vars (version 1.1-9)

logLik: Log-Likelihood method

Description

Returns the log-Likelihood of a VAR, level-VECM or SVAR object.

Usage

## S3 method for class 'varest':
logLik(object, ...)
## S3 method for class 'vec2var':
logLik(object, ...)
## S3 method for class 'svarest':
logLik(object, ...)

Arguments

object

An object of class varest, generated by VAR(); or an object of class vec2var, generated by vec2var(); or an object of class svar

...

Currently not used.

Value

• An object with class attribute logLik.

encoding

latin1

concept

• VAR
• VECM
• Vector autoregressive
• Forecasts of VAR
• Prediction of VAR
• logLik

Details

The log-likelihood of a VAR or level-VECM model is defined as: $$\log l = - \frac{K T}{2} \log 2 \pi - \frac{T}{2} |\Sigma_u| - \frac{1}{2} tr (U \Sigma^{-1}_u U')$$ and for a SVAR-model the log-likelihood takes the form of: $$\log l = - \frac{K T}{2} \log 2 \pi + \frac{T}{2} \log |A|^2 - \frac{T}{2} \log |B|^2 - \frac{T}{2} tr (A'B'^{-1}B^{-1}A\Sigma_u)$$

References

Hamilton, J. (1994), Time Series Analysis, Princeton University Press, Princeton. L�tkepohl, H. (2006), New Introduction to Multiple Time Series Analysis, Springer, New York.

See Also

VAR, vec2var, SVAR, SVAR2

Examples

data(Canada)
var.2c <- VAR(Canada, p = 2, type = "const")
logLik(var.2c)