Our Goals

Crude oil is a risky asset insofar as its price cannot be perfectly predicted. The crude oil market is not operating in isolation: It is embedded in the economy and interacting with other markets. Our project is concerned with measuring and forecasting the joint risk of

• WTI crude oil,
• the Dow Jones Industrial Average (DJIA).

On the basis of our analysis, we provide a weekly forecast for the price of WTI and the DJIA.

Technically speaking, we are interested in the conditional standard deviation of next week's price change, given all the information about the time series available this week. (The standard deviation of price change is a measure for the volatility of a financial instrument. It quantifies the risk associated with this instrument.) The mathematical tool for this is a bivariate GARCH model.

The price of crude oil can go up or down from one week to the next. We speak of

• "good news" if oil becomes less expensive, or the DJIA increases (or both),
• "bad news" if oil becomes more expensive, or the DJIA decreases (or both).

We want to find out what the impact of news means for the future volatility. Positive news need not have the same effect as negative news on next week's volatility. This makes it necessary to use a bivariate asymmetric model for conditional heteroskedasticity. The model specifications are illustrated here.