Disentangling Climate and Non-Climate Determinants of Cereal Productivity: Evidence from India’s Arid and Semi-Arid Regions

Abstract

Assessing the impact of annual climate variability on cereal crop yields is crucial for addressing food security challenges in countries with arid and semi-arid regions. Using district/state panel data for India (1998-2021), this study estimates fixed-effects, fully modified OLS (FMOLS), and dynamic OLS (DOLS) models and applies Pedroni and Kao tests to establish long-run relationships between climate, agricultural inputs, and yields of wheat, rice, and millets (bajra, jowar, and other small millets). Cointegration results indicate a stable long-run association among variables. Across crops, cultivated area is a strong, positive correlate of productivity. For wheat, higher mean maximum temperature and greater rainfall (when excessive) reduce yields, while agricultural credit supports productivity; mean minimum temperature shows limited effects. For rice, area remains strongly positive, temperature effects are generally negative but imprecisely estimated, and aridity substantially depresses yields. For millets, maximum (and in some models minimum) temperature lowers yields, whereas rainfall, fertiliser, and agricultural credit are yield-enhancing. Taken together, the results show that models omitting input and financial controls risk overstating climate impacts; including credit and fertiliser management meaningfully attenuates climate-only effects. Policy implications include targeting heat-risk management and stress-tolerant varieties where temperature pressures are acute, improving water management in rainfall-sensitive systems, strengthening access to well-structured agricultural credit, and promoting fertiliser stewardship and place-specific crop strategies to bolster resilience and sustain cereal production under a changing climate.