Cohen, J., & Dupas, P. (2010). Free distribution or cost-sharing? Flashcards
(13 cards)
🎯 Central Research Question
Q1: What core question does the paper investigate?
A: Whether charging subsidized prices (cost-sharing) for health products like insecticide-treated nets (ITNs) leads to better targeting and usage than free distribution, or whether it simply reduces demand and coverage.
🧪 Data and Methodology
Q2: What is the experimental design used in the study?
A: A randomized controlled trial (RCT) conducted across 20 prenatal clinics in Western Kenya. Clinics were randomly assigned to different ITN price levels: 0, 10, 20, and 40 Kenyan shillings, corresponding to subsidies of 100% to 90%.
Q3: How was usage measured?
A subsample of women who received or bought ITNs were visited at home 3–10 weeks later.
Enumerators checked if the ITN was retained and being used (i.e., hanging over a bed).
Usage was recorded unannounced, and no one knew in advance that home visits would occur.
Q4: What is the purpose of the two-stage randomization (the lottery)?
A: To isolate sunk cost effects. In clinics with cost-sharing, women willing to pay the full posted price were randomly offered a discount through a lottery. This helped test if paying more made them more likely to use the ITN.
Q5: What outcome variables did the authors analyze?
Uptake – whether women acquired an ITN
Usage – whether it was used (hung at home)
Health selection – whether buyers were sicker (measured via hemoglobin)
Effective coverage – proportion using ITNs among all eligible women
📊 Key Findings
Q6: How does price affect demand for ITNs?
A: Demand drops dramatically even with modest prices. Uptake falls by 60 percentage points when moving from free to just $0.60 (40 Ksh). Only 25% of women who would take a free net would pay the standard subsidized price.
Q7: Does charging more lead to better targeting (i.e., sicker women buying)?
A: No. There was no evidence that women who paid more were more anemic or had higher health need than those who got ITNs for free.
Q8: Does paying more increase the likelihood of using the ITN?
A: No. Usage rates among those who received ITNs did not increase with price. There was no significant sunk cost effect, and retention was high (>90%) regardless of price.
Q9: What does the study conclude about psychological effects like sunk cost bias or price signaling?
A: The authors find no evidence of these effects in this context. People didn’t use ITNs more just because they paid for them, contradicting theories that cost-sharing increases perceived value or accountability.
💡 Interpretation and Policy Implications
Q10: What does the study suggest about cost-sharing vs. free distribution?
A: Free distribution reaches more people without compromising usage. Since ITNs have strong positive externalities, broad usage is crucial. Cost-sharing fails to target better and simply excludes the poor.
Q11: How does the study address concerns about “wastage” under free distribution?
A: It finds very little wastage—nets given for free were rarely resold and mostly used. The fear that people don’t value free goods is not supported by the data.
Q12: What model do the authors use to interpret the subsidy effect?
A: A simple Pigouvian subsidy model, where optimal subsidies depend on:
Elasticity of demand
Elasticity of use
Health need of marginal buyers
Externalities from product use
They show that optimal subsidies are highest when usage elasticity is low and externalities are high—just like with ITNs.
Q13: What’s the main conclusion of the cost-effectiveness analysis?
A: Free distribution saves more lives at a lower cost per life saved than cost-sharing programs. Especially given ITNs’ spillover effects, broad, free access is both economically and epidemiologically justified.
