Rawls' difference principle, self-help group, financial inclusion and social cohesion—lore or actuality? Experience of Central Assam

The paper examines whether SHG–Bank Linkage Programme (SBLP)-led microfinance in Central Assam advances its twin goals: financial inclusion and the broader social objective aligned with Rawls’ Difference Principle—social cohesion for the least-advantaged. Against the backdrop of widespread financial illiteracy among women and the prominence of SHGs in India’s inclusion agenda, the study asks if SBLP participation not only enhances access to formal finance but also translates into social inclusion across economic, public service, and civic-political dimensions. The authors articulate a central hypothesis: SHG-led microfinance fails to corroborate social cohesion for the least-advantaged. The study’s purpose is to provide region-specific evidence from three districts in Central Assam where SBLPs are numerous and populations diverse in religion, caste, and class, thereby testing the theoretical promise of Rawls’ Difference Principle embedded in microfinance policy (Aajeevika under DAY-NRLM).

Prior studies widely credit SHGs and microfinance with facilitating self-employment, women’s empowerment, and poverty alleviation, though evidence is mixed on poverty reduction and comprehensive inclusion. Many works find positive links between SHG participation and financial inclusion, entrepreneurship, and social capital; others note partial success or low coverage of the very poorest, questioning poverty alleviation potential. Financial literacy is especially low among women and is associated with empowerment. SHG loans can substitute informal credit and help mitigate shocks. Only two recent studies (Maity 2023a, 2023b) explicitly examine social cohesion and financial inclusion, reporting that SHG participation can enhance both, yet results may depend on context, sample size, and participant characteristics. The authors identify a gap in social cohesion outcomes of SBLP across diverse settings and propose to test: Ho: SHG-led microfinance fails to corroborate social cohesion for the least-advantaged group in society. The study extends the literature by jointly testing financial inclusion and social cohesion in a region with pronounced socio-religious and caste heterogeneity and high SBLP penetration.

Conceptual framework: The study is grounded in Rawls’ Theory of Justice (1971), specifically the Difference Principle—maximizing benefits to the least-advantaged—which underpins microfinance and India’s DAY-NRLM (Aajeevika). Social cohesion is treated as the ultimate goal alongside financial inclusion. Data and variables: The analysis uses primary survey data from Central Assam (Nagaon, Morigaon, Hojai), supplemented with secondary administrative statistics. Primary data were collected April 2019–January 2020 from 825 respondents (335 SBLP participants; 490 non-participants). Socioeconomic covariates include age, education (years), agricultural landholding (bigha), caste (SC/ST dummy), religion (Hindu dummy), distance to nearest bank (km), and monthly household consumption expenditure (Rs). Outcome indices are self-constructed: Financial Inclusion Index (FII) and Social Cohesion Index (SCI). FII aggregates access/usage of formal credit, savings, insurance, transactions, knowledge, and SHG operations (modalities in Appendix Table 8). Social cohesion is derived as 1 − Social Banishment Index (SBI), where SBI aggregates three domains of exclusion—economic, public services, civic/political participation—via multiple dichotomous indicators (Appendix Table 9). Weights for binary indicators are obtained using Multiple Correspondence Analysis (MCA), avoiding subjective weights and acknowledging PCA’s inapplicability to binary data. Sampling: Multistage stratified random sampling: Stage 1 selects three districts. Stage 2 selects intensive blocks (two from Nagaon; all from Hojai and Morigaon). Stage 3 determines SHG sample size using Krejcie and Morgan’s formula. Stage 4 determines participant interviews; control group is twice the target initially. Participants are defined as members who received SBLP loans at least two years prior; controls formed SHGs but had not received loans at survey time. Estimation strategy: Participation equation estimated via Probit to obtain propensity scores P(Z)=Φ(βZ) using the listed covariates. Common support is verified (propensity score range 0.239–0.672; mean 0.407). Matching is performed “without replacement,” yielding 335 treated matched to 335 controls (155 controls dropped). Balance checks show substantial reduction in bias and low post-matching pseudo-R² (0.083) with non-significant LR chi², indicating good balance. Treatment effects are estimated as Average Treatment effect on the Treated (ATT) using Nearest Neighbor Matching (NNM) and Kernel Matching (KM) for robustness. Sensitivity to hidden bias is assessed via Mantel–Haenszel bounds at Γ=1.

