Essays on oil and the macroeconomy

Abstract

This dissertation studies the effects of crude oil on the macroeconomy using a vector autoregression (VAR) modelling framework. In particular, the importance of identifying different types of oil shocks is explored for both small and large open economies. A VAR model is also used to study the information content of oil futures contracts. The dissertation consists of three essays.

The first essay of the dissertation introduces a methodology for identifying oil supply shocks in a small open economy. Financial market information is used to construct an identification scheme that forces the response to an oil shock of the VAR model to be the same as that implied by futures markets. Due to the identification scheme, the model is only partially identified, and confidence intervals for impulse responses are calculated by using a bootstrapping procedure. The methodology is applied in illustrative examples to two small open economies in a VAR model that includes key domestic and international macroeconomic variables. The results suggest that while oil supply shocks have had an effect on domestic inflation during recent past, the effect on domestic GDP is ambiguous.

The second essay of the dissertation studies the existence of risk premiums in crude oil futures prices with simple ordinary least squares regression and Bayesian VAR models. It also studies the importance of three main theoretical risk premium models in explaining and forecasting the risk premiums in practice. Whilst the existence of the premiums and the validity of the models can be established at certain time points, it turns out that the choice of sample period has a considerable effect on the results. Hence, the risk premiums are highly time-varying. The study also establishes a model, based on speculative positions in the futures markets, which has some predictive power for future oil spot prices.

The third essay of the dissertation explores oil market and other macroeconomic shocks in a structural VAR model. It introduces a new indicator for oil demand, and performs a sign restriction set-up with a penalty function approach in an oil market VAR for a large open economy. The model also allows for macroeconomic shocks to interact with the oil market shocks. The results underline the importance of the source of an oil shock for its macroeconomic consequences. Oil supply shocks have been less relevant in driving real oil prices, and had less of an effect on inflation than demand shocks. Overall, the effects of oil shocks on real activity have been relatively limited.