Supersoulty has the right idea. There are two huge problems with forecasts to 2030. One is that the census can project current trends, but can't anticipate new trends very well. Changes in immigration patterns as well as relocation patterens are likely to shift as they have with each generation. I'm comfortable looking at 2010, but my confidence drops a lot when I'm dealing with voters who aren't yet born.
The second factor is in the parties themselves. 2032 is 28 years from the last presidential election. 28 years before 2004 was 1976 (my first chance to vote ). Both elections were decided with the winner holding under 300 EV -- they're close. Compare the maps of those two elections and there are 24 states that voted differently. That's almost half. The South and Pacific Coast, and northern New England flipped. IL and MI were in the GOP camp in 1976. It's an even bet that another shift will occur sometime in the next 28 years.
I agree with distant population predicitions to be weak, but I disagree with your party analysis. The parties switched ideologies after 100 years sticking with on ideology. 1976 was a year when the change was still underway and so the election played out oddly with people voting Republic or Democrat for opposing reasons. I believe the current ideologies are here to stay for a long time.
Realignments dont only happen once every 100 years.
Also, if you look at the electorate today, would you not say that voters vote for the two major parties for vastly different reasons.
That is why this current allignment really bothers me, and I don't think it can hold. I would say the most natural allignment is essencially libertarian vs populist, and that was the way politics behaved for nearly all of the past 200 years. The current parties are an amalgamation of odd interests, that I don't think can hold together for very long.