What's the evidence that demonstrates NSA certificate forgery would be traced back to NSA?
There are two reasons to misappropriate a certificate: (a) impersonating someone else (b) decrypting already-tcpdumped traffic. The people demanding the secret keys to that certificate surely must have known (a) would cause Ladar to issue a revocation for the compromised certificate that renders such a plan null and void, so that leaves us with option (b). Furthermore, option (b) cannot be accomplished by compromising a CA after the fact. So, if what they wanted was to decrypt recorded traffic -- remember they were chasing someone who had used the service in the past, so this is the likely option -- they would have had to (as they did) extort the keys out of Ladar, irrespective of whether a CA had been compromised or not. That's why the Lavabit case disproves nothing w.r.t. CA compromise.
We are discussing people who do textbook malicious stuff in secret, so much of it that it is mind-boggling, and most of it we'll never find out. In fact, the only reason we know about these things is because someone blew the whistle. We have flabbergasting evidence that they're doing things way beyond CA compromise ("find, fix, finish" e.g.) And CA compromise is not at all beyond their capabilities in the slightest, nor is it against their stated interests.
So to insinuate that people who suspect CA compromise might be happening are "conspiracy theorists", defies reason. It is to say "Oh, these people who steal keys for a living would never steal THAT extra-powerful set of keys". ? That's epic special pleading.
I used to have this teacher in college, fiftysomething lady educated in computer science in Soviet Russia, who liked to harp extensively about all these secret U.S. programs (like Echelon, the standard conspiratorial stuff at the time). I of course laughed behind her back because, clearly, she was a conspiracy theorist punch card operator... or so I believed. Turns out the joke was on me -- what she described was but a drop in the bucket compared to what actually happens. I am no longer in the habit of reflexively stigmatizing people with "conspiracy theory!!!" whenever (i) somebody says something critical (ii) about powerful people (iii) who operate in full secrecy (iv) that is entirely within the realm of possibility and (v) aligns with their interests.
The logically consistent position in this case is not to discard the possibility of CA compromise -- it is rather to at least suspect some CAs may be compromised, or in other words, to be reasonably skeptical about claims that CAs haven't been compromised. There is nothing out of the ordinary, conspiratorial, or absurd about holding that position. This does not mean that every CA is untrustworthy, but it does mean caution is warranted.