primary education and less as the base category. Table A20 strengthens our results and show that financial literacy remains positively related to decision making. As a result, the finding that higher financial literacy increases the likelihood that women will participate in household decision making is not only capturing the effect of educational attainment but also rather financial literacy that highly matters. We diverge from homogeneous effects and estimate the link between financial literacy and financial decision making for different sub-samples: (Panel A) only married women, (Panel B) women who report that they do not have to ask for permission to attend or travel to a meeting and (Panel C) those who need permission to do so. Results in Table A21 remain significant for married women. Estimates for women who need permission and those who do not are positive but not exclusively significant, which is potentially caused by a small sample. Next, we further augment our IV identification strategy. We report results adding a second instrument that is borderline not weak in Table A17 and Table A22. Table A23 shows a positive but insignificant correlation between financial literacy and distance to nearest school,18 which is why we focus on the group instrument throughout this manuscript. We also add robustness to the findings in Table 5 and change the measure of financial literacy. Table A24 and A25 show a negative and significant relationship between the female dummy and financial literacy in all regressions, even after adding further controls. Finally as for the decomposition analysis, we show alternative results in Table A26 using (i) a linear probability model with the financial literacy factor score and (ii) a non-linear probability model with the discrete financial literacy dummy.19 These analyses yield similar results as in Table 6.20 If anything, the endowment effect is slightly reduced in the non-linear specification. A possible reason is that the dummy for only correct answers captures less variation and is too short sighted. We also add group variables such as the age of the group, total number of members, yearly share out and the default rate to the decomposition. Table A27 shows that group characteristics, however, do not close the gender gap in financial literacy. 7. Conclusion This article explores the relationship between financial literacy, gender and decision-making power within the household. Using both OLS and IV regression analyses, we first study whether financial literacy has an effect on women to participate in decision-making processes at home. Our findings indicate that women with higher financial literacy are more likely to report that they are involved in income, credit, investment and expenditure decisions. This result is consistent with the hypothesis that financial literacy is a resource of empowerment and enhances women’s involvement in intra-household decision making. Motivated by this and to deepen our understanding why women lack behind men in terms of financial literacy, we examine this gender gap in detail. Using a multivariate decomposition technique, we find that about 46% of the gender gap is explained by different endowments between men and women. The largest part of this is made up of differences in education and personality traits. A total of 54% of the gap can be attributed to gender differences in returns on these endowments. Similar to Bucher-Koenen et al. (2017), Filipiak & Walle (2015), and Grohmann et al. (2016), we argue that it is reasonable to believe that this coefficient effect captures some of the societal and cultural circumstances in women’s lives that may prevent them from achieving higher financial literacy rates. Clear policy lessons can be drawn from this research. First, it provides motivation to improve women’s financial literacy, especially in developing countries. The decomposition analysis shows that improved educational levels should result in higher financial literacy levels. Moreover, a large part of financial literacy differences between men and women is associated with personality traits. Financial literacy trainings should, therefore, take into account gender differences in personality and tailor content and delivery methods accordingly. Further, our results inform policymakers by highlighting that personal characteristics contribute about half to the financial literacy gender gap and that also cultural and societal factors are relevant. It is, therefore, possible that cross country studies or studies that look at personality traits and gender roles in more detail will provide further insights into the origins of the gender gap in financial literacy. Supplementary material Supplementary material is available at JAFECO online. Footnotes 1 See Karlan et al. (2017) for a detailed description of the VSLA model. 