1 / 15

Analysis of the Level and Variance of Gasoline Prices in the U.S. (1991-2014)

Analysis of the Level and Variance of Gasoline Prices in the U.S. (1991-2014). Lorenzo Borghi and Nathan Wiseman STAT 758 – Time Series Analysis Dr. Ilya Zaliapin. Outline. Goals Data Preprocessing Model Results Conclusion Limitations. goals.

bgibbs
Download Presentation

Analysis of the Level and Variance of Gasoline Prices in the U.S. (1991-2014)

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Analysis of the Level and Variance of Gasoline Prices in the U.S. (1991-2014) Lorenzo Borghi and Nathan Wiseman STAT 758 – Time Series Analysis Dr. IlyaZaliapin

  2. Outline • Goals • Data • Preprocessing • Model • Results • Conclusion • Limitations

  3. goals • Find a transformation that makes the TS stationary • Estimate ARMA and GARCH Models • Provide dynamic forecasts that account for volatility clustering • Use the estimated conditional variances as a parameter in a model of vehicle choice and use in the future

  4. DATA

  5. Applying the Box-Jenkins Approach ACF - One first-difference ACF - Two first differences

  6. P values of the Box-LJUNG TEST Various orders of difference operators () log differenced series icorresponds to lags 1 to 53 j corresponds to the order of the operator (up to 10). The only operator that significantly improves the outcome of the test is the lag-2 operator

  7. weekly values of ()ln(Price)

  8. Grid search for arma order

  9. Assessing the ARMA (6, 6) Model Causality -Using thepolyroot command in R, it is found that the AR terms have a root of 0.41 Invertibility -The MA terms have a root of 0.06, which implies non-invertibility

  10. Best Fitting ARMA/GARCH Model

  11. Forecasts Accounting for Volatility Clustering

  12. Assessing Forecast Quality

  13. Conclusions • Adding the GARCH Model leads to a model that can account for how uncertain forecasting becomes in volatile periods • The overall forecasts have a significant positive correlation (r=0.789) with the observed values

  14. Limitations

  15. Questions?

More Related