Hi,

There are a variety type of Hausman tests (e.g. HAC robust or HCCME robust ones). Hausman specification test is also known as Durbin-Wu-Hausman test due to the contributions by Durbin, Wu, and Hausman. The test is used to test whether OLS estimator is consistent when your underlying variables are treated as exogenous. It is often used in GMM estimations. Of course, in linear models, it is generally used in 2SLS estimator. Note that Hausman test often involves Moore-Penrose generalized inverse.

Sargan test (Sargan 1958) is equivalent to Hansen's (1982) J-test in linear models. In more general Data generating process, the J-test is used to test whether a specific set of instrumental variables are orthogonal to the error term in the underlying model.

However, Sargan or Hansen's J-test often performs poorly, particularly in models with serrially correlated error term.

These tests are very much easier to be implemented in Gauss. Eviews is not very straightforward though it can achieve it. I have Gauss programms for these tests upon request

Hope this is of some help .

Best wishes.

Chengsi

Sargan Test:

The Sargan test is a test of the validity of instrumental variables. It is a test of the overidentifying restrictions. The hypothesis being tested with the Sargan test is that the instrumental variables are uncorrelated to some set of residuals, and therefore they are acceptable, healthy, instruments.

If the null hypothesis is confirmed statistically (that is, not rejected), the instruments pass the test; they are valid by this criterion.

In the Shi and Svensson working paper (which shows that elected national governments in 1975-1995 had larger fiscal deficits in election years, especially in developing countries), the Sargan statistic was asymptotically distributed chi-squared if the null hypothesis were true

The Hausman Test:

Given a model and data in which fixed effects estimation would be appropriate, a Hausman test tests whether random effects estimation would be almost as good. In a fixed-effects kind of case, the Hausman test is a test of H₀: that random effects would be consistent and efficient, versus H₁: that random effects would be inconsistent. (Note that fixed effects would certainly be consistent.) The result of the test is a vector of dimension k (dim(b)) which will be distributed chi-square(k). So if the Hausman test statistic is large, one must use FE. If the statistic is small, one may get away with RE.