I don't think there really was a period in the 90s when such a statement holds true at all.
Unfortunately for those times I've only found figures for global market share, so I don't know exactly how what I said can be extended to US market share, which is what this thread is about and perhaps what interest you most. I also admit that I had underestimated the current global market share of Mac OS X, so the words I used, "most of its lifetime", should rather be "part of its lifetime".
But anyway, although most published statistics for 2012 are still preliminary, the estimated global market share of Mac computers for 2012 is about 5%. Maybe they just had a fabulous quarter that will prove this estimation wrong when Apple publishes its results, but for now I'm taking 5% as good. In the times of classic Mac OS, Macs had an annual market share bigger than that in eight occasions:
Mac global market share
1984: 6%
1987: 6%
1988: 6%
1991: 11.20%
1992: 12%
1993: 10%
1994: 9.30%
1995: 9%
2012: ~5% (provisional)
My main source for these numbers is Jeremy Reimer's "
Total Share: Personal Computer Market Share". Concerning the reliability of those data, I just can say that they are the best I was able to find. He's been keeping those statistics for a number of years now, so I guess that if he had said something that departs essentially from what statistical researchers say someone would have raised an alarm.
About the matter of US market share throughout those years, I believe that the difference from global market share was not as pronounced as today, but I still think that at least during the period 1991-1995 Macs had bigger market share than now even in the United States. I would estimate about 18% Mac US market share for the year 1992, but I concede that this is a wild guess on my part. Please feel free to share if you find US PC market share reports on those years.