Sample size (or number of clusters) by treatment arms
In the first randomization step, the field offices are randomly split in two groups.
(1) Hence, we have 58 field offices in which clients can win the risk-coverage treatment. Within these 58 field offices, we consider 700 applicants. Of those 700 applicants, 350 will be randomly assigned to win the lottery, i.e., to receive the risk-coverage treatment and another 350 will not win the lottery and will thus be considered as the control group. The 350 individuals in the control group can still take up the loan.
(2) Moreover, we have another 58 field offices in which clients can win the subsidy treatment. Within these 58 field offices, we again consider 700 applicants. Of those 700 applicants, 350 will be randomly assigned to win the lottery, i.e., to receive the subsidy treatment and another 350 will not win the lottery and will thus be considered as the control group. The 350 individuals in the control group can still take up the loan.