Higher-Order Risk Preferences and Risk-Management Behavior among Italian Winegrowers

2026-01-23

2026-04-30

Key independent variables (experimental measures)
- Risk aversion: Number of “safe” choices in decisions 1-5 (range 0-5)
- Prudence: Number of choices classified as prudent in decisions 6-10 (range 0-5)
- Temperance: Number of choices classified as temperate choices in decisions 11-15 (range 0-5)

Primary behavioral outcomes:
Reported behavior:
- Insurance status (current and past 5 years; yes/no)
- Mutual fund participation (current and past 5 years; yes/no)
- Adoption of active defense practices (current; check-list items)
Stated intentions:
- Interest in taking up green insurance (Likert 1-7)
Observed behavior:
- Participation in the weather-station loan draw (yes/no)
- Administrative records that may include data such as insurance participation (yes/no), total insured area and value, premium paid by the farmer, type of insured product, participation to mutual funds, depending on data availability at local consortium level.

Covariates:
Age, gender, education, years of experience in viticulture, farm size (in hectares), and region. Climate-belief measures (Section 3) may be included as controls and/or explored as moderators.

The study consists of four sections.
In Section 1, farmers state their preferences for an innovative “green insurance” product and selected attributes. Green insurance is described as an insurance product offering economic benefits conditional on the adoption of selected sustainable practices.
In Section 2, risk aversion, prudence, and temperance are elicited via 17 incentivized binary lottery choices following Noussair et al. (2014). The choices are grouped in four consecutive parts: decisions 1-5 elicit risk aversion, 6-10 elicit prudence, 11-15 elicit temperance, and 16-17 test relative risk aversion and relative prudence under expected utility. Participants are presented with one choice at a time and they always have to choose between two options (no indifference option). The left–right position of lotteries is randomized. Lotteries are presented in compound form and framed as independent coin tosses.
In Section 3, we elicit subjective beliefs about the frequency of extreme climatic events affecting grape production.
In Section 4, participants complete a short questionnaire on socio-demographics, farm characteristics, and self-reported risk-management practices.
In Section 5, participants have the option to enter a draw to receive a weather station, provided free of charge under a two-years loan agreement (for use and return). Adopting this type of smart technology can be viewed as a preventive measure that helps manage climate-related risks and can therefore be considered a risk-management strategy.
Survey data will be integrated, when available, with administrative data on real risk-management behavior (e.g., insurance coverage, insured area, insured value), provided by farmers’ defense consortia.

Incentives: All participants receive a one-year subscription to a professional agricultural magazine as a non-monetary show-up reward. In addition, 10% of participants are randomly selected to receive a monetary payment ranging from €5 to €170 based on their decisions and luck. The payment is determined by one randomly selected lottery choice from Section 2 and the corresponding simulated coin toss outcome. Payments are made via fuel vouchers.
Additionally, at the end of the study (Section 5), participants have the option to enter a draw to receive a weather station on free loan. 35 who enter the draw will be randomly selected to receive the device. Participants are informed about this opportunity only at the end of the study, since the decision to enter the draw is used as an outcome variable in the analysis.

Not available

Randomization done by the computer software (oTree or R): random selection of the payoff-relevant lottery choice and random selection of participants receiving monetary incentives (oTree) and weather stations (R).

Individual participants (winegrowers)

No

Not applicable (individual-level randomization)

150 individual winegrowers

Not applicable (no treatment arms; correlational analysis with experimental measures)