2 Only Burundi comprises more VSLAs relative to its population (CARE, 2017). 3 Even though VSLA members are predominantly rural, poor women; in Africa, groups can be gender mixed. Mali is the only exception with almost no male members. In remaining Africa, the female share ranges from 62% in Angola and Mozambique to 95% in Benin and Togo. The average female share across 30 African countries listed is 81% (CARE, 2017). 4 This list was acquired by an international non-governmental organization that has been promoting and expanding the VSLA model in Rwanda. 5 In Rwanda, poor people are selected into the first or second ‘Ubudehe category’. 6 If no more than two% of the covariate’s values are missing, we recode missing values to the overall mean for the relevant VSLA group. 7 We use the widely known ‘Center for Epidemiologic Studies Depression Scale Revised’ (CESD-R). It is standard battery of 20 questions that measure depression and depressive disorders in nine different groups: sadness, loss of interest, appetite, sleep, thinking and concentration, guilt, tired, movement and suicidal ideation (Eaton et al., 2004, Radloff, 1977). 8 We use expenditures to proxy for income. All expenditure categories were aggregated on a yearly base and further divided into fourths. 9 The sample size varies for this index because not everyone in our sample stated to have dependents. 10 We abstract from the subscript i in all equations. 11 We are unable to control for spouse’s characteristics, as no data on is available. 12 In Table A7, we regress financial literacy against whether a woman reports that she makes financial decisions alone. Results show a negative relationship, probably because women who make household decisions themselves are more likely to be widowed and these tend to have lower financial literacy. 13 Table A8 shows results for OLS regressions on expenditure decisions at home. There is a significant positive correlation between a woman’s financial literacy and her involvement in household’s food and energy consumption as well as her own health and clothing decisions. Her decisions regarding children’s health and clothing are not significantly associated with financial literacy. 14 Results separated by expenditure decisions are listed in Table A10 in the Appendix. 15 VAs are local VSLA members who facilitate and train VSLAs to conduct their savings and lending activities. They work on a fee for service basis in their community and neighboring areas. 16 Depending on the context of the research question, the coefficient effect has been interpreted in different ways. In the gender wage gap literature, for instance, this effect has often been used as a measure for discrimination (Blinder, 1973, Oaxaca, 1973). 17 For non-linear models, however, results are sensitive to the order in which independent variables enter the decomposition. Yun (2004) proposes a convenient solution for the so-called ‘path dependence’. He obtains weights from a first-order Taylor expression to linearize the endowments and coefficients effects in equation (5) around |$\overline{X}_A\beta _A$| and |$\overline{X}_B\beta _B$|. 18 The National Statistics Office of Rwanda publishes latitude and longitude data for all schools in Rwanda; we use this to calculate the distance between the residence of each respondent and the nearest school. 19 For probit decomposition analysis, the mapping function, |$F(.)$|, is the cumulative distribution function (CDF) of the standard normal distribution. 20 This holds for both in total and in detail. The detailed results can be provided upon request. References Abbott P. , Mugisha R., Sapsford R. ( 2018 ) ‘ Women, land and empowerment in Rwanda ’, Journal of International Development , 30 : 1006 – 22 . Google Scholar OpenURL Placeholder Text WorldCat Allendorf K. ( 2007 ) ‘ Do women’s land rights promote empowerment and child health in Nepal? ’, World Development , 35 : 1975 – 88 . Google Scholar Crossref Search ADS PubMed WorldCat Alsop R. , Heinsohn N., Somma A. ( 2005 ) ‘ Measuring empowerment: An analytic framework ’, in R. Alsop (ed.) , Power, Rights and Poverty: Concepts and Connections . Washington, DC : World Bank . