# Timeseries: Evaluating Var models

#### studentbelg

##### New Member
I have build some var models with different types of variables and now I want to try to evaluate them and see wich one is the best

But when I look at the R² which i want to use as a criteria to choose the best model I get results like 0.9997 or 0.9999. Some combination even give an R of 1, which isn't really realistic or usable to select and appropriate model.

Is there another way to compare or what 'm I doing wrong?

model
summary(model)

Output is:

> summary(model)

VAR Estimation Results:
=========================
Deterministic variables: none
Sample size: 54
Log Likelihood: 118.939
Roots of the characteristic polynomial:
1.002 0.9372
Call:
VAR(y = cbind(Loglagond, LogIndexlonen), p = 1, type = "none")

Estimation results for equation Loglagond:
==========================================

Estimate Std. Error t value Pr(>|t|)
Loglagond.l1 0.91211 0.03326 27.425 < 2e-16 ***
LogIndexlonen.l1 0.17058 0.06031 2.829 0.00662 **
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.1059 on 52 degrees of freedom
Multiple R-Squared: 0.9999, Adjusted R-squared: 0.9999
F-statistic: 1.988e+05 on 2 and 52 DF, p-value: < 2.2e-16

==============================================

Estimate Std. Error t value Pr(>|t|)
Loglagond.l1 -0.01326 0.02783 -0.476 0.636
LogIndexlonen.l1 1.02730 0.05047 20.355 <2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.08863 on 52 degrees of freedom
Multiple R-Squared: 0.9997, Adjusted R-squared: 0.9997
F-statistic: 8.586e+04 on 2 and 52 DF, p-value: < 2.2e-16

Covariance matrix of residuals: