Historical data on the percentage of women and minorities in philosophy in the U.S. are hard to find. It's clear that the numbers have increased since the 1970s, but it's possible that some of the trends have flattened since the 1990s. Without good historical data, it's hard to evaluate this possible flattening.
So I put in a request to the National Center for Education Statistics, and they generated estimates of the percentage of women and the percentage of non-Hispanic white people in philosophy vs. other disciplines based on governmental surveys conducted from 1988, 1993, 1999, and 2004. These data concern only full-time faculty in 4-year institutions.
With their permission, I've made these NCES estimates available here (XLSX format).
The 1988 and 1993 are aggregate data for philosophy, religion, and theology (which appear not to have been coded separately in the NCES survey raw data). However, I believe that these aggregated data can serve as reasonable estimates for philosophy in those two years, as I'll explain in point (4) below.
1988: philosophy*: 91% male (vs. 75% for all fields).
1993: philosophy*: 88% male (vs. 70% for all fields).
1999: philosophy: 80% male (vs. 67% for all fields).
2004: philosophy: 86% male (vs. 64% for all fields).
1988: philosophy*: 93% white non-Hispanic (vs. 89% for all fields).
1993: philosophy*: 95% white non-Hispanic (vs. 86% for all fields).
1999: philosophy: 91% white non-Hispanic (vs. 84% for all fields).
2004: philosophy: 89% white non-Hispanic (vs. 80% for all fields).
(1.) Unsurprisingly to anyone familiar with the demographics of philosophy in the U.S., philosophy was very white and very male throughout the period -- more so than academia as a whole. The trends have been toward increasing diversity, but only slowly.
(2.) The fluctuations suggest statistical error of at least a few percent. On the second page of the worksheet, you can also see the calculated standard errors from the NCES, generally about 1-4%. The dip from 88% to 80% back up to 86% in particular seems unlikely to reflect real demographic change. Turnover in full-time positions is slow; the field doesn't change that fast.
(3.) For comparison, the U.S. Survey of Earned Doctorates from the 1990s shows 73% of U.S. philosophy PhDs going to men and 91% going to non-Hispanic whites (the latter number excludes temporary residents and decline-to-state). So PhD recipients during the period had a bit more gender and ethnic diversity than faculty as a whole, consistent with a slow trend toward diversification.
(4.) As mentioned, the 1988 and 1993 numbers are actually numbers from "philosophy, religion, and theology". I nonetheless believe that these are also decent estimates for philosophy alone. My reasoning is this: (a.) In the chart linked to above, we can see that philosophy and religion are treated separately and together in 1999 and 2004. The aggregate estimate is always somewhat closer to philosophy than to religion, suggesting that there are somewhat fewer faculty in religion -- which also fits with recent data from the Survey of Earned Doctorates (e.g., in 2009 423 doctorates were awarded in philosophy and 323 in religion). (b.) The race and gender percentages for religion and philosophy in 1999 and 2004 aren't hugely different, though religion tends to be a little closer to parity than philosophy in both respects (which also fits with recent SED data, which finds religion to be almost as race and gender-skewed as philosophy). For example, averaging 1999 and 2004, philosophy has 83.2% men to religion's 81.0%, and 90.0% white non-Hispanic to religion's 84.6%. So, combined with point a, the percentages for philosophy alone are likely to be similar to the percentages for philosophy and religion together, perhaps philosophy alone a bit more white and male. Of course, because of the nature of this estimate, as well as the fluctuations noted in (2) above, we should keep in mind that these are rough estimates, probably good only plus or minus a few percentage points.