What explains why grants to women-owned microenterprises often fail to yield improvements in business outcomes, especially as compared to grants to men’s microenterprises? This is the topic tackled in the new working paper “Mitigating the Impact of Household Expropriation on Female Entrepreneurship: Experimental Evidence from Ghana,” by Francisco Campos, Adriana Conconi, Elwyn Davies, Marine Gassier, and Markus Goldstein. They look within households to understand the role of “expropriation” of resources (also referred to in the paper as intra-household sharing pressure) which diverts these grants away from their intended use.
This RCT studies four alternative interventions: [a] an unconditional grant provided through a mobile money account (=2 months of median profits), [b] the unconditional grant disbursed to the female entrepreneurs’ spouses; [c] the grant conditional on participating with their spouses in a training on joint decision-making; and [d] the grant conditional on the woman reaching a savings goal (a third of the grant amount) with a dedicated bank account over 6 months. The study population is a sample of married/co-habiting women business owners, average age 39, in Accra, Ghana. About half of these women are selling goods, 22% in food transformation or restaurants, 12% in the textile sector (tailors I assume), and 11% in beauty services. These are very clearly microenterprises. The mean number of paid employees in these businesses is 0.2 (0.5 workers when counting paid and family workers, in addition to the women themselves). The sample are women and their husbands who were willing to participate in the study, drawn from a random walk in over 600 enumeration areas which identified women-owned businesses. There are about 3,100 women (along with their husbands) in the baseline. Take-up for treatment [c], completing the couples training and getting a grant, was 76%; take-up for treatment (d), the savings program with a grant incentive, was 90%.
A short list of some of the findings 1.5 years later: the unconditional grants do not significantly impact business sales and profits for women or men; the grant plus couples decision-making intervention [c] increased husbands’ support for the women’s businesses but had no impact on her business’ outcomes; the grant plus 6-months own savings resulted in a 15 percent increase in sales and a 10 percent increase in profits. This last result is even larger for women who reported greater intra-household sharing pressure at baseline. They find that self-control, liquidity constraints, and access to savings options do not explain these impacts. And a key conclusion (and last sentence of the abstract) is: “The findings substantiate that intrahousehold dynamics matter for women’s investment decisions, and highlight the importance of promoting autonomy in the face of expropriation pressures, for increased growth and investment.”
On a personal level, I often find that blogging about one paper, inevitably leads me to looking up lots of other papers. And I swirl around and struggle with what to focus on in the blog (trying to get beyond just a summary of the paper). That was especially the case with this paper. Eventually, after a lot of false starts and meandering, I landed on three things that I will focus on here. Enjoy!
Beyond the quant
In addition to reading the Campos et al. paper, I encourage you to read an in-depth qualitative study in World Development: “Competing priorities: Women’s microenterprises and household relationships,” by Sophia Friedson-Ridenour and Rachael S. Pierotti, which was conducted with women entrepreneurs from the pilot sample of the RCT. If only every impact evaluation had such useful work to inform RCT design! They unpack the details around the idea that “women and their partners generally had uncooperative financial management arrangements” and the many nuances related to how a grant for the woman’s business might get re-directed. I suggest making much (much) more use of this paper in the RCT paper. Friedson-Ridenour and Pierotti provide strong and compelling rationale for treatment arms [b], [c], and [d]. My hunch (confirmed informally) is that this intentional and well-designed qual work was indeed used to inform the RCT design, and so let’s acknowledge this important imput. My nudge to Campos and coauthors: tell us more about this link, draw on the qual study in your paper!
Is it intrahousehold dynamics, or is it that women lack of self-control?
Fafchamps et al. (2014) study a very similar population of microenterprises in Accra (with a sample of both men and women-owned businesses). And they come to a very different conclusion on why cash grants do not have impacts on business outcomes for women microenterprise owners (and moderate impacts for male counterparts). They conclude that these business women owners lack self-control rather than the constraints they face from external pressures. This is especially (or mainly) the case for high-profit women. Among these women, those with low self-control do not yield improved business outcomes from the cash grant, whereas the in-kind grants do grow firm profits.
Clearly this is a very different finding for the same context (albeit a decade apart). And worthy of some exploration in the new study.
But… how do you measure self-control?
And this then lead me down a deep rabbit hole with both papers, especially in regards to the term and concept of self-control. Bottom line: I really (really really) dislike that term as applied (and measured) in this context.
Campos et al. capture self-control with an indicator for having an above-the-median discount rate based on five indicators of time preferences. In Fafchamps et al. (2014), self-control is measured drawing on four components: whether the respondent used a susu collector (informal mobile banker, resulting in saving at negative interest rate in exchange for safekeeping) as an indicator or having self-control issues; whether they agree with the statement “I save regularly”; having a high discount rate (based on a standard hypothetical discounting question about waiting a month for extra cash); and a standard measure of hyperbolicity, based on whether their preferences switch to be more patient when asked to choose between amounts at 5 and 6 months. In related work, Riley (2024) also measures self-control based on time preferences in her study of women businesses in Uganda.
After reading in Friedson-Ridenour and Pierotti about the complex decision-making facing these women business owners (often related to their intrahousehold circumstances) in regards to finances, investments, businesses, and savings, it is hard to see how any of the above measures tell me much about self-control, at least as defined when I googled it (among other phrases “In economics, self-control is the capacity to override immediate impulses in favor of long-term goals”). The complex conditions facing women suggest to me that savings behaviors and time preferences capture far more than the one-dimensional notion of impulse control: Yikes! That is not to dismiss the idea that impulse control/self-control matters for them and for all of us – I am a firm believer in the value of commitment devices to keep me in check in terms of my own self-control. I also think all of us probably could “be better” with more self-control. But it strikes me as reductive, naïve, and misguided to say that we know anything about self-control in the way it is measured for a sample of women for whom a lot of aspects of their lives – including business decisions – seem partially or very much out of their control.
One might say that this is all semantics and tell me not to get so worked up. But I think that words matter. Self-control, to me, points to one’s internal psychological state (‘if only she had more self-control’), and downplays significant external realities that influence decision-making. If we mismeasure self-control, we risk ignoring the influence of these realities.
Source : World Bank
