Model determination for high-dimensional longitudinal data with missing observations: an application to microfinance data

Fall F.S., Tchakoute Tchuigoua H., Vanhems A., Simar L.

The main objective of this study is to assess the impact of unobserved heterogeneity on microfinance social efficiency analysis. Based on recent nonparametric techniques and directional distances, we identify a latent heterogeneity factor related to the microfinance institute (MFI) manager’s ability to promote women, independent of MFI size. We test for the significance of this unobserved factor and analyze the impact of MFI social inefficiency measures. Using a cross-country sample of 501 MFIs in 2011 from six main regions of the world, our findings reveal a significant effect of unobserved heterogeneity on the frontier and hence stress the importance of subjective factors in defining the set of production possibilities. We assess the robustness of our findings with the considered profit-oriented status and analyze the link between our unobserved heterogeneity factor and institutional and socioeconomic indicators.

Quayes S., Joseph G.

The paper analyzes the determinants of social outreach of microfinance institutions (MFIs), using three measures of outreach – depth of outreach, breadth of outreach, and outreach to women, and its...

Buera F.J., Kaboski J.P., Shin Y.

Abstract What is the aggregate and distributional impact of microfinance? To answer this question, we develop a quantitative macroeconomic framework of entrepreneurship and financial frictions in which microfinance is modelled as guaranteed small-size loans. We discipline and validate our model using recent empirical evaluations of small-scale microfinance programs. We find that the long-run general equilibrium impact is substantially different from the short-run effect. In the short-run partial equilibrium, output and capital increase with microfinance but total factor productivity (TFP) falls. In the long run, when general equilibrium effects are considered, as should be for economy-wide microfinance interventions, scaling up microfinance has only a small impact on per-capita income, because an increase in TFP is offset by lower capital accumulation. However, the vast majority of the population benefits from microfinance directly and indirectly. The welfare gains are larger for the poor and the marginal entrepreneurs, although higher interest rates in general equilibrium tilt the gains toward the rich.

Efron B., Narasimhan B.

Awaworyi Churchill S.

The financial sustainability of microfinance institutions (MFIs) is crucial for the continual existence of the microfinance industry. As a result, emphasis has been placed on the financial sustainability of MFIs over the past few years. However, with the primary goal of the industry being the attainment of social outreach, the emphasis on financial sustainability has raised concerns about a potential adverse effect on outreach. Using data on 1595 MFIs in 109 countries, we examine if there is a trade-off between financial sustainability and outreach. The evidence shows that there is a trade-off between sustainability and outreach depth, but complementarity between sustainability and outreach breadth.

Hermes N., Hudon M.

Microfinance institutions (MFIs) generally aim at improving the access of the poor to financial services while at the same time being financially sustainable. But what do we know about how MFIs reach and combine these two goals? We carry out a systematic review of close to 170 papers discussing the determinants of the financial and social performance of MFIs. The review shows that the most important determinants addressed in the literature are MFI characteristics (size, age and type of organization), their funding sources, the quality of organizational governance and the MFIs' external context such as macro-economic, institutional and political conditions. The evidence on these issues is rather mixed. Moreover, the direction of the relationship between these drivers and MFI performance depends on the context, particularly the country-specific context. Finally, there is a lack of consensus in the literature on the measurement of financial and social performance. Due to the complexity of the concept, we argue that social performance should only be assessed by using a multidimensional perspective. This can be done either by applying recent and holistic social performance measures such as the SPI4, or at least by using a combination of proxies, such as outreach, gender and rural measures.

Zhao Y., Long Q.

Variable selection plays an essential role in regression analysis as it identifies important variables that associated with outcomes and is known to improve predictive accuracy of resulting models. Variable selection methods have been widely investigated for fully observed data. However, in the presence of missing data, methods for variable selection need to be carefully designed to account for missing data mechanisms and statistical techniques used for handling missing data. Since imputation is arguably the most popular method for handling missing data due to its ease of use, statistical methods for variable selection that are combined with imputation are of particular interest. These methods, valid used under the assumptions of missing at random (MAR) and missing completely at random (MCAR), largely fall into three general strategies. The first strategy applies existing variable selection methods to each imputed dataset and then combine variable selection results across all imputed datasets. The second strategy applies existing variable selection methods to stacked imputed datasets. The third variable selection strategy combines resampling techniques such as bootstrap with imputation. Despite recent advances, this area remains under-developed and offers fertile ground for further research.

Belloni A., Chernozhukov V., Hansen C., Kozbur D.

We consider estimation and inference in panel data models with additive unobserved individual specific heterogeneity in a high-dimensional setting. The setting allows the number of time-varying regressors to be larger than the sample size. To make informative estimation and inference feasible, we require that the overall contribution of the time-varying variables after eliminating the individual specific heterogeneity can be captured by a relatively small number of the available variables whose identities are unknown. This restriction allows the problem of estimation to proceed as a variable selection problem. Importantly, we treat the individual specific heterogeneity as fixed effects which allows this heterogeneity to be related to the observed time-varying variables in an unspecified way and allows that this heterogeneity may differ for all individuals. Within this framework, we provide procedures that give uniformly valid inference over a fixed subset of parameters in the canonical linear fixed effect...

Liu Y., Wang Y., Feng Y., Wall M.M.

