The paper uses the autoregressive distributed lag approach to cointegration to test the Harberger-Laursen-Metzler (HLM) effect in the context of Bangladesh. The HLM effect predicts that a rise in the terms of trade from an exogenous shock to a small open economy will lead to an improvement in the country’s trade balance. The study findings confirm the existence of a long run relationship in the case of Bangladesh. The Granger test suggests unidirectional causality from income terms of trade to trade balance. The results are found consistent with the theoretical predictions.