They do it that way because that's the way they've always done it, at least as far back as 1550, and probably earlier. Many clock historians claim that IIII is supposed to provide artistic balance, since you mentally pair it off with VIII on the other side of the dial. (Presumably you see how the otherwise economical IV would have trouble holding its own in this respect.) The only problem with this theory is that the Romans apparently never used IV--it's a relatively modern invention. It's possible, in other words, that old-time clock makers used IIII because it was considered perfectly proper usage for all purposes, horological or otherwise, at the time. My friend David Feldman, in his book "Why Do Clocks Run Clockwise," cites an expert who says medieval clockmakers used IIII so as not to confuse the illiterate. You could count, "One, two, three, four! Hey, it's four o'clock!" Whereas having to subtract I from V to arrive at the same result was beyond your mental capabilities. Well, maybe. But let's think about this. The peasants couldn't handle IV, but somehow the IX for 9 posed no problems? Did only literate people go out after eight o'clock? Actually, as I read Dave more closely, he seems to be saying that at one time clockmakers used VIIII for 9. OK, but why are we using IIII and IX NOW? Tragically, we may never know the truth. History can be like that.
It goes on even more.... found at http://www.straightdope.com/classics/a2_153.html