Working Paper: NBER ID: w30905
Authors: Arya Gaduh; Rema Hanna; Benjamin A. Olken
Abstract: Concerns about fraud in welfare programs common arguments worldwide against such programs. We conducted a survey experiment with over 28,000 welfare program administrators and over 19,000 beneficiaries in Indonesia to elicit the ‘marginal disutility from corruption,’ i.e., the trade-between more generous social assistance and losses due to corruption and fraud. Merely mentioning corruption reduced perceived program success, equivalent to distributing more than 20 percent less aid. However, respondents were not sensitive to the amount of corruption—respondents were willing to trade off $2 of additional losses for an additional $1 distributed to beneficiaries. Program administrators and beneficiaries had similar assessments.
Keywords: Corruption; Social Programs; Welfare; Indonesia; Survey Experiment
JEL Codes: D73; I38; O15
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Corruption (D73) | Perceived Program Success (I23) |
| Corruption (D73) | Program Satisfaction (C88) |
| Missing Funds due to Corruption (H84) | Perceived Program Success (I23) |
| Corruption (D73) | Tolerance for Corruption (D73) |
| Corruption Sensitivity (D73) | Perceived Program Success (I23) |