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

How to cite

Barba, Francesca Romana; Cerroni, Simone; Palma, Marco; Drichoutis, Andreas. 2026. "Higher-Order Risk Preferences and Risk-Management Behavior among Italian Winegrowers." AEA RCT Registry. AEARCTR-0017606. https://doi.org/10.1257/rct.17606-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Francesca Romana Barba* , Simone Cerroni , Marco Palma , Andreas Drichoutis

* Primary investigator

Registry ID
AEARCTR-0017606
DOI
https://doi.org/10.1257/rct.17606-2.0
First registered
Last updated
Status
In development
Trial dates
2026-01-23 – 2026-05-28
Countries
Private
JEL codes
C99, D81, Q12, Q54
Registry record
http://www.socialscienceregistry.org/trials/17606

Abstract

This study elicits risk and higher-order risk (HOR) preferences of Italian winegrowers. More specifically, we elicit risk aversion, prudence, and temperance, and test their external validity by correlating them to both stated intentions and real-world agricultural risk-management behavior. Previous work suggests a link between HOR and real world behavior (see for example Noussair et al 2014, and Schneider and Sutter 2026). Data are collected through an online survey comprising four components: (i) stated preferences regarding an innovative “green insurance” product that provides economic benefits conditional on the adoption of sustainable practices; (ii) an incentivized experimental elicitation of risk and higher-order risk preferences using 17 lottery choices adapted from Noussair et. al (2014); (iii) subjective beliefs about climate-related production losses affecting grape production; and (iv) a short questionnaire on socio-demographics, farm characteristics, and self-reported risk-management practices; v) at the end of the study, farmers can become eligible for a weather-station free-loan draw by choosing to attend a 90-minute online workshop held after data collection; entry is conditional on full workshop attendance, which serves as the implementation cost. Adopting a smart climate technology such as a weather station can be interpreted as a mild preventive measure against climate-related risks and, therefore, as a risk-management strategy. Survey and experimental data will be integrated, where available, with administrative records on observed risk-management behavior, such as insurance participation and insured area and value.The study is implemented using the oTree open-source framework (https://www.otree.org/).

References:
Noussair, C. N., Trautmann, S. T., & van de Kuilen, G. (2014). Higher order risk attitudes, demographics, and financial decisions. The Review of Economic Studies, 81(1), 325–355.
Sutter, Matthias, and Sebastian O. Schneider. "Risk preferences and field behavior: The relevance of higher-order risk preferences." American Economic Review (2026).

Intervention

The study does not involve any intervention. It aims to test correlation between HORPs and stated intentions and real-world agricultural risk-management behavior. The study elicits three higher-order risk preferences (“treatment conditions”): first-order risk preferences (risk aversion), second-order risk preferences (prudence), and third-order risk preferences (temperance). All participants complete each elicitation (within-subject design); therefore, there is no random assignment to conditions and the study does not include experimental arms.

Experimental design

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. All participants complete all four parts in a within-subject design; there is no random assignment to treatment conditions and the study does not include experimental arms. 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 can become eligible to enter a draw to receive a weather station, provided free of charge under a two-years loan agreement (for use and return). Eligibility is conditional on full attendance of a 90-minute online workshop on how the station works and the relevance of on-farm data monitoring. 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. Participants are informed about this opportunity only at the end of the study, since workshop attendance (and thus draw entry) is used as an outcome variable in the analysis. Two follow up questions are also included: (i) for those who decide not to attend the workshop, their main reason for not attending (e.g., not interested in the station/workshop, time constraints, or already owning a station); (ii) participants’ motivation for taking part in the study, including the option “I wanted enter the draw for the weather station”, to assess potential spillover effects.

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 can become eligible to enter a draw to receive a weather station on free loan by attending a 90-minute online workshop. 35 among those who attend the full workshop (and thus enter the draw) will be randomly selected to receive the device.

Design details

Randomization method
There is no random assignment to treatment arms as treatment arms do not exist in this study. Randomization is used only for incentive implementation and is done by the computer software: random selection of the payoff-relevant lottery choice, random selection of participants receiving monetary incentives (oTree) and random selection of weather-station recipients among participants who attend to the full workshop and enter the lottery (R). Randomization of the left-right position of lotteries is implemented by distributing two survey versions (links) that differ only in the left-right placement of the options.
Randomization unit
Individual participants (winegrowers)
Clusters
Not applicable (individual-level randomization)
Observations
150 individual winegrowers
Arms
Not applicable (no treatment arms; correlational analysis with experimental measures)
Primary outcomes
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), operationalized as full workshop attendance measured using the workshop attendance records - 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.

Cited by

No citation data available for this registration yet.

Keywords