Study On Energy Demand Medium-and-long Forecast

This thesis quantitatively analyzes the general influencing factors on the energydemand forecast error after integrating the data related to the energy demandmedium-and-long forecast published almost annually by the world’s two big energyforecast agencies: IEA and EIA. First, compare the two agencies’ forecast data of energydemand in key areas and countries in the world by energy type and region division. Andmeanwhile, compare the specification error of the basic assumption data such as populationgrowth rate, economy growth rate and energy price. Then, make comprehensive analysis ofthe possible influencing factors, and measure the relevant factors with regional or energytype characteristics by using region index and energy type index. And then, add the basicassumption error variance, forecast time, region dummy variance and energy type dummyvariance to the linear model, and then by using ordinary least square method, analyze theinfluencing factors on the medium and long term energy demand forecast error. To find theaccuracy of the assumed value of the forecast model and whether its influence on theenergy demand forecast, the forecast error of the energy demand has the feature of regionand type. Finally, conclude the influencing factors on energy demand forecast level anderror, and make a propose to improve the energy demand forecast level.The conclusions are as follows: the specification error of the basic assumptionsdecides more than40%of IEA’s energy demand forecast error averagely; the forecastdegree ranks from top to down by region are the developed zone, developing countries andtransition countries; oil demand forecast level is the highest in the type forecast, while thecoal demand is always underestimated largely. So the paper gives the following suggestionsfinally: IEA should focus on the accuracy of the assumption data especially; theadvancement of the forecast skills on developing and transition countries’ economic growthrate for both of IEA and EIA helps reducing energy demand errors on those countries; it isnecessary to classify the regions basic on more energy related details to develop theforecast degree in regions; to reduce the forecast error by energy sorts, agencies should addsome energy structure related indexes, such as the old and new energy replacement rate.