The Scorpion Who Stings the Dog Who Bites: The Effect of Women's Different Job Positions on Gender Discrimination in Microfinance

"I prefer male enemies to female ones; I can survive a dog's bite better than a scorpion's sting."(Matshona Dhliwayo)IntroductionEquitable access for women to employment, entrepreneurship and credit cannot be considered simply a "women's issue" (Joshi et al., 2015). In light of the causes of the recent financial crisis, the question can be raised of whether a more balanced representation of women, in industrial and financial firms, would have avoided corporate scandals and poor performance (Adams and Funk, 2012). Recent institutional policy interventions, such as gender quotas in corporate leading positions, show clear proof that gender equality and discrimination are now strategic issues.These reasons contributed to the success of microfinance institutions (MFIs) that are considered women paladins in the financial industry giving them access to credit. MFIs are a valid alternative to a conventional bank when providing financing to disadvantaged and poor people (Yunus et al., 2010; Quayes, 2012). Through smaller loans (outreach depth) granted to as many clients as possible (outreach breadth), MFIs promise to fight poverty, contribute to pro-poor growth (Mayoux, 2010) and to the empowerment of women who are the natural borrowers of these organizations (Strom et al., 2014). This evidence is driven by at least four reasons. First, in poor countries, women are poorer than men (Agier and Szafarz, 2013) and consequently they are natural candidates for MFIs to the point that their percentage on total borrowers has been used as a proxy of outreach depth (Mori et al., 2015). Second, women are usually considered credit constrained (Fletschner, 2008; 2009), investing in small-scale projects that are unattractive to conventional banks (Armendariz and Morduch, 2010). Third, capital access to women ensures more social (Aggarwal et al., 2015) and economic (Fletschner, 2008) benefits to their households than would occur if borrowers were men. Finally, women are better re-payers, enhancing the persistence of the usually fragile business model of the MFIs (Boehe and Cruz, 2013; D'Espallier et al., 2011).While women are the main target of MFIs and their empowerment has become crucial in microfinance, a still unanswered question is whether they face gender discrimination in credit conditions. In fact, in an attempt to grant smaller loans to as many clients as possible, credit rationing could potentially damage women, moving MFIs away from their mission of improving the condition of women and promoting their rights. Although gender discrimination cannot be treated as a marginal issue in microfinance (Mayoux, 2010), there are few studies on the topic and evidence ispuzzling, to say the least. For example, Storey (2002) does not find any denial rate divergence between men and women, as well as Corsi and De Angelis (2017) who find no evidence of gender discrimination against women borrowers in loan size. Agier and Szafarz (2013) document that women are discriminated in the size of loan, while Sagamba et al. (2013) observe a slight preference for female applicants.Our paper provides a perspective with the potential for accommodating the divergences found in the literature. We investigate the discrimination suffered by women borrowers, looking at the effect enhanced by women lenders in mitigating the phenomenon. The main idea is straightforward and fairly intuitive. We expect that the percentage of women employed in the MFIs shrinks the effect of smaller loans granted to women on the demand side for a sort of solidarity in sharing the "same traits." The failure to consider this aspect could be the reason for the mixed findings documented so far. To investigate this issue, we discriminate MFIs based on the roles of women inside the organization: top roles (CEO or board members) and operational roles (managers, officers, and staff). As we will show, a study which abstains from considering this distinction would be missing any robust explanatory power. …