Replication Package for "Do Private Consultants Promote Savings and Investments in Rural Mozambique?"
This repository contains the data and codes that replicate the figures and tables presented in the paper "Do Private Consultants Promote Savings and Investments in Rural Mozambique?" by Paul Christian, Steven Glover, Florence Kondylis, Valerie Mueller, Matteo Ruzzante and Astrid Zwager, published in Agricultural Economics (Journal of the International Association of Agricultural Economists), January 2022.
[Abstract], [Published Paper], [Online Appendix], [BibTeX]
Advice from management professionals can help small- and medium-sized firms reach complex financial goals in the Global South. We apply lessons learned in the firm literature to determine the degree in which farmer associations face constraints to management and planning capacity that can be alleviated by the provision of advice from external consultants. In particular, we conducted a randomized control trial in 42 water user associations (WUAs) in Mozambique to examine whether more intensive attention from financial consultants through repeated follow-up visits prompts households to save and invest in agricultural equipment. All WUAs received a financial literacy training and were eligible to receive a matching grant. Twenty-one WUAs were randomized into the treatment group that additionally were visited by private consultants quarterly, who tailored their advice to meet individuals’ own savings and investment objectives. We find the follow-up visits increase ‘hidden savings’ in the form of new capital investments on farmers’ own account. Thus, the visits may have changed savings’ habits by leading farmers to invest in technologies that were not directly subsidized. Our ability to detect an additional effect on the type of investments farmers targeted through the matching grant and, hence, the savings for the respective investments is limited given the power of our study design. Although the proportion of households saving increased, the intervention was likely less cost-effective than other modalities aimed to enhance the proclivity to save.The whole analysis in the paper can be rerun by using the script MAIN_proirri.do, which is in the Dofiles subfolder. It is only necessary to add your computer's username and path to the cloned replication folder(s) in line 94-96 of such do-file in PART 1.
You can select which section(s) to run by editing the locals in the preamble of the do-file. Make sure to run the packages section – PART 0 to install all necessary packages before running the other sections.
The main script will take around a full day on a reasonable cluster. Without considering the do-files using bootstrapping replications (see section Code Process below), it would take around 3 minutes.
The individual do-files with their respective inputs and outputs are explained below. The do-files employ finalized datasets, which are constructed from various data sources, listed and described below.
Computational reproducibility was verified by DIME Analytics. Details of the reproducibility checklist can be found here.
The final dataset PROIRRI Financial Literacy - Savings paper data.dta contains information on all 42 associations and 3,081 households in the experimental sample.
The corresponding ID variables are associd and hhid.
As presented in Section 3 of the paper, multiple sources of data were collected to assess the impact of the study program. Every variable in the final dataset has a prefix, which specifies the origin of the information it contains. Namely, we apply the following pattern to inform variable naming:
ad_for project administrative data;bp_for box pick-up data;bl_for association and household baseline census;el_for household endline survey.
You can find a more detailed description of all the variables employed in the data analysis, especially with regard to variables constructed using survey data, here.
The survey instrument is also available in Portuguese here and the related CTO form here.
The name of the do-files, which are run by the principal script MAIN_proirri.do, corresponds to the .tex or .png file to be created in the output folder, with the exception of tab03-4,A04-6-hh_multiple_hypothesis_testing.do and tabA11-12-hh_multiple_hypothesis_testing_controls.do.
The latter do-files estimate the program impact on secondary outcomes from household survey data, such as mechanization use and ownership, credit, input use and other costs. These variables were not part of our pre-analysis plan for the experiment filed in the AEA RCT Registry as "Group Interventions for Agricultural Transformation in Mozambique" (RCT ID: AEARCTR-0000937) [BibTeX], and therefore we adjust the p-values for family-wise error rate using the free step-down procedure by Westfall and Young (1993).
All do-files use the final dataset, PROIRRI Financial Literacy - Savings paper data.dta.
All tables and figures were included – without further editing – in the TeX document containing the current version of the paper.
You can find a more detailed description of each do-file's inputs and outputs here.
If you have any comment, suggestion or request for clarifications, you can contact Matteo Ruzzante at matteo.ruzzante@u.northwestern.edu or directly open an issue or pull request in this GitHub repository.
This work is licensed under a Creative Commons Attribution 4.0 International License.