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC American Psychological Association ( 1994 ) Diagnostic and Statistical Manual of Mental Disorders , Washington : American Psychological Association Publishing . Anderson S. , Eswaran M. ( 2009 ) ‘ What determines female autonomy? Evidence from Bangladesh ’, Journal of Development Economics , 90 : 179 – 91 . Google Scholar Crossref Search ADS WorldCat Ashraf N. , Karlan D., Yin W. ( 2010 ) ‘ Female empowerment: Impact of a commitment savings product in the Philippines ’, World Development , 38 : 333 – 44 . Google Scholar Crossref Search ADS WorldCat Aterido R. , Beck T., Iacovone L. ( 2013 ) ‘ Access to finance in sub-Saharan Africa: Is there a gender gap? ’, World Development , 47 : 102 – 20 . Google Scholar Crossref Search ADS WorldCat Bartlett M. S. ( 1937 ) ‘ The statistical conception of mental factors ’, British Journal of Psychology. General Section , 28 : 97 – 104 . Google Scholar Crossref Search ADS WorldCat Bernard T. , Doss C., Hidrobo M., Hoel J., Kieran C. ( 2020 ) ‘ Ask me why: Patterns of intra-household decision making ’, World Development , 125 : 104671 . Google Scholar Crossref Search ADS WorldCat Blinder A. S. ( 1973 ) ‘ Wage discrimination: Reduced form and structural estimates ’, Journal of Human Resources , 4 : 436 – 55 . Google Scholar Crossref Search ADS WorldCat Bobonis G. J. ( 2009 ) ‘ Is the allocation of resources within the household efficient? New evidence from a randomized experiment ’, Journal of Political Economy , 117 : 453 – 503 . Google Scholar Crossref Search ADS WorldCat Bucher-Koenen T. , Lusardi A., Alessie R. J. M., van Rooij M. C. J. ( 2017 ) ‘ How financially literate are women? An overview and new insights ’, Journal of Consumer Affairs , 51 : 255 – 83 . Google Scholar Crossref Search ADS WorldCat CARE ( 2017 ) ‘ An overview of the global reach of CARE’s village savings and loans association programming ’, Available at: https://insights.careinternational.org.uk/media/k2/attachments/CARE-VSLA-Global-Outreach-Report-2017.pdf. [Accessed: 24 January 2020] . Cole S. , Sampson T., Zia B. ( 2011 ) ‘ Prices or knowledge? What drives demand for financial services in emerging markets? ’, The Journal of Finance , 66 : 1933 – 67 . Google Scholar Crossref Search ADS WorldCat Connelly R. , Roberts K., Zheng Z. ( 2010 ) ‘ The impact of circular migration on the position of married women in rural China ’, Feminist Economics , 16 : 3 – 41 . Google Scholar Crossref Search ADS WorldCat Costa P. , Jr., Terracciano A., McCrae R. R. ( 2001 ) ‘ Gender differences in personality traits across cultures: Robust and surprising findings ’, Journal of Personality and Social Psychology , 81 : 322 – 31 . Google Scholar Crossref Search ADS PubMed WorldCat De Brauw A. , Gilligan D. O., Hoddinott J., Roy S. ( 2014 ) ‘ The impact of Bolsa Família on women’s decision-making power ’, World Development , 59 : 487 – 504 . Google Scholar Crossref Search ADS WorldCat DFID ( 2010 ) ‘ Department for international development. Agenda 2010: The turning point on poverty. Background paper, London ’, Available at: http://www.oecd.org/dac/gender-development/45249900.pdf. [Accessed: 29 July 2019] . Doi Y. , McKenzie D., Zia B. ( 2014 ) ‘ Who you train matters: Identifying combined effects of financial education on migrant households ’, Journal of Development Economics , 109 : 39 – 55 . Google Scholar Crossref Search ADS WorldCat Doss C. ( 2006 ) ‘ The effects of intra-household property ownership on expenditure patterns in Ghana ’, Journal of African Economies , 15 : 149 – 80 . Google Scholar Crossref Search ADS WorldCat Doss C. ( 2013 ) ‘ Intrahousehold bargaining and resource allocation in developing countries ’, The World Bank Research Observer , 28 : 52 – 78 . Google Scholar Crossref Search ADS WorldCat Doss C. R. ( 2001 ) ‘ Is risk fully pooled within the household? Evidence from Ghana ’, Economic Development and Cultural Change , 50 : 101 – 30 . Google Scholar Crossref Search ADS WorldCat Duflo E. ( 2003 ) ‘ Grandmothers and granddaughters: Old-age pensions and intrahousehold allocation in South Africa ’, The World Bank Economic Review , 17 : 1 – 25 . Google Scholar Crossref Search ADS WorldCat Duflo E. ( 2012 ) ‘ Women empowerment and economic development ’, Journal of Economic Literature , 50 : 1051 – 79 . Google Scholar Crossref Search ADS WorldCat Duflo E. , Udry C. ( 2004 ) ‘ Intrahousehold Resource Allocation in Cote d’Ivoire: Social Norms, Separate Accounts and Consumption Choices ’, National Bureau of Economic Research Working Paper No. 10498 . Eagly A. H. ( 2013 ) Sex Differences in Social Behavior: A Social-Role Interpretation . Psychology Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Eaton W. W. , Smith C., Ybarra M., Muntaner C., Tien A. ( 2004 ) ‘ Center for Epidemiologic Studies Depression Scale: review and revision (CESD and CESD-R) ’, in M. E. Maruish (ed.) , The Use of Psychological Testing for Treatment Planning and Outcomes Assessment , 3rd edn. Mahwah, NJ : Lawrence Erlbaum , pp. 363 – 77 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Fiala N. , He X. ( 2017 ) ‘ Unitary or noncooperative intrahousehold model? Evidence from couples in Uganda ’, The World Bank Economic Review , 30 : S77 – 85 . Google Scholar OpenURL Placeholder Text WorldCat Filipiak U. , Walle Y. M. ( 2015 ) ‘ The Financial Literacy Gender Gap: A Question of Nature or Nurture? ’, Courant Research Centre: Poverty, Equity and Growth Discussion Paper No. 176 . Google Scholar FinScope ( 2013 ) ‘ Finscope Rwanda 2012: Access to finance Rwanda ’, Available at: http://statistics.gov.rw/publication/finscope-survey-report-2012. [Accessed: 4 July 2019] . FinScope ( 2016 ) ‘ FinScope Rwanda 2016: Financial inclusion in Rwanda ’, Available at: http://www.statistics.gov.rw/datasource/finscope-survey-2016. [Accessed: 4 July 2019] . Fonseca R. , Mullen K. J., Zamarro G., Zissimopoulos J. ( 2012 ) ‘ What explains the gender gap in financial literacy? The role of household decision making ’, Journal of Consumer Affairs , 46 : 90 – 106 . Google Scholar Crossref Search ADS PubMed WorldCat Gathergood J. ( 2012 ) ‘ Self-control, financial literacy and consumer over-indebtedness ’, Journal of Economic Psychology , 33 : 590 – 602 . Google Scholar Crossref Search ADS WorldCat Grohmann A. , Hübler O., Kouwenberg R., Menkhoff L. ( 2016 ) ‘ Financial Literacy: Thai Middle Class Women Do Not Lag Behind ’, DIW Working Paper No. 1615 . Hashemi S. M. , Schuler S. R., Riley A. P. ( 1996 ) ‘ Rural credit programs and women’, empowerment in Bangladesh ’, World Development , 24 : 635 – 53 . Google Scholar Crossref Search ADS WorldCat Hsu J. W. ( 2016 ) ‘ Aging and strategic learning: The impact of spousal incentives on financial literacy ’, Journal of Human Resources , 51 : 1036 – 67 . Google Scholar Crossref Search ADS PubMed WorldCat Kabeer N. ( 1999 ) ‘ Resources, agency, achievements: Reflections on the measurement of women’s empowerment ’, Development and Change , 30 : 435 – 64 . Google Scholar Crossref Search ADS WorldCat Kaiser T. , Menkhoff L. ( 2017 ) ‘ Does financial education impact financial literacy and financial behavior, and if so, when? ’ The World Bank Economic Review , 31 : 611 – 30 . Google Scholar Crossref Search ADS WorldCat Karlan D. , Savonitto B., Thuysbaert B., Udry C. ( 2017 ) ‘ Impact of savings groups on the lives of the poor ’, Proceedings of the National Academy of Sciences , 114 : 3079 – 84 . Google Scholar Crossref Search ADS WorldCat Lusardi A. , Mitchell O. S. ( 2007 ) ‘ Baby boomer retirement security: The roles of planning, financial literacy, and housing wealth ’, Journal of Monetary Economics , 54 : 205 – 24 . Google Scholar Crossref Search ADS WorldCat Lusardi A. , Mitchell O. S. ( 2011 ) ‘ Financial literacy and planning: Implications for retirement wellbeing ’, National Bureau of Economic Research Working Paper No. 17078 . Lusardi A. , Mitchell O. S. ( 2014 ) ‘ The economic importance of financial literacy: Theory and evidence ’, Journal of Economic Literature , 52 : 5 – 44 . Google Scholar Crossref Search ADS PubMed WorldCat Lusardi A. , Tufano P. ( 2015 ) ‘ Debt literacy, financial experiences, and overindebtedness ’, Journal of Pension Economics and Finance , 14 : 332 – 68 . Google Scholar Crossref Search ADS WorldCat Mahdavi M. , Horton N. J. ( 2014 ) ‘ Financial knowledge among educated women: Room for improvement ’, Journal of Consumer Affairs , 48 : 403 – 17 . Google Scholar Crossref Search ADS WorldCat Manda D. K. , Mwakubo S. ( 2014 ) ‘ Gender and economic development in Africa: an overview ’, Journal of African Economies , 23 : i4 – i17 . Google Scholar Crossref Search ADS WorldCat McElroy M. B. , Horney M. J. ( 1981 ) ‘ Nash-bargained household decisions: Toward a