The contribution of different agricultural subsectors to economic growth in Nigeria is investigated and further suggests policy implications for investing in each of these subsectors. To this end, Johansen cointegration test and Gregory–Hansen test for cointegration with regime shift, vector error correction model (VECM), dynamic ordinary least squares (DOLS), fully modified ordinary least squares (FMOLS), Granger causality, and frequency domain causality test are employed for data from 1981 to 2016. This paper further highlights the long and causal dynamics between the selected agricultural subsector, namely forestry, crop production, fishery and livestock, and economic growth. Findings from time and frequency domain causality tests indicate a one‐way causality running from various subsectors of agriculture to economic growth in Nigeria, meaning how various subsectors of agriculture are important for predicting economic growth. In addition, there is 54% speed of adjustment from the error correction model, suggesting a need for diversification of the economy into the agricultural sector as a means for sustainable economic growth in the face of the continuous plunge in the global oil price. In the long‐run, the effect of forestry, crop production, and fishery on economic growth is statistically significant and positive. These outcomes have inherent policy implication(s), which are elucidated in the concluding section. [ABSTRACT FROM AUTHOR]