| Energy price fluctuations affect government and business policies, effective planning and implementation of strategic decisions and industrial growth and competitiveness of business enterprises. Energy is a major input in national and global economies and business growth and expansion, and thus, the ability to forecast energy prices is a critical component of major budgetary policies. Governments and industry continue to struggle in dealing with energy prices. Toward, the solution of this problem, many efforts are currently being made to provide some tools for guiding governments and industry in this domain. Energy prices are affected by a number of unforeseeable future events that are hardly predictable. However, modern science and economics have provided tools that can be used to provide forecasts with a reasonable degree of confidence. This research contributes toward this important issue of energy price forecasts. A number of statistical and econometric methods, including GARCH, ARIMA, PCR and neural networks modeling techniques, have been used to develop energy forecasts models. These energy models include electricity, coal, crude oil and natural gas prices and total energy consumption for Alberta and Canada. The determinants of energy prices in these models are the energy production, OPEC prices, the price of other energy products, personal income, GDP, number of oil and gas wells drilled (westca), personal income, unemployment and number of degree days. The models are verified and validated with data from CANSIM, Alberta Energy Library, EUB, Energy Prices and Taxes periodical, Annual Oil Market Report and OPEC bulletin. The results show that the PCR and neural networks techniques provide the best forecasts and could be used for developing reasonable energy price forecast for guiding regional, national and business policies. |
