European farmers split on agroforestry as policy support proves decisive

European farmers are far from united in their willingness to keep or start planting trees alongside crops and livestock, and the future of agroforestry on the continent may hinge less on persuasion than on the survival of existing subsidy schemes. A new survey-based study of 376 farms in Germany, Italy, Greece and Serbia finds that farmers fall into four distinct behavioural groups, and that removing policy support would sharply reduce both the intention to maintain existing agroforestry systems and the intention to adopt new ones. The findings, published in Agricultural and Food Economics, suggest that compensatory payments are effective at keeping current adopters on board but do relatively little to win over farmers who have never tried the practice.

The research team, led by Maria Raimondo with colleagues including Fabio Bartolini and Daniele Vergamini, set out to move beyond the simple adopter-versus-non-adopter divide that has dominated earlier work. Building on a classification first proposed by Morris and Potter in 1995, they sorted respondents into active adopters, who currently farm with trees and intend to continue; passive adopters, who currently use agroforestry but do not intend to persist; conditional non-adopters, who do not practise agroforestry today but might in the future; and resistant non-adopters, who reject it under any of the proposed conditions. A fifth group of indifferent farmers served as the statistical reference category in the econometric analysis.

Data came from an online questionnaire distributed across four case study areas spanning northern, southern and eastern European climatic zones. Because Serbia operates outside the European Union’s Common Agricultural Policy while Italy, Germany and Greece fall within it, the sample also allowed the researchers to compare farmers under different policy regimes. The authors are careful to stress that the sample is not nationally representative; it was drawn to capture heterogeneity across countries rather than to mirror the wider farming population, so the results are exploratory and indicative rather than definitive.

Before answering the main questions, respondents received a standardised definition of agroforestry as a system combining trees or shrubs with annual crops, meadow plants or grazing on the same land. Farmers then rated, on a five-point Likert scale running from extremely unlikely to extremely likely, how probable they would find it to maintain or adopt agroforestry over the next ten years under four scenarios: keeping the current Basic Payment Scheme and Countryside Stewardship payments, introducing new schemes to reimburse the costs of adopting agroforestry, creating a carbon market that pays for carbon storage, and abolishing all Basic Payment and Countryside Stewardship support entirely.

The scenario results reveal a clear hierarchy of incentives. Among current adopters, the highest stated intention to maintain agroforestry appeared under the compensation-cost schemes, with a mean score of 4.04, statistically indistinguishable from the 4.02 recorded under the status quo of existing payments, but significantly higher than the 3.90 observed for the carbon market. The abolition scenario produced a dramatic drop to 2.51. A similar pattern held for potential adopters: the lowest score, 2.45, occurred under abolition, and no statistically significant differences separated the first three scenarios. Notably, the mean scores of current non-adopters stayed below the neutral threshold of 3 in every scenario, indicating a generally low propensity to take up agroforestry regardless of the policy or market conditions on offer.

To translate these responses into behavioural categories, the researchers applied a multiple correspondence analysis, an exploratory statistical technique for examining associations among categorical variables. The first two dimensions of this analysis captured a substantial share of the total inertia, 66.1 percent and 15.1 percent in one panel and 56 percent and 39 percent in the other, showing that the two-dimensional representation reflects much, though not all, of the variability in the data. The first dimension broadly contrasted favourable with unfavourable stated intentions, while the second separated indifferent responses from more clearly defined positions. The authors emphasise that this analysis is a geometric representation of predefined categories rather than an empirical derivation of new behavioural groupings.

The core of the study lies in a multinomial logit model that estimates the probability of a farmer belonging to each category as a function of household, farm and personal characteristics, with indifferent farmers as the baseline. The authors caution that the uneven distribution of observations across categories, with some groups such as active adopters being small, reduces statistical power, so the estimates should be read as indicative rather than conclusive. Marginal effects computed after estimation translate the coefficients into changes in predicted probabilities, and the researchers note that differences in sign and magnitude between the two tables of results reflect the non-linear nature of the model.

Several structural variables emerged as significant. Larger farms were significantly less likely to remain indifferent, with a marginal effect of minus 0.076, suggesting that structural capacity pushes farmers toward some definite strategy, whether maintaining existing systems or contemplating future adoption. Greater farm diversification raised the probability of being an active adopter by 0.049. The presence of livestock was associated with a higher probability of belonging to the passive-adopter group, at 0.031, hinting that certain production systems shape short-term decisions. A higher share of farming income in household revenue, by contrast, reduced the probability of being a conditional non-adopter by 0.094, which the authors interpret as a sign that farms more dependent on agriculture may be more risk-averse about new practices.

Age and perception proved equally consequential. Older farmers were less likely to fall into the passive-adopter and conditional-non-adopter groups, with marginal effects of minus 0.023 and minus 0.029 respectively, and more likely to be resistant non-adopters, at 0.047. Formal agricultural education showed no statistically significant effect, indicating that training alone does not explain maintenance or adoption behaviour. Attitudes toward the supply chain mattered considerably: a positive perception increased the probability of being a conditional non-adopter by 0.058 and lowered the probability of resistance by 0.046, while a negative perception reduced the probability of being a passive adopter by 0.043 and raised the probability of resistance by 0.066. Access to technical advice was one of the strongest predictors, increasing the likelihood of being a passive adopter by 0.086 and a conditional non-adopter by 0.133, while reducing indifference by 0.108. Policy payments and organic farming, in contrast, showed no statistically significant effects.

Country differences added a further layer. Compared with Italian farmers, German farmers were less likely to be active or passive adopters, with marginal effects of minus 0.120 and minus 0.101, a pattern the authors link to more specialised and homogeneous production structures that may limit the integration of diversified practices. Serbian farmers were less likely to be passive adopters, at minus 0.115, but significantly more likely to be conditional non-adopters, at 0.248, pointing to a stronger orientation toward future adoption despite lower current engagement. Greece showed no significant differences from Italy. Taken together, the results suggest that while structural characteristics such as farm size and diversification facilitate the maintenance of agroforestry, the transition toward new adoption is driven largely by farmers’ attitudes and their access to technical advice, and that a one-size-fits-all policy approach is unlikely to succeed across such heterogeneous profiles.

Subject of Research: Farmers' stated intentions to maintain or adopt agroforestry in Germany, Italy, Greece and Serbia under alternative policy scenarios

Article Title: Understanding motivations for agroforestry maintenance and adoption: a comparative analysis in Europe

Article References: Raimondo, M., Bartolini, F., Dumper-Pollard, R., Reinhard-Kolempas, M., Roganovic, D., Schmutz, U., Thiesmeier, A. I. M., Zander, P., & Vergamini, D. (2026). Understanding motivations for agroforestry maintenance and adoption: a comparative analysis in Europe. Agricultural and Food Economics, 14(1), Article 64. https://doi.org/10.1186/s40100-026-00506-6

Image Credits: AI Generated

DOI: 10.1186/s40100-026-00506-6

Keywords: agroforestry, farmer behaviour, policy scenarios, multinomial logit, multiple correspondence analysis, Common Agricultural Policy, carbon market, technical advice, farm diversification, Europe, adoption barriers, sustainable farming