Meager 2019 - Bayesian Hierarchical Analysis of Microcredit

Routing Summary

This folder fully captures Meager (2019): Bayesian hierarchical aggregation of seven microcredit RCTs - models, priors, pooling metrics, average effects, heterogeneity, and covariate analyses. It contains 9 notes covering Sections I-VI, Appendix A, Tables 1-5 and Figures 1-10.

Concept Map

ConceptNoteTypeDepends OnKey Result
Studies and outcomesMeager 2019 - Studies, Data and Outcomesconcept-7 RCTs; USD PPP per fortnight; ITT of access
Rubin model (eq. 1)Meager 2019 - Hierarchical Model Specificationsdefinition-,
Full-data and covariate models (eq. 2-4)Meager 2019 - Hierarchical Model Specificationsdefinitioneq. 1control means, subgroup and site covariates
ExchangeabilityMeager 2019 - Hierarchical Model Specificationsdefinitioneq. 1 applies only to exchangeable sites
Priors (eq. 5), HMCMeager 2019 - Bayesian Estimation, Priors and Computationconcepteq. 2LKJ(3) correlation, half-Cauchy(0,10) scale;
Pooling factors (eq. 6-8)Meager 2019 - Pooling Factors and Measuring External Validitydefinitioneq. 1omega, omega-breve, lambda in [0,1]
Table 2 average effectsMeager 2019 - Average Treatment Effects - Resultsexampleeq. 2profit 6.8, revenues 14.5; all 95% intervals include 0
Tables 3-4; predictiveMeager 2019 - Heterogeneity, Pooling Results and Predictive Distributionsexampleeq. 6-8about 60% sampling variation; 25% adverse sign
Covariates and ridgeMeager 2019 - Covariates - Prior Business Experience and Site-Level Ridgeexampleeq. 3-4prior business: larger, more heterogeneous; APR, loan size, women
Conclusion and marketing mappingMeager 2019 - Conclusion, Limitations and Marketing Analoguesconceptallmoderate external validity

Notes

Sources

  • Meager (2019), AEJ: Applied Economics 11(1):57-91, doi:10.1257/app.20170299 - Meager 2019 - Understanding the Average Impact of Microcredit Expansions (LSE Research Online published version).

See Also