https://orcid.org/0000-0002-1391-1714
![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
A multitude of NGOs form informal savings and lending associations in developing countries, called savings groups. Group members are responsible for the day-to-day governance, operating within the framework provided by the NGOs. As NGOs differ in the model they provide, savings groups are endowed with different levels of autonomy over their operations. Drawing on theories related to autonomy and performance, we test whether more autonomous savings groups perform better. A two-way correlated random effects regression on data for 97,653 savings groups formed by five NGOs in 16 countries shows that savings groups with more autonomy over their financial operations demonstrate greater financial performance. Additionally, the financial performance of more autonomous savings groups improves more in subsequent cycles, suggesting that savings groups with more autonomy have a greater ability to adapt. We argue that groups thus benefit from being granted more autonomy over their day-to-day operations. Our study may have broader implications for similar development interventions by NGOs, in which participants could benefit from increased autonomy.
Acknowledgement
We are grateful to the editor and two anonymous referees for the valuable feedback they provided.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability
The data that support the findings of this study are available from the corresponding author upon request.
Notes
1 In other contexts sometimes referred to as accumulating savings and credit associations (ASCAs) or village lending and savings associations (VSLAs). These are distinct from other institutions sometimes referred to as savings groups, such as rotating savings and credit associations (ROSCAs).
2 NGOs often note that the SGs they form are, in principle, largely autonomous, and can decide on their own rules and regulations. However, the manuals and other training tools they use to form the groups illustrate that NGOs steer their groups differently, some more strictly than others.
3 See Mersland, D’Espallier, Gonzales, and Nakato (Citation2019) for an elaborate description of the database.
4 Whereas a different participant of the ROSCA receives the pot each meeting, the SG allows members to request loans during the cycle. Loans are reimbursed with interest, which accumulates over time, making the SGs under study more similar to ASCAs than ROSCAs.
5 Some SGs may have as few as 15 or as many as 30 members. Different (grassroots) groups or institutions (for example ROSCAs and self-help groups), may be much larger and meet either more or less frequently.
6 Newly formed SGs increasingly store their funds digitally, and some groups may be linked to formal financial institutions, storing their funds using a deposit account.
7 In cases where all members contribute an equal amount per week, they all receive the same share-out. This, however, is still a proportional payment, as all have an equal proportion.
8 In our data, only very few members leave the group during this time.
9 The share-out is weighed by savings contributed, regardless of timing. Early contributions can be lent out multiple times and thus generate more profit. Saving more later in the cycle generates less profit, but allows members to capture profit while losing access to their funds for a relatively short time.
10 Present-biased individuals place a higher value on money in the present, while time-consistent participants maintain consistent valuation over time.
11 These indicators are winsorized on both sides at 2 per cent to remove outliers. Savings, credit, and loans per members are only winsorized at the right-hand side, as their minimum, both theoretically and in the data is 0.
12 Such a period in which borrowing is not allowed is uninformative about SG-performance which is why we treat this as an outlying observation.
13 Complete freedom to set interest rates can either be supplemented by a suggested range or left completely open.
14 Due to the relatively low number of NGOs, principal component analysis is not feasible.
15 Tables 1–5 in the Supplementary Materials provide a robustness check through running the same model while removing an individual NGO.
16 For robustness, Tables 6–13 in the Supplementary Materials perform the same model for the continuous index, as well as the continuous index after removing individual autonomy dimensions. The largest difference occurs when removing interest rate freedom (Table 9), likely due to the heterogeneity in interest rate policies and the importance of interest rates in balancing out returns and fund-utilization.
17 For example, the development of the rate of female participants over time or whether groups decide to link to a formal financial institution, may depend on specific unobserved group-level characteristics.
18 Table 14 in the Supplementary Materials reports our outcomes after replacing the cycle dummies with a single dummy for either the first or subsequent cycles, which represents the main effect of the time dimension to our interaction.
19 This story remains consistent when the autonomy index is considered as a continuous variable. Using a pooled-OLS without CRE, the results are qualitatively equivalent as well, except for the interaction effect for loans per member, ROS, and fund-utilization, which become insignificant.
20 This analysis is performed including the cycle dummy, which represents the main effect of time. A table with these effects reported can be found in Table 15 of the Supplementary Materials.