Abstract
Kazakhstan is warming at roughly twice the global rate, and its grain and livestock sectors remain exposed to recurrent drought. However, national evidence on climate action (SDG 13) and zero hunger (SDG 2) is largely descriptive. This study examines how annual temperature anomalies, precipitation and drought years shape cereal yield and agricultural value added, whether these links are long-run or transitory, and whether yield transmits climate shocks to sectoral income. It employs a 20-year national time series (2006–2025) from World Bank, FAO/EDGAR and Copernicus ERA5 sources and applies unit-root and Zivot–Andrews break tests, ARDL bounds cointegration, Johansen–VECM, Granger causality, VAR impulse responses and variance decomposition, mediation analysis, Durbin–Wu–Hausman and 2SLS estimation, and robustness checks. The results show a cointegrating relationship for cereal yield (bounds F = 10.62, p < 0.01), a long-run precipitation elasticity of yield of 0.45–0.56, an insignificant contemporaneous temperature effect (0.008), and yield losses of 27% in drought years; drought cuts agricultural value added by 8.3% with the effect fully transmitted through yield (indirect effect −0.104, Sobel z = −2.75, p = 0.006). Precipitation explains 10% and temperature 6% of the forecast-error variance of yield, and yield explains 54% of the variance of value added. In comparison, a 1% rise in value added raises agricultural greenhouse gas emissions by 0.67%. The findings suggest that moisture, not mean warming, is currently the binding climatic constraint on Kazakhstan’s agricultural economy, that climate risk reaches sectoral income almost entirely through crop yields, and that output growth without decoupling will enlarge the sector’s emissions footprint. JEL Classification Codes: Q54, Q11, Q18, O13, C32.