Experimental Design
The impact study aims to make causal statements regarding the effects of matching and cash grants on business performance, productivity, and capacity to create jobs in enterprises. The study is designed as a randomized controlled trial (RCT) with three groups each with 400 entrepreneurs either beneficiaries of grants (matching or cash) or non-beneficiaries (the control group). Each beneficiary was allocated a grant amount up to US$ 8,000. The actual grant amount was determined from the business plan of every firm. The amount varies across firms, but all firms received their requested cash amount, which corresponds to the cost of inventory, working and physical capital budgeted in their business plans.
The selection started off with a business plan competition to shortlist 1,600 entrepreneurs located in nine municipalities within the project area. These were selected out of 2,200 applicants. In shortlisting participants in the different communities, women were particularly favored thanks to selection quotas. The elaborated targeting at entrepreneurs that have growth potential and a vision on how to develop their business distinguishes this policy experiment from many other previous experiments. All shortlisted entrepreneurs were trained on business management. The trainers delivered a standard and simplified training certified by the International Labor Organization (ILO). Although the relevance of the business plans guided the shortlisting of participants, the training intended to ensure that all participants would be able to implement their business plans if offered treatment. Noteworthy is that this training does not address the skills constraints specific to each eligible entrepreneur, so the matching grants remain a valuable asset for them.
The randomization was conducted through public lotteries held in the nine municipalities. To ensure cost-effectiveness and tailor the support to the actual needs of the firms, the randomization was performed within tranches of the requested cash grant amounts. This design ensured an identical distribution of the grant amounts in the groups of matching and cash beneficiaries.
The eligible activities as well as the disbursement rules are specific to each intervention. Thus, the impact assessment will not be able to disentangle the role of each single feature. The evaluation will rather assess the effects of each ‘intervention package’ along with their respective procedures of implementation.