Which loop gain are you referring to?
1. The loop gain of a single integrator?
2. The loop gain of the filter alone (generally multiple loops)?
3. The loop gain of the converter?
In the last case, at any instant, the loop is open,  so you must be thinking of the loop gain in some average sense.

Jitter Man

Spectre has a conventional small signal analysis, stb,  that computes the loop gain. That will work fine if you believe you can turn off the clock while measuring loop gain. Otherwise you will need to use the techniques described in Section 3.4.1 of Kundert's book. Of course, with the clock being active, you will have to substitute SpectreRF's PAC analysis for the AC analysis used in the book.

Jitter Man