Sims (1980) introduced an exciting and ground-breaking new framework which would prove to be extremely insightful for macroeconomic analysis. This is known as vector autoregression (VAR). The VAR model combines the one variable autoregression with a single-variable linear model and the result, is a model with 'n' equations and 'n' variables. From this, Sims proved that the present value of each variable can be explained by its own lagged values and also current and previous values of the remaining variables in the model. This framework offers a method to capture relationships in multiple time series data. Sims initially thought that this framework could become a key tool for analysing economic policies, forecasting and describing data.
Since his paper the main critique is that, whilst VARs are indeed a useful model for forecasting and data description, one must be extremely careful when interpreting the data. However the idea that the model can analyse policy has been widely disputed. Stock and Watson (2001) analysed the framework which Sims created and they concluded that the VAR model cannot be used to analyse policy as 'economic theory and institutional knowledge is required'. Since the VAR model is a-theoretic, it would not be accurate at analysing economic policies.
An initial glance at historical data and graphswhich have been collected for this study, they indicate that oil price shocks are cyclical and that they occur during most decades and are typified by a sudden increase in the price of oil. These shocks are then followed by a reduction in price as the economy stabilises through time. This sparked the interest of economists and econometricians who were keen to analyse the impacts that these shocks were having on the major economies in the world. As mentioned, oil price shocks appeared cyclical, however from the early 1970's and throughout the 1980's the shocks were occurring much more often.
Buckley (2009) writes that the UK was in recession for several short periods during this time, which placed further emphasis on researchingrelationships between the price of oil and key macroeconomic variables in the hope that once a relationship was found, correct measures could be taken to deal with a future oil price shock. Hamilton (1983) produced a paper which analysed the effects of oil price shocks on Gross National Product (GNP) in the United States of America (USA) using data from the period 1948 to 1972. He concluded that GNP would decrease after a period of a sudden increase in oil prices. Further to this, Hamilton claimed that attributing his results to completely random correlations between the variables would be incredibly naïve and irrational. Hamilton's conclusions were unchallenged and were actually supported by other economists. Significantly Burbridge et al (1984) found similar evidence for this relationship in Japan. The fact that this had been proven in a completely different economy supports Hamilton's idea that one must not assume these correlations are random.