Jonathan Rauch has written a fascinating article for National Journal using data from the Pew Research Center. His argument is complicated, so let's reduce it to bullet points.

1) The electorate has grown more conservative over the last two years.

2)  Yet the share of the electorate that identifies with the GOP remains low.

3) The reason is a rise in the number of "de-branded Republicans" -- conservatives who have left the GOP and become independents because they believe it is not conservative enough.

4) The "de-branded Republicans" -- many of whom are Tea Partiers -- may dislike the GOP, but they absolutely hate the Obama Democrats.

5) Hence the de-branded Republicans will help the GOP make major gains in 2010.

6) However, 

Republicans' problem is that core conservative constituencies -- particularly white working-class and Christian voters -- are shrinking as a share of the electorate. Core center-left constituencies -- minorities, left-leaning women, professionals, and socially liberal Millennial Generation voters -- are growing. The demographic trends appear to require Republicans to expand beyond their conservative base just to keep from losing ground.

7) So, while the return of the de-branded Republicans to the GOP will help the party in the short term, the changing composition of the electorate will help Democrats in the long term.

Fair enough. But what if Points Six and Seven aren't quite accurate?

Rauch presents plenty of data showing that independents are becoming more numerous and more conservative. But he simply asserts that "core conservative constituencies" are "shrinking" while "core center-left constituencies" are "growing."

Now, it's true that America is becoming more diverse. But when you look at the composition of the electorate, things really didn't change too much between 2004 and 2008. Consider exit poll data. Latinos went from 8 percent of the electorate in 2004 to 9 percent in 2008. African Americans went from 11 percent in 2004 to 13 percent in 2008. Professionals, which I define here as voters with at least a college degree, went from 42 to 44 percent. And under-30s, or Millennials, went from 17 to 18 percent.

Those changes are not what you'd call statistically significant. (Evangelical voters, for example, went from 23 percent of the electorate to 26 percent. They didn't shrink, and McCain won them overwhelmingly.)

More important were the margins by which Obama won his groups: 95 percent of African Americans, two-thirds of Hispanics and Millennials, and (using another approximation of "professional") 58 percent of voters with graduate degrees.

For the sake of argument, let's attribute those margins to the particularities of the 2008 election. It was a very peculiar election, after all. Obama was the first African-American presidential nominee. There was a historically unpopular incumbent, two unpopular wars, and the economy had been in recession since December 2007. Maybe Obama will be able to maintain his high margins among key groups after four years of rule, during which time conditions will change and the historical novelty of his presidency will become a matter of course. Maybe not.

More important, in my opinion, is the behavior of independent voters. Like evangelicals, independents as a share of the electorate grew more than African Americans, Latinos, or Millennials; they went from 26 percent of the electorate in 2004 to 29 percent in 2008. And this shift helped Obama too: In 2004, independents went for Kerry over Bush by 49 percent to 48 percent, but in 2008 they backed the Democrat by a decisive 52 percent to 44 percent margin. Obama won the Democrats and the independents, and thus the presidency.

But the independents are now slipping away and becoming, of all things, more conservative. And if the Republicans can hold them in 2010 and 2012, and the composition of the electorate remains about the same as it was in 2000, 2004, and 2008, then Obama will have his work cut out for him.

Yes, those are assumptions. But it seems to me that Rauch makes assumptions of his own  -- that African Americans, Latinos, and under-30s will continue to grow to significant proportions of the electorate, and that current partisan allegiances will persist indefinitely -- when he starts talking about politics beyond 2010.