• Determinants of SBLP participation (Probit): Model fit is strong (pseudo-R²≈0.428; LR χ² significant). Higher age and longer bank distance reduce participation; education, Hindu religion, SC/ST caste status, and higher consumption expenditure increase participation. Agricultural landholding is not significant in the reported Probit table.
• Propensity score: Range 0.239–0.672; mean 0.407; adequate overlap and balance achieved post-matching (post-match pseudo-R²=0.083; LR χ² p=0.966).
• Financial inclusion (FII): ATT (NNM): Treated 0.450 vs Controls 0.387; Difference 0.063; SE 0.024; t=2.63 (significant at 1%). ATT (Kernel): Treated 0.450 vs Controls 0.385; Difference 0.065; SE 0.021; t=3.07 (significant at 1%). Conclusion: SBLP participation significantly improves financial inclusion by about 6–7 percentage points.
• Social cohesion (SCI): ATT (NNM): Treated 0.662 vs Controls 0.437; Difference 0.225; SE 0.237; t=0.952 (not significant). ATT (Kernel): Treated 0.662 vs Controls 0.447; Difference 0.215; SE 0.159; t=1.352 (not significant). Conclusion: No statistically significant effect of SBLP participation on social cohesion.
• Sensitivity (Mantel–Haenszel bounds, Γ=1): Results indicate no hidden bias undermining main inferences. Overall: The study accepts the null hypothesis that SBLP fails to corroborate social cohesion for the least-advantaged, while confirming gains in financial inclusion.

While SBLP participation clearly fosters financial inclusion—through bank account ownership, loan access/repayment discipline, use of transaction services, and growing financial literacy—these gains do not translate into statistically significant improvements in social cohesion. The paper attributes this to entrenched social structures and norms in rural India: patriarchal control over women’s financial decisions; persistent caste and religious boundaries limiting inter-group interaction; and political-religious influences affecting civic and political participation. Group formation often occurs through personal familiarity rather than inclusive processes supervised by Panchayats, potentially reinforcing homophily and limiting cross-caste or cross-religion engagement. Consequently, even as women gain financial practices and access, broader dimensions of social inclusion—economic security, public service access, and civic/political participation—remain constrained. Contextual heterogeneity in Assam, sample composition, and local implementation practices may explain divergence from prior studies that found positive social inclusion effects.

The study contributes novel, region-specific evidence on whether SBLP, grounded in Rawls’ Difference Principle and implemented under DAY-NRLM (Aajeevika), achieves its twin aims. It finds robust, positive impacts on financial inclusion but no statistically significant improvement in social cohesion among participants in Central Assam. Policy implications include: scaling SBLP enrollment under DAY-NRLM; prioritizing outreach and preparatory training (potentially via NGOs) to improve loan use; considering complementary MFI presence; and, crucially, reforming group formation to ensure inclusive participation across religion, caste, and class. Panchayats should oversee transparent, inclusive group formation, and NGOs can provide counseling to build inter-group trust. Future research should expand geographic scope, increase sample sizes, and adopt dynamic frameworks (e.g., panel data) capturing group maturity and participation duration to assess long-term social cohesion effects.

• Geographic scope limited to three districts in Central Assam; findings may not generalize to all of Assam or India.
• Cross-sectional primary data; absence of panel/dynamic measures limits assessment of long-run social cohesion.
• Lack of data on group maturity and duration of participation restricts exploration of dynamic effects.
• Potential selection on unobservables mitigated via PSM and MH bounds, but residual bias cannot be entirely ruled out.
• Differences in local implementation, cultural heterogeneity, and sampling constraints may influence external validity.