We measure the effects of receiving information on financial consumer protection through multiple channel treatment (consisting of training sessions, interactive primers and SMS messages) on the levels of financial trust over formal institutions.
Experimental design
In order to understand the way information about consumer protection and financial literacy affect trust, a randomized controlled trial will be deployed. For both treatment and control groups, an interactive primer with financial literacy information in simple and practical terms will be deliver. The primer contains sections on production, workers, investment, machinery, debts, savings and planification accountability, including information on how to properly define monetary values to have a rigorous following of the finance of each component and interactive spaces for producers to carry out such accountability. For the treated group, a multiple treatment strategy will be implemented. Apart from the control primer, the treated group will receive an additional primer with financial consumer protection information, containing recommendations, information and tools to exert and defend their rights as consumers in the financial market, as well as some practical examples of common situations and recommendations on how to act if they face such cases. The topics of the flyout will be exposed in a training session of 3 hours in which participants will understand the material through a video and a capacitation on the topic . Lastly, after delivering the primers, a series of text messages will be send to the treated group, reinforcing the information delivered in the initial treatment.
Design details
Randomization method
Randomization done in office by a computer.
Randomization unit
In order to define treatment and control groups, aggrupation of villages in macro clusters was realized. The criteria for aggrupation were the minimization of distance of the centroid from observations within the same municipality. The baseline survey allowed us to have the precise location of each observation, allowing us to identify it up to 400 village. three observations that didn´t have the GPS location of the baseline survey were grouped in a macro village of its own defined as ‘others’.
In order to assign the macro-clusters into treatment and controll group, they are going to be selected by a stratified random selection, based on stablishing groups depending on relevant factors that can affect results, this is possible because of the previous knowledge from baseline survey on exclusive and relevant categories. The two categories in which we based the stratification was production levels and farm size. For the production levels of coffee, three groups are define based on the information of the 209-macro village as follows:
Low production: macro village with a total coffee production level equal or less than the 33.3rd percentile in the overall sample distribution.
Normal production: macro village with a total coffee production level between the 33.3rd and 66.3rd percentile in the overall sample distribution.
High production: Macro village with a total coffee production level equal or above than the 66.3rd percentile in the overall sample distribution.
For the farm size levels, the following three groups where defined:
Small areas: macro village where the total sum of farm area is equal or less than the 33.3rd percentile in the overall sample distribution.
medium areas: macro village with a total sum of farm area between the 33.3rd and 66.3rd percentile in the overall sample distribution.
Large areas: Macro village with the total sum of farm area equal or above than the 66.3rd percentile in the overall sample distribution.
This way, the stage stratification creates nine groups in which a simple randomization is applied to assign each macro village into treatment or control group based on the strata, where 312 households were assigned to control group and 342 were assigned to treatment.
Clusters
209 clusters
Observations
654 households
Arms
312 households control, 342 household treatment, 105 clusters treatment and 104 clusters control