Chib (1995) Calculation Query

Hello Everyone:

I am attempting to program in MATLAB the Chib (1995) marginal
likelihood calculation from the Gibbs output but am running into
problems regarding the correct numerical value. I'm utilizing an
independent normal-inverse gamma prior (i.e. multivariate normal for
Betas and inverse gamma for the variance sigma^2) in a simple multiple
regression setting. According to Koop (2003 book) this model has no
analytical solution for the marginal likelihood and I was wondering if
anyone out there has simulated some data and has a correct marginal
likelihood value that I can use to see if my code is working, at least
within a certain tolerance given the random nature of the Gibbs
sampling algorithm.

Any help in this matter would be greatly appreciated.

Regards,
Donald Lacombe

.