Zhen Tian Jeff Lee Visut Hemithi Huan Zhang Diana Aguilar Yuli Yan A Deep Analysis of A Random Walk
Identification
Unit Root Null Hypothesis: PRICE has a unit root Exogenous: Constant Lag Length: 3 (Fixed) t-Statistic Prob.* Augmented Dickey-Fuller test statistic Test critical values:1% level % level % level
Pre-Whitening Log Transformation Trend in Var. Difference Trend in Mean
Pre-Whitening
Unit Root Null Hypothesis: DLNPRICE has a unit root Exogenous: Constant Lag Length: 4 (Fixed) t-Statistic Prob.* Augmented Dickey-Fuller test statistic Test critical values:1% level % level % level
Dependent Variable: DLNPRICE Method: Least Squares Sample (adjusted): 5/17/ /22/2008 Included observations: 815 after adjustments Convergence achieved after 6 iterations Backcast: 1/11/1993 5/10/1993 VariableCoefficientStd. Errort-StatisticProb. C AR(1) AR(2) AR(3) AR(5) MA(8) MA(18) R-squared Mean dependent var Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood F-statistic Durbin-Watson stat Prob(F-statistic)
Model Validation-1
Model Validation-2 Breusch-Godfrey Serial Correlation LM Test: F-statistic Prob. F(2,806) Obs*R-squared Prob. Chi-Square(2) ARCH Test: F-statistic Prob. F(1,812) Obs*R-squared Prob. Chi-Square(1)
ARCH GARCH (1) Dependent Variable: DLNPRICE Method: ML - ARCH (Marquardt) - Normal distribution MA backcast: 1/11/1993 5/10/1993, Variance backcast: ON GARCH = C(8) + C(9)*RESID(-1)^2 + C(10)*GARCH(-1) CoefficientStd. Errorz-StatisticProb. C AR(1) AR(2) AR(3) AR(5) MA(8) MA(18) Variance Equation C6.94E E RESID(-1)^ GARCH(-1)
Model Validation-ARCH GARCH (1)
ARCH GARCH (2) Dependent Variable: DLNPRICE Method: ML - ARCH (Marquardt) - Normal distribution MA backcast: 1/11/1993 5/10/1993, Variance backcast: ON GARCH = C(8) + C(9)*RESID(-1)^2 + C(10)*GARCH(-1) CoefficientStd. Errorz-StatisticProb. C AR(1) AR(2) AR(4) AR(5) MA(9) MA(18) Variance Equation C7.30E E RESID(-1)^ GARCH(-1)
ARCH GARCH (3) Dependent Variable: DLNPRICE Method: ML - ARCH (Marquardt) - Normal distribution MA backcast: 1/11/1993 5/10/1993, Variance backcast: ON GARCH = C(8) + C(9)*RESID(-1)^2 + C(10)*GARCH(-1) CoefficientStd. Errorz-StatisticProb. C AR(1) AR(2) AR(4) MA(18) Variance Equation C7.55E E RESID(-1)^ GARCH(-1)
Model Validation-ARCH GARCH (3)
Correlogram
Correlogram of Residual 2
Histogram
ARCH Test ARCH Test: F-statistic Prob. F(1,812) Obs*R-squared Prob. Chi-Square(1)
Forecast
Recolor
Comparison
A Little Bit Further
Story Behind the Scene To Investigate the Sources of Shock Geopolitical Events (War & Disasters) GDP / Mean Personal Income Vehicle Sales (SUV Sales) China Petro Consumption Speculation (Future Contract Price) Key Bibliography “Causes and Consequences of the Oil Shock of ” James D. Hamilton, UCSD (2009)