Conference Agenda

Session motivation and relevance

Evidence-based agricultural policy design requires not only credible causal estimates, but also advanced methodological tools that ensure policy incentives are measured accurately, results are comparable across contexts, and insights extend beyond past outcomes. Despite substantial progress in empirical methods, important methodological gaps continue to limit the translation of research into real-world agricultural policy solutions.

First, agricultural policy incentives are often measured using aggregated indicators that obscure commodity-level dynamics and within-country spatial heterogeneity, particularly in large and non OECD economies. Second, empirical findings on policy impacts remain difficult to integrate, as studies rely on heterogeneous indicators, outcomes, and identification strategies that hinder systematic synthesis and cumulative learning. Third, policy-relevant evidence extends beyond ex-post evaluation: understanding present public acceptance of policy instruments requires experimental approaches, while anticipating future structural and distributional effects calls for simulation and forecasting tools.

This session brings together five studies that cultivate methodological advances across the entire policy-evidence cycle. They combine harmonized and transparent measurement of agricultural policy incentives, spatially disaggregated analysis of within-country policy incidence, systematic synthesis of ex-post evidence, experimental analysis of current policy preferences, and forward-looking modelling of future policy impacts. Together, the papers demonstrate how advanced quantitative, computational, and simulation-based methods can generate policy-relevant insights that are robust, interpretable, and actionable.

Conceptual focus

The session is structured along the policy-evidence cycle. The first two papers focus on ex-post measurement of agricultural policy incentives, moving from harmonized cross-country indicators to a spatially disaggregated within-country perspective. The third paper synthesizes this ex-post evidence through a systematic literature review, enabling cumulative learning across heterogeneous studies. The fourth paper shifts the focus to the present, using experimental methods to analyze public support and political constraints shaping agricultural policy choices. The fifth paper extends the analysis to the future, applying simulation-based modelling to anticipate long-term structural and distributional policy impacts under alternative policy scenarios. Together, the contributions demonstrate how advanced methods can be combined to generate policy-relevant insights across time and levels of analysis.

Paper contributions

Paper 1: Harmonized measurement of agricultural policy incentives

This paper introduces a harmonized, observation-based global dataset of Market Price Support (MPS) covering 119 countries from 1992 to 2023. While widely used indicators such as the OECD Producer Support Estimate and the World Bank Nominal Rate of Assistance provide important benchmarks, they remain limited in commodity coverage and conceptual consistency, particularly outside OECD economies.

The dataset integrates FAOSTAT producer prices and quantities with border prices constructed from UN Comtrade trade flows, explicitly accounting for preferential tariffs under regional trade agreements. Data collection and harmonization are implemented using advanced API-based workflows and AI-supported validation tools across multiple aggregation levels. Commodity-level price gaps at the HS-4 level capture short-term distortions linked to policy reforms, market shocks, and global price movements. The resulting dataset provides a transparent and reproducible foundation for cross-country policy analysis and global modelling.

Paper 2: How representative are national-level protection measures?

Traditional protection indicators were designed to capture policy instruments affecting international agricultural trade and therefore typically adopt a national-level perspective. This paper shows that, particularly in large and spatially diverse countries, such aggregation can mask substantial within-country variation in policy incidence.

Using Kazakhstan as a case study, the analysis exploits between-oblast variation in producer prices for key agricultural commodities. Panel data estimators are applied to examine factors associated with regional price gaps, including market access and structural characteristics. The results demonstrate that policy conclusions derived from national-level protection measures depend critically on aggregation assumptions, highlighting the importance of spatially disaggregated evidence for policy assessment.

Paper 3: Synthesizing evidence through LLM-supported meta-analysis

This paper addresses the limited comparability of empirical findings on agricultural policy impacts through a large-scale quantitative meta-analysis of more than 900 studies. While meta-analysis offers a framework for cumulative learning, its application in agricultural economics has been constrained by the scale and methodological diversity of the literature.

The study combines PRISMA-guided review protocols with large language model–assisted screening and data extraction, enabling systematic coding of policy indicators, outcomes, and methodological features at scale. Effect sizes are standardized using partial correlation coefficients, and meta-regression is used to assess how reported policy effects vary with indicator choice and identification strategy. The analysis demonstrates how advanced computational tools can enhance transparency, comparability, and policy relevance in evidence synthesis.

Paper 4: Political feasibility and the interpretation of policy impacts

This paper examines how empirically identified agricultural policy effects translate into public support for concrete policy instruments in food supply chains. Using experimental survey data from Germany and the United States, the analysis studies how individuals allocate a consumer food dollar and evaluate alternative redistribution mechanisms.

The paper introduces the concept of a “Food Dollar Trap,” describing situations in which strong preferences for redistribution toward farmers coexist with limited willingness to accept higher food price increases. By linking policy preferences to beliefs about markets and bargaining power, the study explains why empirically effective policies may face political resistance. The results highlight the importance of integrating experimental and political economy approaches into agricultural policy assessment.

Paper 5: Anticipating agricultural policy impacts with agent-based simulation

This paper extends agricultural policy assessment beyond ex-post evaluation by applying the agent-based model AgriPoliS for ex-ante analysis of policy scenarios targeting long-term structural change and system resilience. The model represents individual farms as heterogeneous agents that adapt production, investment, land-market, and exit decisions over time, with aggregate outcomes emerging endogenously through market interactions.

AgriPoliS is used to compare alternative policy designs, such as payment schemes, eligibility rules, and market regulations, under varying economic and behavioral assumptions. Rather than producing point predictions, the analysis explores distributional, structural, and dynamic policy effects, highlighting trade-offs, unintended consequences, and path dependencies. The paper demonstrates how simulation-based modelling complements empirical evidence and supports forward-looking agricultural policy design under uncertainty.