generalization of the theory of demand ’, International Economic Review , 333 – 49 . Google Scholar OpenURL Placeholder Text WorldCat Moursund A. , Kravdal Ø. ( 2003 ) ‘ Individual and community effects of women’s education and autonomy on contraceptive use in India ’, Population Studies , 57 : 285 – 301 . Google Scholar Crossref Search ADS PubMed WorldCat NISR ( 2012 ) ‘ National Institute of Statistics of Rwanda. Rwanda Population and Housing Census 2012 ’, Available at: http://microdata.statistics.gov.rw/index.php/catalog/65. [Accessed: 23 February 2016] . OpenURL Placeholder Text WorldCat Nolen-Hoeksema S. ( 1987 ) ‘ Sex differences in unipolar depression: Evidence and theory ’, Psychological Bulletin , 101 : 259 . Google Scholar Crossref Search ADS PubMed WorldCat Oaxaca R. ( 1973 ) ‘ Male-female wage differentials in urban labor markets ’, International Economic Review , 14 : 693 – 709 . Google Scholar Crossref Search ADS WorldCat Oaxaca R. L. , Ransom M. R. ( 1999 ) ‘ Identification in detailed wage decompositions ’, Review of Economics and Statistics , 81 : 154 – 7 . Google Scholar Crossref Search ADS WorldCat Radloff L. S. ( 1977 ) ‘ The CES-D scale: A self-report depression scale for research in the general population ’, Applied Psychological Measurement , 1 : 385 – 401 . Google Scholar Crossref Search ADS WorldCat Sayinzoga A. , Bulte E. H., Lensink R. ( 2015 ) ‘ Financial literacy and financial behaviour: Experimental evidence from rural Rwanda ’, The Economic Journal , 126 : 1571 – 99 . Google Scholar Crossref Search ADS WorldCat Thayer J. F. , Rossy L. A., Ruiz-Padial E., Johnsen B. H. ( 2003 ) ‘ Gender differences in the relationship between emotional regulation and depressive symptoms ’, Cognitive Therapy and Research , 27 : 349 – 64 . Google Scholar Crossref Search ADS WorldCat Thomas D. ( 1990 ) ‘ Intra-household resource allocation: An inferential approach ’, Journal of Human Resources , 25 : 635 – 64 . Google Scholar Crossref Search ADS WorldCat Thomas D. ( 1993 ) ‘ The distribution of income and expenditure within the household ’, Annales d’Economie et de Statistique , 29 : 109 – 35 . Google Scholar Crossref Search ADS WorldCat UN ( 2015 ) ‘ Transforming Our World: The 2030 Agenda for Sustainable Development ’, UN, General Assembly, A/RES/70/1, 21 October . van Rooij M. C. J. , Lusardi A., Alessie R. J. M. ( 2011 ) ‘ Financial literacy and stock market participation ’, Journal of Financial Economics , 101 : 449 – 72 . Google Scholar Crossref Search ADS WorldCat van Rooij M. C. J. , Lusardi A., Alessie R. J. M. ( 2012 ) ‘ Financial literacy, retirement planning and household wealth ’, The Economic Journal , 122 : 449 – 78 . Google Scholar Crossref Search ADS WorldCat Xu L. , Zia B. ( 2012 ) ‘ Financial literacy around the world: An overview of the evidence with practical suggestions for the way forward ’, World Bank Policy Research Working Paper No. 6107 . Yun M.-S. ( 2004 ) ‘ Decomposing differences in the first moment ’, Economics Letters , 82 : 275 – 80 . Google Scholar Crossref Search ADS WorldCat Yun M.-S. ( 2008 ) ‘ Identification problem and detailed Oaxaca decomposition: A general solution and inference ’, Journal of Economic and Social Measurement , 33 : 27 – 38 . Google Scholar Crossref Search ADS WorldCat Author notes †The authors gratefully acknowledge support of a special grant (Sondertatbestand) from the German Federal Ministry for Economic Affairs and Energy and the Ministry of Innovation, Science, and Research of the State of North Rhine-Westphalia and financial support by the German Research Foundation through CRC TRR 190 (project number 280092119). The data underlying this research was collected in cooperation with Genesis Analytics and ikapadata for an impact evaluation commissioned by CARE Canada and MasterCard Foundation. The authors are grateful to Noel Verrinder for valuable input to the study design and fruitful cooperation in implementing the survey and to participants of a number of conferences, as well as Olaf Huebler, Lukas Menkhoff, Jörg Peters and Sebastian Schneider for useful comments. Special thanks to the editor, Jenny Aker and two anonymous referees. © The Author(s) 2020. Published by Oxford University Press on behalf of the Centre for the Study of African Economies, all rights reserved. For Permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
Journal
Journal of African Economies – Oxford University Press
Published: Oct 20, 2020