We propose a Multiple Imputation Random Lasso (mirl) method to select important variables and to predict the outcome for an epidemiological study of Eating and Activity in Teens. In this study 80% of individuals have at least one variable missing. Therefore, using variable selection methods developed for complete data after listwise deletion substantially reduces prediction power. Recent work on prediction models in the presence of incomplete data cannot adequately account for large numbers of variables with arbitrary missing patterns. We propose MIRL to combine penalized regression techniques with multiple imputation and stability selection. Extensive simulation studies are conducted to compare MIRL with several alternatives. MIRL outperforms other methods in high-dimensional scenarios in terms of both reduced prediction error and improved variable selection performance, and it has greater advantage when the correlation among variables is high and missing proportion is high. MIRL is shown to have improved performance when comparing with other applicable methods when applied to the study of Eating and Activity in Teens for the boys and girls separately, and to a subgroup of low social economic status (ses) Asian boys who are at high risk of developing obesity.

Zhou H., Elliott M.R., Raghunathan T.E.

Multiple imputation (MI) is a well-established method to handle item-nonresponse in sample surveys. Survey data obtained from complex sampling designs often involve features that include unequal probability of selection. MI requires imputation to be congenial, that is, for the imputations to come from a Bayesian predictive distribution and for the observed and complete data estimator to equal the posterior mean given the observed or complete data, and similarly for the observed and complete variance estimator to equal the posterior variance given the observed or complete data; more colloquially, the analyst and imputer make similar modeling assumptions. Yet multiply imputed data sets from complex sample designs with unequal sampling weights are typically imputed under simple random sampling assumptions and then analyzed using methods that account for the sampling weights. This is a setting in which the analyst assumes more than the imputer, which can led to biased estimates and anti-conservative inference. Less commonly used alternatives such as including case weights as predictors in the imputation model typically require interaction terms for more complex estimators such as regression coefficients, and can be vulnerable to model misspecification and difficult to implement. We develop a simple two-step MI framework that accounts for sampling weights using a weighted finite population Bayesian bootstrap method to validly impute the whole population (including item nonresponse) from the observed data. In the second step, having generated posterior predictive distributions of the entire population, we use standard IID imputation to handle the item nonresponse. Simulation results show that the proposed method has good frequentist properties and is robust to model misspecification compared to alternative approaches. We apply the proposed method to accommodate missing data in the Behavioral Risk Factor Surveillance System when estimating means and parameters of regression models.

Buuren S.V., Groothuis-Oudshoorn K.

The R package mice imputes incomplete multivariate data by chained equations. The software mice 1.0 appeared in the year 2000 as an S-PLUS library, and in 2001 as an R package. mice 1.0 introduced predictor selection, passive imputation and automatic pooling. This article documents mice, which extends the functionality of mice 1.0 in several ways. In mice, the analysis of imputed data is made completely general, whereas the range of models under which pooling works is substantially extended. mice adds new functionality for imputing multilevel data, automatic predictor selection, data handling, post-processing imputed values, specialized pooling routines, model selection tools, and diagnostic graphs. Imputation of categorical data is improved in order to bypass problems caused by perfect prediction. Special attention is paid to transformations, sum scores, indices and interactions using passive imputation, and to the proper setup of the predictor matrix. mice can be downloaded from the Comprehensive R Archive Network. This article provides a hands-on, stepwise approach to solve applied incomplete data problems.

Basharat B., Hudon M., Nawaz A.

Pricing is a central strategic decision for all companies, and is particularly sensitive for social enterprises with both financial and social objectives. High interest rates in microfinance are a topic of intense debate. Using an original database of 291 MFIs, this paper provides empirical evidence of the impact the efficiency of an MFI has on its microcredit interest rate. We use the non-parametric Data Envelopment Analysis (DEA) framework to calculate efficiency and differentiate financial and social efficiency. The results show that financial efficiency has a positive impact on interest rates, with more financially efficient MFIs having lower interest rates, while social efficiency has no impact on microcredit interest rates.

Quayes S.

Using a panel of 764 microfinance institutions (MFIs) from 87 countries, this study analyses the possible trade-off between outreach and performance and shows that greater depth of outreach has a positive impact on the financial performance of an MFI. The empirical results of this study should dispel the widely held apprehension that the recent emphasis on attainment of financial sustainability by the MFIs could seriously impair their outreach efforts and shows that outreach to the poor can actually bolster financial performance.

Banerjee A., Duflo E., Glennerster R., Kinnan C.

This paper reports results from the randomized evaluation of a group-lending microcredit program in Hyderabad, India. A lender worked in 52 randomly selected neighborhoods, leading to an 8.4 percentage point increase in takeup of microcredit. Small business investment and profits of preexisting businesses increased, but consumption did not significantly increase. Durable goods expenditure increased, while “temptation goods” expenditure declined. We found no significant changes in health, education, or women's empowerment. Two years later, after control areas had gained access to microcredit but households in treatment area had borrowed for longer and in larger amounts, very few significant differences persist. (JEL G21, G31, O16, O12, L25, I38)

Berge L.I., Bjorvatn K., Tungodden B.

Microenterprises constitute an important source of employment, and developing such enterprises is a key policy concern in most countries. But what is the most efficient tool for microenterprise development? We study this question in a developing country context (Tanzania), where microenterprises are the source of employment for more than half of the labor force, and we report from a field experiment that jointly investigated the importance of a human capital intervention (business training) and a financial capital intervention (business grant). Using data from three survey rounds, a lab experiment, and administrative records of the microfinance institution, we present evidence on business performance, management practices, happiness, business knowledge, and noncognitive abilities. Our study demonstrates strong effects of the combination of the two interventions on male entrepreneurs, while the effects on female entrepreneurs are much more muted. The results suggest that long-term finance is an important constraint for microfinance entrepreneurs, but that business training is essential to transform financial capital into productive investments. Our study also points to the need for more comprehensive measures to promote the businesses of female entrepreneurs. Data, as supplemental material, are available at http://dx.doi.org/10.1287/mnsc.2014.1933 . This paper was accepted by John List, behavioral economics.