UnitRoot: Unit Root Work Flow

In efriedland/friedland: Making Functions Do Your Econometrics For You

Description

UnitRoot requires adf() and MacKinnon and allows the user to quickly determine whether a variable is stationary or non-stationary. The user enters the variable of interest, the model specifications (constant, time trend) and then the number of variables within the relationship. Often 1 variable is used, but ADF tests of residuals will consist of the number of dependent and independent variables in the model. Thus the function can be used to test for cointegration. Calculates ADF statistic for a defined number of models varying by lag number, determines the smallest sized AIC, checks the ADF statistic against MacKinnon Table 3 and determines stationarity.

Usage

UnitRoot(variable, k, drift = TRUE, trend = TRUE, cointvariables = 1, startfrom = 0)

Arguments

variable	The Time Series variable.
k	The maximum lags desired, Use only a single positive integer. This will determine the number of models in the selection process.
drift	Logic statement. If TRUE, includes an intercept/constant in the model.
trend	Logic statement. If TRUE, includes a time trend in the model.
cointvariables	Cointegrating variables in the relationship. For the stationarity of a single variable this will be 1. For cointegrating relationships, this equals the number of independent variables + the dependent variable.
startfrom	The start date of the sample in the test, can be used to limit observations

Details

Used to pick out which variables are stationary and nonstationary.

Value

selection	Matrix showing the Lags tested from 1:k and the results at each lag
Lag	Which lag was chosen based on AIC score
StartDate	the starting date the regresison was run from
AIC	the criterion function results used to select the best model
ADFstatistic	the t-value of the augmented dickey fuller test's lagged variable.
significance	the calculated critical value using the MacKinnon table. If the ADF statitics is too high or low to be used, this value will be NA.
results	The resulting unit root. "I(1)" implies non stationary and "I(0)" implies stationary
reason	A sentence explanation of the result in the context of the MacKinnon table.

See Also

adf, aic, UnitRootApply

Examples

Asymmetry.ts <- ts(Asymmetry[,-1], start = c(2003,6), frequency = 12)
UnitRoot(Asymmetry.ts[,"Pcrude"])
UnitRoot(Asymmetry.ts)

efriedland/friedland documentation built on Feb. 9, 2021, 11:53 a.m.