The real issue is cost. If you can really only afford one fairing size right now, a 5m-class one is the one you want because you can handle wider payloads, should any come your way. Also building a 4m-class PLF would be an extra cost that isn't really a necessity. Not all companies have pockets as deep as ULA.
I seriously doubt that developing a smaller fairing is very expensive. In fact, it might well be the cheapest part of the rocket. And it is quite commercial. Ariane 5, Vega, and Atlas V all use Ruag and Antares uses fairings, I know Delta IV offered the SYLDA, so it seems to be a COTS piece of equipment. Please note that the separation mechanism, sound suppression inlay, connectors, etc. are mostly the same.
In any case, it might well be the cheapest upgrade available for the money. Remember that a 4m fairing has 38% less drag losses. That's a lot. So, in every 5.2m launch that would fit a 4m they are paying that cost. That also translates into strain on the thrust structure. Less strain allows for more aggressive ascent profiles. And since Falcon can fly with 3.6m Dragon, and 5.2m fairing, flying with 4m should be easy. An there's weight issues. In the case of the Atlas V, the 5m medium fairing is 1700kg heavier than the 4m. And the F9 separates the fairing while the US is firing, so it does adds an extra weight.
What's more, I think very few commercial satellites need 5m. If you need that you can only use Arian 5 (120M), Atlas V(180M) or H-IIA (I don't know how much). If you can use 4m, then you can use Proton (80M), Sea Launch (around 70M?), Soyuz-ST, LM-3BE. What do you think satellite operators usually chose? In particular, Atlas V and H-IIA/b have zero share in the commercial market. And only five satellites per year use the upper part of the Arian 5. So at most (and I guess is much less) there are five commercial satellites that might need 5m.
Look into the fairing sizes for DoD, too. Delta IV did 11 4m launches and 2 5m launches, plus the 5 heavies, of course. Atlas IV is 19 of 4m vs 9 of 5m. In just 3 (the two OTV plus one NROL) of the 11 launches (sans heavy) of 5m F9 could match the performance of the LV. So overall it would seem that they would rather get the extra performance
Regarding the PLF length, I read somewhere that SpaceX is claiming a variable-length PLF, should there be customer interest to justify its development. Additionally, has SpaceX ever claimed it wishes to do Ariane-5-style dual launches or is this just a forum assumption of what they will do with FH?
I think it was proposed somewhere. But I'm stating that as an economist. The FH is too expensive unless you dual manifest for at least 80% of the GTO market (that's about 16 satellites/yr, btw). Without dual manifest it's too expensive. Will be dirty cheap if they do, though.