The AUTOREG procedure provides regression analysis and forecasting of linear models with autocorrelated or heteroscedastic errors. The AUTOREG procedure includes the following features:
estimation and prediction of linear regression models with autoregressive errors
autoregressive or subset autoregressive processes of any order
optional stepwise selection of autoregressive parameters
choice of the following estimation methods:
exact maximum likelihood
exact nonlinear least squares
Yule-Walker
iterated Yule-Walker
tests for any linear hypothesis that involves the structural coefficients
restrictions for any linear combination of the structural coefficients
forecasts with confidence limits
estimation and forecasting for A of ARCH (autoregressive conditional heteroscedasticity), and the following variations:
GARCH (generalized autoregressive conditional heteroscedasticity)
IGARCH (integrated GARCH)
EGARCH (exponential GARCH)
QGARCH (quadratic GARCH)
TGARCH (threshold GARCH)
PGARCH (power GARCH)
GARCH-M (GARCH-in-mean)
combination of ARCH and GARCH models with autoregressive models, with or without regressors
estimation and testing of general heteroscedasticity models
variety of model diagnostic information, including the following:
autocorrelation plots
partial autocorrelation plots
Durbin-Watson test statistic and generalized Durbin-Watson tests of any order
Durbin h and Durbin t statistics
Godfrey LM test
Ramsey’s RESET test
McLeod-Li portmanteau Q test for ARCH disturbances
Engle’s LM test for ARCH disturbances
Lee and King’s for ARCH disturbances
Wong and Li’s test for ARCH disturbances
Chow test
Akaike’s information criterion
Schwarz information criterion
Phillips-Perron stationarity test
Phillips-Ouliaris cointegration test
Kwiatkowski-Phillips-Schmidt-Shin (KPSS) test
Shin cointegration test
augmented Dickey-Fuller test
Engle-Granger cointegration test
Elliot, Rothenberg, and Stock test
Ng and Perron test
tests for statistical independence
Jarque-Bera test for normality
CUSUM and CUMSUMSQ statistics
exact significance levels (p-values) for the Durbin-Watson statistic
embedded missing values