I have a question about residual autocorrelation tests on eviews 8 student version.

I estimated a var model, and I'd like to test residual autocorrelation, which test do I have to choose on Eviews? Portmanteau or LM test?

I obtain very different results to one to each other, below the outcomes:

VAR Residual Portmanteau Tests for Autocorrelations

Null Hypothesis: no residual autocorrelations up to lag h

Sample: 1985Q1 2008Q2

Included observations: 94

Lags Q-Stat Prob. Adj Q-Stat Prob. df

1 1.257024 NA* 1.270540 NA* NA*

2 3.443273 NA* 3.504317 NA* NA*

3 7.354968 NA* 7.544969 NA* NA*

4 17.99864 NA* 18.66169 NA* NA*

5 31.96562 0.0002 33.41333 0.0001 9

6 39.80693 0.0022 41.78928 0.0012 18

7 42.43163 0.0298 44.62517 0.0178 27

8 51.81122 0.0427 54.87727 0.0228 36

9 54.70707 0.1522 58.07974 0.0913 45

10 62.43809 0.2014 66.73112 0.1144 54

11 68.60129 0.2932 73.71113 0.1676 63

12 79.93779 0.2439 86.70663 0.1140 72

*The test is valid only for lags larger than the VAR lag order.

df is degrees of freedom for (approximate) chi-square distribution

VAR Residual Serial Correlation LM Tests

Null Hypothesis: no serial correlation at lag order h

Sample: 1985Q1 2008Q2

Included observations: 94

Lags LM-Stat Prob

1 8.830470 0.4531

2 6.764551 0.6616

3 7.897322 0.5445

4 14.19993 0.1154

5 14.83966 0.0954

6 8.583225 0.4766

7 2.959672 0.9659

8 10.30467 0.3264

9 3.056786 0.9620

10 9.538542 0.3891

11 7.204604 0.6158

12 12.64424 0.1794

Probs from chi-square with 9 df.

How could I interpret these results? Which one do I have to choose?

thanks in advance. ]]>