Same seed, same regression, different eviews version and different result

[akbasege]

Hello, I am using v13 and trying to replicate a markov switch regression model runned by v12 with the same data, settings and seed. However I get slightly different results (which I cannot share here). I am wondering if the two eviews versions have different techniques used for markov switch.

[EViews Gareth]

No, but the random number generator isn't guaranteed to produce the same numbers between versions (even with the same seed)