Preferences of extensive livestock farmers for wildfire prevention grazing programs: Evidence from a discrete choice experiment in Spain

This study investigates livestock farmers’ preferences for participation in public programs that use controlled grazing as a wildfire prevention tool. Respondents complete a survey including a discrete choice experiment in which they are asked to choose between alternative program designs for wildfire prevention grazing.
The program alternatives vary in terms of operational requirements and financial compensation. Farmers evaluate different program designs and may also choose not to participate in the program.

The intervention consists of a discrete choice experiment embedded in a survey administered to extensive livestock farmers in Spain. Respondents are presented with a hypothetical public program that compensates farmers for providing wildfire prevention services through controlled grazing in strategic areas such as fuel breaks.
In each choice task, respondents choose between two alternative program contracts and an opt-out option (“no participation”). The program alternatives are defined by four attributes:
1. Share of the herd enrolled in the program (less than 30%, 30–60%, more than 60%)
2. Distance between the farm and the designated grazing area (<5 km, 5–10 km, >10 km)
3. Cooperation requirements with another livestock farmer (no cooperation, cooperation with a farmer chosen by the respondent, cooperation with a farmer assigned by the administration)
4. Maximum payment per hectare (€50–€300)
The experimental design includes 18 choice situations divided into two blocks, with each respondent completing 9 choice tasks. Respondents are randomly assigned to one of the two blocks while order of choice cards within the blocks is randomized across respondents.
The experimental design was generated using NGENE with a D-efficient design algorithm based on prior parameter assumptions. The resulting design achieved a D-error of 0.08658.

2026-04-12

2026-09-30

The observed outcome variable is the choice made by the respondent in each choice task.
For each choice situation, respondents select one of three alternatives:
• Program A
• Program B
• Opt-out (no participation)

These choices constitute the dependent variable used to estimate farmers’ preferences for the attributes of wildfire prevention grazing programs.
The main quantities of interest derived from these models are:
• taste parameters associated with program attributes
• willingness-to-accept (WTA) estimates for participation conditions.

Secondary outcomes include:
• stated reasons for choosing the opt-out alternative
• attitudinal variables related to cooperation, risk preferences, and farm management objectives
• farm structural characteristics.

These variables will be used to explore observed heterogeneity in preferences.

This study uses a survey-based discrete choice experiment to elicit livestock farmers’ preferences for wildfire prevention grazing programs. Respondents are presented with a series of choice tasks in which they select their preferred option among two alternative program contracts and an opt-out alternative (no participation).



Each respondent completes nine choice tasks and is randomly assigned to one of two blocks of the experimental design.

The program alternatives vary along four attributes: the share of the herd enrolled in the program, the distance to the grazing area, the cooperation requirements with another livestock farmer, and the payment offered per hectare.

The experiment is implemented as a discrete choice experiment embedded in a survey administered to extensive livestock farmers in Spain.

Each choice task presents respondents with three alternatives: two program alternatives and an opt-out option representing non-participation.

Program alternatives are defined by four attributes:

share of the herd enrolled in the program (three levels),
distance between the farm and the grazing area (three levels),
cooperation requirements with another livestock farmer (three levels),
maximum payment per hectare (six levels ranging from €50 to €300).

The experimental design consists of 18 choice situations divided into two blocks, with each respondent completing nine choice tasks.

The design was generated using NGENE with a D-efficient design algorithm based on prior parameter assumptions. The resulting design achieved a D-error of 0.178.

The sequence of choice tasks within each block is fixed for respondents assigned to that block.

Computer-generated random assignment of respondents to blocks by the survey platform.

The unit of randomization is the individual respondent (livestock farmer). Each respondent is randomly assigned to one of two blocks of the experimental design.

No

Not applicable

The study aims to collect 200 responses from livestock farmers while 100 responses is considered as the minimum threshold required to estimating discrete choice models with the number of attributes included in the experiment. Recruitment will occur through livestock cooperatives, farmer associations, and professional networks.

Block 1 of the experimental design assigned to half of the sample
Block 2 of the experimental design assigned to half of sample