Tropentag, September 16 - 18, 2026, Göttingen "Towards multi-functional agro-ecosystems promoting climate-resilient futures"

Breeding for portfolios: A risk-oriented perspective on yield stability and food security

¹Tilburg University, Tilburg School of Economics and Management, Zero Hunger Lab, The Netherlands
²Bioversity International, Information services and seed supplies, Costa Rica

Abstract

Breeding programs usually aim to develop new crop varieties that outperform the current market leader. In practice, however, farmers often grow multiple varieties on the same farm, which tends to mitigate environmental risk through response diversity: varieties respond differently to environmental hazards, such as heat. Breeders tend to introduce new varieties into this system without considering their contribution to the overall portfolio available to farmers, potentially missing a benefit for climate adaptation as climate risk increases. To explore this foregone benefit, we combined modelling with an analysis of empirical data.
Our model represents breeding programs and farmers as sequential decision-makers: breeding programs select new genotypes for release within genetic and budgetary constraints, while farmers choose among available varieties to maximize risk-adjusted yield. We evaluated four breeding strategies: one that maximized yield and yield stability of the total portfolio, and three that maximized, for the new variety only, expected yield, expected yield and stability, or market share. Strategy performance was measured as the gain in total farmer utility. We evaluated each strategy across 195,000 scenarios.
Portfolio-oriented breeding outperforms the stability-seeking strategy by at least 20 points of normalized gain in 24% of scenarios, and the yield-maximizing strategy in 42%. Gain is highest when existing varieties already respond differently to environmental conditions or are strongly biased toward one set of conditions.
An analysis of Kansas wheat data from 1993 - 2006 shows that the most widely planted varieties ranked 6th, 7th, and 11th in mean yield but had low mutual yield correlation. Even against high pool-wide correlation, portfolio-oriented allocation delivers a 10% utility premium over any other strategy. This suggests decades of replacement breeding narrowed environmental responses, so the 10% premium reflects what remains extractable from a narrowed pool. Breeding explicitly for complementary responses could yield considerably more.
Our findings suggest that portfolio-oriented breeding could substantially reduce climate risk, a benefit that may grow as climate variability increases, though current strategies limit it. Capturing it requires explicit risk analysis by breeding programs, and policy makers may wish to revise release criteria to balance yield gain against correlation with varieties on the market.

Keywords: Climate risk, food security, genotype × environment interaction, plant breeding strategies, varietal portfolios, yield stability

Contact Address: Paul Deuker, Tilburg University, Tilburg School of Economics and Management, Zero Hunger Lab, Tilburg, The Netherlands, e-mail: p.deukertilburguniversity.edu