The Granger-causality (GC) and error correction (ECM) techniques were applied on 1970-1999 data for Malawi to examine cointegration and causality between electricity consumption (kWh) and, respectively, overall GDP, agricultural-GDP (AGDP) and non-agricultural-GDP (NGDP). Cointegration was established between kWh and, respectively, GDP and NGDP, but not with AGDP. The GC results detect bi-directional causality between kWh and GDP suggesting that kWh and GDP are jointly determined, but one-way causality running from NGDP to kWh. The ECM results detect causality running one-way from GDP (also from NGDP) to kWh suggesting that a permanent rise in GDP may cause a permanent growth in electricity consumption. (C) 2003 Elsevier Science B.V. All rights reserved.