This study examines the role of total factor productivity with other major determinants on economic growth in the SSA over the period 1990–2022 using panel ARDL, GMM and their extended asymmetric non-linear ARDL and GMM models. The asymmetric non-linear ARDL and GMM are more superior to the counterpart ARDL and GMM symmetric models in the long-run than the short-run estimation, ensuring more efficient and reliable information regarding economic growth analysis. Finally, the robust dynamic inter-temporal Granger causality tests show that asymmetrically there are bi-directional causality between the growth rate of an upward movement in total factor productivity and real GDP as well as between a downward movement in the growth rate of terms of trade and real GDP growth rate. These empirical results are extremely important indicators for the SSA governments that should enhance total factor productivity and terms of trade is a policy suggestion of this research.