Model building in Agricultural Economics with Machine Learning - Echoes from the past - Abstract and introduction

Aug 2, 2026

Upload by Sage

Machine Learning tools are transforming empirical research in agricultural economics, but the shift also raises concerns that modern methods may become purely data-driven. This paper examines the recurring tension between economic theory and data across the history of econometric modelling. It reviews theory- and data-driven approaches, studies key econometric evidence from agricultural economics, evaluates machine learning applications, then synthesizes findings to recommend how to integrate machine learning effectively while preserving theoretical relevance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Abstract

1. Introduction

Theory-driven vs data-driven modelling

Research questions and paper objective

The World’s Largest Open Access Agricultural & Applied Economics Digital Library This document is discoverable and free to researchers across the globe due to the work of AgEcon Search . Help ensure our sustainability. Give to AgEcon Search AgEcon Search http://ageconsearch.umn.eduaesearch@umn.edu Papers downloaded from AgEcon Search may be used for non-commercial purposes and personal study only.No other use, including posting to another Internet site, is permitted without permission from the copyrightowner (not AgEcon Search), or as allowed under the provisions of Fair Use, U.S. Copyright Act, Title 17 U.S.C. No endorsement of AgEcon Search or its fundraising activities by the author(s) of the following work or theiremployer(s) is intended or implied. Model building in Agricultural Economics with Machine Learning: Echoes from thepast C. Gardebroek 1, * and M. Kornelis2 1Agricultural Economics and Rural Policy group, Wageningen University, Hollandseweg 1,6716KN, Wageningen, The Netherlands 2 Wageningen Social and Economic Research, Droevendaalsesteeg 4, 6708PB, Wageningen, The Netherlands Contributed Paper prepared for presentation atthe 99th Annual Conference of theAgricultural Economics Society, Bordeaux School of Economics, University ofBordeaux, France 14 – 16 April 2025 Copyright 2025 by C. Gardebroek and M. Kornelis. All rights reserved. Readers may makeverbatim copies of this document for non-commercial purposes by any means, provided thatthis copyright notice appears on all such copies. Please note this isa preliminary draftversion of this paper. *Agricultural Economics and Rural Policy group, Wageningen University, Hollandseweg 1,6716KN, Wageningen, The Netherlands, koos.gardebroek@wur.nl # Abstract Machine Learning tools are currently transforming empirical research in agriculturaleconomics. However, a concern with these new tools is that they are purely data-driven. Thehistory of economic science reveals a recurring tension between the roles of economic theoryand data. The objective of this paper is to describe lessons from the apparent divergencebetween theory-driven and data-driven modelling approaches that can guide to the current riseof machine learning modelling in agricultural economics. We first discuss different views onusing theory and data in economic building in general terms. Next, we review several keyeconometric papers in agricultural economics in order to show how economic theory and dataare used. This is followed by an evaluation of agricultural economics publications that haveemployed machine learning techniques. Finally, we synthesize these findingsin the discussionsection and provide recommendations for the effective integration of machine learning inagricultural economics. Keywords Machine Learning, econometric modeling, economic theory, data analysis JEL code B41, C18, C51, Q00 # 1. Introduction A new methodological wave in the form of Machine Learning tools currently freshens upempirical research in agricultural economics. Classic econometrics toolboxes are rapidlycomplemented with new data-driven algorithms that search for patterns in often large datasetsto yield better forecasting or classification models (Storm et al., 2020; Brignoli et al., 2024) .This data-driven approach makes economists wonder whether thereis still a role for economictheory in model construction, and how these learning tools can be reconciled with well -established economic theory. Interestingly, these concerns are not new. In economic research, two distinct quantitativeapproaches are commonly recognized: theory-driven and data-driven. Theory-driven modelsbase their functional formson behavioural theories, hypotheses, ora priori assumptions, whichare then tested against real-world data. Conversely, data-driven models derive their structurefrom empirical observations, statistical tests, or data generalizations, from which theory isinferred. Both approaches have long been the

FAQs

Why does the rise of machine learning in agricultural economics raise concerns?

Because many ML tools are perceived as primarily data-driven, which prompts questions about whether economic theory still has a role in model construction and how it can align with established theory.

What is the core distinction between theory-driven and data-driven modelling approaches?

Theory-driven models start from behavioral theories or assumptions and test them with real data, while data-driven models derive structure from observations, statistical tests, or data generalizations and infer theory from the results.

How does the paper plan to guide the effective integration of machine learning?

It discusses different views on using theory and data, reviews econometric papers in agricultural economics, evaluates publications using machine learning, then synthesizes those findings in the discussion to provide recommendations.

Aug 2, 2026

Upload by Sage