My glib objection was not to the notion of there being less entropy in one state than the other, but in calling the result "the 2nd law of thermodynamics" in cases like this.

The phase space argument means that it is less likely for the system to wind up in the e-/e+ state than in the 2-photon state, and calling that a manifestation of entropy is entirely sensible. But in general, there's a non-negligible chance that the system will wind up in the e-/e+ state anyway: it's very much a possibility (and becomes even more so when the total energy is greater).

I usually reserve the term "2nd law" for cases where the probabilities are tiny enough that it's essentially impossible for the entropy to decrease by any remotely significant amount. That level of certainty typically requires systems of many particles, not just one.