Energy-output nexus in Bangladesh: A two-sector model analysis

The relationship between energy and output is one of the key issues in economic literature. Existing empirical studies have mainly used the neoclassical one-sector aggregate production function to examine the energy-output nexus and revealed inconclusive results. To address this issue, in this paper, we develop a two-sector model differentiating the energy sector from the conventional production sector to study the energy-output linkage for the Bangladesh economy. Using the annual data from 1985 to 2018, we conduct several unit root tests and reveal that our variables are stationary at their first difference. We further use the Johansen-Juselius cointegration and the Autoregressive Distributed Lag (ARDL) bounds tests and find that the variables are cointegrated in the long-run. By using three robust estimators (Dynamic Ordinary Least Squares (DOLS), Fully Modified OLS (FMOLS), and Dynamic ARDL), we reveal that a 1% increase in energy can increase the output by 0.06–0.10% in the long-run. We also run the Granger causality test and Vector Error Correction Model (VECM) to check the long-run and short-run causality. Our results find a unidirectional causality running from energy to output in the long-run but not in the short-run. We recommend that a reliable supply of energy needs to be ensured for future sustainable development in Bangladesh. •We develop a two-sector model differentiating energy and non-energy sector.•We find the estimated long-run coefficient of energy is positive and significant.•Unidirectional causality is running from energy to output growth in the long run.•No causality between energy and output growth in the short run.•We suggest proper energy supply-side management policies for development.