Stated Versus Actual Demand Disconnect
RCTReview
Hypothetical contingent valuation surveys fail to predict actual market adoption of novel financial products among smallholders. Farmers overstate willingness to pay during hypothetical interviews compared to real cash purchasing choices.
Picture this
Imagine asking restaurant patrons in a survey if they would pay ten dollars for a new exotic dessert, and most say yes. When the waiter brings the actual menu with real prices, almost nobody orders the dessert unless given a free discount coupon.
What the evidence says
Stated willingness to pay and actual insurance purchase were uncorrelated or negatively correlated (probit marginal effect = -0.0777, p < 0.10). Among households stating ex-ante willingness to pay at actuarially fair price, actual uptake was 37.2%, compared to 44.3% among those stating no ex-ante willingness. Unsubsidized demand was 0% in the study sample.
- Who was studied
- N = 2,399 rural households in baseline contingent valuation survey; N = 460 households tracked across 23 treatment villages with active insurance sales in Amhara, Ethiopia.
- How
- Randomized control trial comparing baseline randomized contingent valuation willingness to pay against actual sales with randomized subsidy vouchers (0 to 500 Birr).
What to do
Evaluate market demand for agricultural microinsurance through real-money field trials with randomized price vouchers rather than hypothetical willingness-to-pay surveys.
From the source
"The study compares a survey-based estimate of willingness to pay with actual uptake for the weather insurance, finding the stated and actual demand to be almost completely uncorrelated."
Productivity, credit, risk, and the demand for weather index insurance in smallholder agriculture in Ethiopia