Microequity for women led microenterprises - Experimental evidence from India

Working with Buzz Women, an NGO, and Yunus Social Business (YSB), a financing institution, we are running a field experiment that offers microequity contracts in two sectors of particular relevance for rural and peri-rural areas: (i) animal husbandry and (ii) tailoring. By randomizing access to microequity contracts, we will quantify the impact of microequity contracts on business and socioeconomic outcomes by comparing the microenterprises that were offered the contract to those that were not.

Microequity contracts combine innovative financial contracting with technology. The idea behind such contracts is that they allow for performance-contingent repayments that provide implicit insurance against risks. Microequity contracts provide a ``more natural” and more continuous process of mutual sharing of losses and profits. Unlike a standard insurance product, the downside risk sharing is less ``binary” (either paying out if some event is triggered or not). This more regular and continuous sharing of downside risk also has the potential to be more valued by clients, given the significant evidence of extremely low take-up rates for conventional microinsurance even when heavily subsidized. We hypothesize this can effectively stimulate growth of microenterprises beyond what we see in traditional microcredit interventions.

Working with Buzz Women, an NGO, and Yunus Social Business (YSB), a financing institution, we are running a field experiment that offers microequity contracts in two sectors of particular relevance for rural and peri-rural areas: (i) animal husbandry and (ii) tailoring. By randomizing access to microequity contracts, we will quantify the impact of microequity contracts on business and socioeconomic outcomes by comparing the microenterprises that were offered the contract to those that were not.

Through ongoing discussions with Buzz Women, we have identified several kinds of assets that have high returns but –- because of their lumpiness –- are beyond the financial reach of many microfinance clients. Specifically, for clients working in animal husbandry, this is true of purchases of cows (with accompanying cow feed); for clients working in tailoring, this is true of sewing and embroidery machines (with accompanying textile inventories). On average, we anticipate these assets each costing approximately Rs.100,000 (USD 1300). We have designed a microequity contract that facilitates the purchase of these assets through a performance-contingent repayment schedule. Monthly repayments under the equity contract will be calculated based on self-reported revenues; these self reports will then be verified using several monitoring technologies. Specifically, entrepreneurs can choose to self-report revenues using either (i) a digital app (called ‘MeraBills’), or (ii) via text messages to their respective field officers. Then –- with the prior consent of participants –- we provide credible monitoring mechanisms to guard against systematic under-reporting. For those in tailoring, we will use motion sensors attached to tailoring machines; these monitor the hours of operation each day. For those engaged in animal husbandry, entrepreneurs will be required to submit receipts that they receive from the dairy, on a weekly basis. In addition, for both sectors we will send (randomly assigned) `secret shoppers' to clients' workplaces (and clients will know in advance that this is a possibility). The revenue earned from the transaction with the secret shopper should be reflected in that day's revenue report.

2022-07-06

2025-07-15

Firm revenues, expenditure and profit; Repayment data

Daily self-reported data on revenues, expenditure and profit are aggregated on a weekly basis; Repayment data from YSB is collected on a monthly basis

Compliance, work days, household consumption, employment, fixed capital investments, missed reporting

Compliance - Weekly measure of compliance on daily reports;
Work days - Number of days worked per week collected using daily self reported data;
Household consumption - Calculated using 30 day recall from quarterly phone surveys
Employment -Number of workers employed from quarterly phone surveys
Fixed capital investments - For HH or business from quarterly phone surveys
Missed reporting - Difference between self-reported revenue and monitoring revenue

In this study, we have two groups - a treatment group that is offered a microequity contract and a control group that is offered no funding (they have business as usual access to regular loans). We have designed a microequity contract that facilitates the purchase of these assets through a performance-contingent repayment schedule. Monthly repayments under the equity contract will be calculated based on self-reported revenues; these self reports will then be verified using several monitoring technologies. Using daily self reported data, quarterly phone surveys and two in person follow-up surveys we study the impact of these contracts on our key outcomes of interest.

This pre-analysis plan was registered shortly after the loans were disbursed. At the time of writing this pre-analysis plan we were blinded to any post-intervention data.

Individual level randomization. For entrepreneurs engaged in animal husbandry we randomise with stratification. The sample is stratified based on median revenue in the three months before contract signing. Within tailoring, the sample is very small and therefore, no stratification was done.

Individual

No

28 entrepreneurs

28 entrepreneurs

15 entrepreneurs in the treatment group and 13 entrepreneurs in the control group

For variables where we have daily pre-intervention data, we present power calculations based on simulated data. We have collected over one year of pre-intervention daily data (on revenue, expenditure and profit) for all participants in the study. We use this pre-intervention data to simulate a 50% increase in revenue due to the intervention over a period of three years. Based on simulated post intervention data, we check for power by estimating the primary ANCOVA specification using OLS and clustered poisson and conducting inference by randomization inference. We cluster at the firm level, and re sample according to the underlying stratification scheme. Despite the small sample size, we are powered to detect treatment effects equivalent to a 50\% increase in revenue (which is a reasonable effect size given the large capital injection). For variables collected using the quarterly phone surveys, preliminary power calculations suggest that with full take-up, a sample size of 30 is powered (at 77\%) to detect effect sizes of greater than 0.45sd.