tsDyn: logLik.nlVar – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

logLik.nlVar

Extract Log-Likelihood

Log-Likelihood method for VAR models.

## S3 method for class 'nlVar'
logLik(object, ...)

`object`	object of class `VAR` computed by `lineVar`.
`...`	additional arguments to `logLik`.

The Log-Likelihood is computed as in Luetkepohl (2006) equ. 3.4.5 (p. 89) and Juselius (2006) p. 56:

LL = -(TK/2) \log(2π) - (T/2) \log|Σ| - (1/2) ∑^{T} ≤ft [ (y_t - A^{'}x_t)^{'} Σ^{-1} (y_t - A^{'}x_t) \right ]

Where Σ is the Variance matrix of residuals, and x_t is the matrix stacking the regressors (lags and deterministic).

However, we use a computationally simpler version:

LL = -(TK/2) \log(2π) - (T/2) \log|Σ| - (TK/2)

See Juselius (2006), p. 57.

(Note that Hamilton (1994) 11.1.10, p. 293 gives + (T/2) \log|Σ^{-1}|, which is the same as -(T/2) \log|Σ|).

Log-Likelihood value.

Matthieu Stigler

Hamilton (1994) Time Series Analysis, Princeton University Press

Juselius (2006) The Cointegrated VAR model: methodology and Applications, Oxford Univesity Press

Luetkepohl (2006) New Introduction to Multiple Time Series Analysis, Springer

data(zeroyld)
data<-zeroyld

#Fit a VAR
VAR<-lineVar(data, lag=1)
logLik(VAR)

Nonlinear Time Series Models with Regime Switching

v10-1.2

GPL (>= 2)

Authors

Antonio Fabio Di Narzo [aut], Jose Luis Aznarte [ctb], Matthieu Stigler [aut], Ho Tsung-wu [cre]

Initial release

2020-02-04