【正文】 anagement. In Tanzania, it is rather the inability to sustain the huge energy import bills that is largely responsible for energy supply related debt, as a proportion of Tanzania39。 Anderson and Fishwick, 1984). Given the virtual dependence by rural population on these fuel sources, the bined environmental impact on the one hand, and on the other, supply difficulties are having negative impacts on output and growth. Thus, the need for, and the speed by which solutions must be sought have never been greater. From 10 such urgency to act, the need to to determine causal directions between the variables is born. Causality test for energyine relationship As discussed above, since regression analysis tells us nothing about the direction of causality in the relationship between two variables, it remains simplistic. In this regard, the causality estimating technique offers certain advantages. Understanding the causal direction between energy consumption and economic growth is crucial to effective energy management, especially in the developing countries where difficult choices have to made because of the peting demands for scarce resources. The causality test developed by Granger (1969) is employed. Our results stand to achieve two things. First, it will help to indicate causal directions between the two variables at discussion. For example, if a unidirectional causality exists from GDP to energy, it would mean that increases in energy consumption is affected by rising GDP. For an en ergy analyst, a case may exist for focusing on the ponents and structure of GDP in order to minimize the adverse effect of energy constraints on its sustainability. This is of policy significance to a country at an early stage of development. During this period, lifestyles are simple, produc tion is labour intensive and subsistence and transport infrastructure and urban planning are at an incipient stage of development. Most developing countries, especially the least developed, are presently at this stage of development. On the other hand, a simultaneous causality will validate the a priori assumed plementary relationship between energy and capital. It is during this stage that a country bees locked into habits of high energy consumption, when economic growth cannot be achieved without a mensurate amount of energy consumption. At this stage, only those countries with the right institutions, energy sav ing investments and capacity to formulate and implement policies can effectively manage its energy problems. It can be seen that a knowledge of the causal directions 11 between energy consumption and economic growth is vital to seeking solutions to the problems posed by energy supply con straints in the developing countries. Model specification Granger39。 the United Nation39。 and the United Nations Yearbook of National Accounts, various years. As discussed above, first difference was taken and used to detrend the data in order to achieve covariancestationarity or white noise process. The Fstatistic was used as a standard rule for accepting or rejecting the null hypothesis in favour of the alternative hypothesis. Rejecting the null hypothesis would indicate that the inclusion of past values of E in the regression provide us with a better explanation of current values of Y than when excluded and vice versa. The DWstatistic was used to test the presence of autocorrelation in the model. 13 Results From Table 2, we can observe that the null hypothesis (increased economic activity does not 39。 an increase in energy consumption) cannot be rejected for Tanzania in view of the estimated standard errors. This means that the inclusion of past values of Yand E in the regressions provided a better explanation of current values of E and Y than when excluded. Indeed, by this method, a simultaneous causal relationship can be seen to exist between energy consumption and economic activities, proxied by GDP and GNP. Given the predominance of cash crops and agricultural exports in total economic activities, this es as no surprise. However, a priori observation may suggest low energy use。spurious causality39。 on one or all the variables used. Such innovations are always present in the field of technology not least energy. For example, effective conservation measures, efficient energy end use appliances and changes in the laws affecting transport and urban planning will affect energy intensity relative to economic activities. Accordingly, Pierce and Hughes (1977) suggested that one way of accounting for 39。 is to test for instantaneous causality. Using this method, we have tested whether the inclusion of a current value of Y with its past values will provide a better prediction of current values of E and similarly if we are testing for instantaneous causality between E and Y. From the results shown in Table 3, the causality between energy and GNP is not instantaneous, but that between GNP and energy, GDP and energy as well as energy and GDP are instantaneous. This is not surprising since a dislocated sequence in energy consumption was unavoidable in the aftermath of the oil price increases in the 1970s. This had an adverse effect on GNP since more than % of the country39。s energy structure casts huge doubts on its ability to achieve sustainable economic growth and development, given its energy problems. Despite its huge mercial energy reserve (hydroelectricity, crude oil, gas and coal), percapitaenergy consumption ranked among the lowest in the world. Conventional or biomass energy (mainly wood and charcoal) account for more than 70% of total energy consumption whereas in Asia and Latin America, this source accounts for only 35% and 25% respectively (Davidson, 1992). While rural energy demand is totally met from this source, only 10% of ur