0
$\begingroup$

I have these results when I estimate a VAR with two variables:Growth and Debt and p=2.How to interpret the result for each equation? Thank you.

VAR Estimation Results:

Endogenous variables: Growth, Debt 
Deterministic variables: const 
Sample size: 21 
Log Likelihood: -106.915 
Roots of the characteristic polynomial:
0.4555 0.4555 0.3236 0.1897
Call:
VAR(y = newdata, p = 2, type = "const")

Estimation results for equation Growth:

Growth = Growth.l1 + Debt.l1 + Growth.l2 + Debt.l2 + const 

          Estimate Std. Error t value Pr(>|t|)  
Growth.l1 -0.51486    0.31247  -1.648   0.1189  
Debt.l1   -0.83960    0.29895  -2.809   0.0126 *
Growth.l2  0.09635    0.29514   0.326   0.7483  
Debt.l2    0.10281    0.31715   0.324   0.7500  
const      6.12191    2.24173   2.731   0.0148 *
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


Residual standard error: 3.976 on 16 degrees of freedom
Multiple R-Squared: 0.3729, Adjusted R-squared: 0.2162 
F-statistic: 2.379 on 4 and 16 DF,  p-value: 0.09521 

Estimation results for equation Debt:

Debt = Growth.l1 + Debt.l1 + Growth.l2 + Debt.l2 + const 

           Estimate Std. Error t value Pr(>|t|)  
Growth.l1  0.731897   0.314552   2.327   0.0334 *
Debt.l1    0.619108   0.300940   2.057   0.0563 .
Growth.l2  0.007834   0.297111   0.026   0.9793  
Debt.l2    0.140517   0.319265   0.440   0.6657  
const     -3.443299   2.256689  -1.526   0.1466  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


Residual standard error: 4.003 on 16 degrees of freedom
Multiple R-Squared: 0.2981, Adjusted R-squared: 0.1227 
F-statistic: 1.699 on 4 and 16 DF,  p-value: 0.1993 

Covariance matrix of residuals:
       Growth   Debt
Growth 15.809 -9.857
Debt   -9.857 16.021

Correlation matrix of residuals:
        Growth    Debt
Growth  1.0000 -0.6194
Debt   -0.6194  1.0000
$\endgroup$
1
$\begingroup$

The normal way to interpret a VAR Model would be to calculate the impulse response functions and plot them. They show you how one variable reacts when a shock hits the system. (see ?irf)

imp<-irf(YOUR_MODEL)
plot(imp)

by the way: Maybe you cant estimate a valid model with such a low number of observations.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.