This empirical paper intends to examine the causal link and elasticity between paddy productivity and poverty rates (poverty headcount ratio) using balanced panel data of ten provinces in Sumatra during the 2010 to 2022 period. Dynamic heterogeneous panels, i.e., Dumirescu-Hurlin (DH) non-causality test and Pooled Mean Group (PMG) model, are applied to examine the direction of the causality and the short- and long-run elasticity. This study also incorporates the homogeneity slope test of Pesaran-Yamagata, panel IPS and CIPS unit root tests, and Pedroni cointegration check. Results reveal that variables are integrated of order one, , and they are strongly cointegrated. The DH causality test denotes a unidirectional causality from paddy productivity toward poverty rates, which implies the absence of a feedback effect. Following PMG results, there is a positive impact of paddy productivity (= 0.29) on poverty rates in the short run; however, this connection switches to become negative (= -0.48) in the long run. A 1% increase in paddy productivity has a beneficial role in declining poverty rates of around 0.48%, ceteris paribus. This finding is robust; the estimated parameter is in line with the alternative methods of Fully Modified OLS (FMOLS) and Dynamic OLS (DOLS).