[Forecasting Oil Price using AFVAR model]

By

Acknowledgement

I would take this opportunity to thank my research supervisor, family and friends for their support and guidance without which this research would not have been possible.

DECLARATION

I, [type your Full first names and surname here], declare that the contents of this dissertation/thesis represent my own unaided work, and that the dissertation/thesis has not previously been submitted for academic examination towards any qualification. Furthermore, it represents my own opinions and not necessarily those of the University.

Signed __________________ Date _________________

Abstract

The price of oil affects everyone, everyday. This thesis investigates two different approaches of oil price models for prediction and forecasting. The first is a structural model and is an application of the AFVAR model. This model is extended to a structural systematic model. The method and model are justified with corresponding econometric tests and procedures. The results are very conclusive, as they confirm what was shown in earlier research on different time periods. It is confirmed that in terms of forecasting the nonlinear models perform the best, however the structural models help explaining effects and importance of single variables.

Table of Content

Chapter 1: Introduction6

Motivation6

Chapter 2: Literature Review8

Oil Forecasts8

Theory8

Returns to Storage9

Scarcity Rent10

Chapter 3: Methodology11

Vector Autoregressive Model11

Structural Vector Autoregressive Model12

Data13

References14

Chapter 1: Introduction

Oil is a commodity that unlike any other, affects everyone's daily life in a plethora of ways. The modern world is as dependent on oil as the human body is dependent on blood. Oil prices and availability affects transportation be it everyday driving or flights, as well as economical growth, since goods have to be transported and oil is used almost everywhere in the secondary industry. Machines have to be oiled, engines need fuel and some products are entirely oil-based. The insulation on the wires in computers and the circuit boards are made from crude oil. Common products like shampoo, detergent, solvents, paint, ink, tires, lubricants, candle wax and hundreds of thousands of other products are made from oil (Elliott, 2005, 1081). Heating and of course the military is heavily dependent on oil prices and availability. The uniqueness of oil and its use in the global economy makes it challenging to find another comparable resource. Furthermore, one must not forget that oil is a non-renewable fossil resource. Presently economic activity and oil consumption are fairly correlated. It is not only for the aforementioned reasons that the interest in oil prices and in particular, in the ability of being able to forecast oil prices is crucial.

Motivation

In essence there are two different types of modelling approaches. The first is a structural model of the price of oil, depending on fundamental data such as demand and supply and is implemented through the use of a linear regression. The second is a time series approach, which includes linear and nonlinear time series analysis. The nonlinear time series analysis includes a neural network autoregressive model, where the complexity of the relationship is modelled with the help of neural network regressions. Modelling the price of oil is difficult because of the changing variability over ...