lazyCoint: Determining the Cointegrating Relationships in the laziest...

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

Description

A quick way to determine all of the possible cointegrating relationships in your dataset, give the dependent variable.

The function will automatically determine a maximum number of lags based on the number of observations in the dataset. It will then use the UnitRootApply function (currently set to only include a constant, no trend, using AIC) to determine which variables are nonstationary. Using the available nonstationary variables, it assembles every possible unique combination of variables to build formulas from.

The result is a matrix that returns only the formulas that appeared to cointegrate with the selected dependent variable.

Usage

 lazyCoint(Y, data) 

Arguments

Y	A case sensitive character string of the dependent variable.
data	The time series matrix containing the variables to be used.

Details

The UnitRootApply and UnitRoot functions used have drift = TRUE and trend = FALSE. Relies on MacKinnon 1994 table, which does not specify explicitly what to do when the number of cointegrating variables in the relationship exceeds 6.

Value

Formula	The full formula in character format.
significance	The resulting P-value from the UnitRoot function, looks directly to the MacKinnon 1994 table.
result	I(0) == stationary.

See Also

UnitRootApply, UnitRoot, MacKinnon, combn

Examples

lazyCoint("logprice", as.ts(M1Germany))

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