McCloud's basic story: If some genres start to dominate in the comics market, stores will emphasize those genres. But this discourages people with minority tastes from frequenting comic book shops, which in turn leads stores to offer fewer titles that appeal to minority tastes, which in turn alienates these potential customers even further. Thus, suppose superhero comics aimed at young males overshadow romance comics for young females. This makes females less eager to stop by their local comics shop, which encourages vendors to put more emphasis on superheroes, which narrows selection... etc.
I'm not buying this story. I'll admit that it could explain why places called "comic book stores" focus on superheroes. But if there were really a lot of demand for non-superhero comics, they would simply be sold in other venues. Romance comics could be stocked by grocery stores next to Harlequin romances, for example... if they would sell. More plausibly, places like Borders would not only carry the disfavored genres, but shelve them outside the "comics" section to reach their eager but alienated audience.
McCloud's right, of course, to point out that the comics industry is far bigger and more diverse outside of the U.S. But if path-dependence is at work, it's path-dependence of consumer demand itself. Part of the reason why Japanese adults read comics today is that when they were kids, the Japanese saw adults reading comics; in the U.S., in contrast, many still believe that comics are "just for kids." This is just another example of the human urge for conformity - but it's conformity over time, rather than at a given moment.
The bottom line: Changing the way vendors sell comics isn't going to make much difference. If there's a "lock-in" problem, its source is the consumer, not the producer.
Of course, if Scott could get my graphic novel illustrated and published, I might be willing to rethink my position. :-)