Match play for ratings I wouldn't think would be that problematic, especially for the IM title. You have to earn norms for the IM title (with I guess at least on exception) so just having a high rating won't do you much good.
The Natinal Master title given by the USCF may be a different story but I would think even if match play was used to gain a higher rating then later tournament results would quickly make it obvious if something nefarious was being done with the match play.
Nicholas Nip is a good example of ratings manipulation through rated matches. In order to become the youngest National Master in history, he played in about 4 matches versus people who were probably familiar and perhaps friendly towards him (Fuentes and Perez), and then played in some "Quads" which boosted his rating into the 2200's. After that...he played in one tournament in December 2008, rose 4 rating points, and then dissappeared from tournament play.
I agree about the International Master thing, though. There are too many restrictions and requirements for an easy match-setup to boost somebody's title that high.
Matches between chess players have always been controversial. Sure, it's more exciting for many to watch two powerhouses go at it one on one, but is it the most fair way? Outside interference outside of gameplay contributes greatly to who wins the match, so it becomes a question of other factors rather than the players' chess skills.
Matches can also be "fixed" much easier than tournaments with multiple players. If a player was close to achieving the required rating for an International Master title, and then arranged matches against some highly rated friends, won 12 games, drew 3, and lost 2, was it fair? Wouldn't the whole rating system be better off if USCF dissallowed 2 player matches?