Demography is destiny in politics, or so we have heard. In 2004, the growth of the exurbs was said to be generating a permanent Republican majority. Now the strong support for the Democrats by young people, Hispanics, and non-Christians is said to be creating an unstoppable trend toward liberalism.

The demographic trends are real. National Journal columnist Ronald Brownstein recently illustrated how much they matter with a neat exercise. He divided the electorate into six broad demographic groups -- e.g., college-educated white voters and Hispanics -- and noted how each had voted in the McCain-Obama contest. "If each of these groups voted as it did in 2008 but constituted the same share of the electorate as in 1992, McCain would have won. Comfortably."

Yet demography isn't everything. The shift from unified Republican control of the government in 2005 to unified Democratic control in 2009 was not produced only or even mainly by demographic trends. The make-up of the country did not change that quickly. A lot of people who had voted for Republicans started voting for Democrats.

Those people are not easily categorizable. Republicans lost ground among Hispanics, whites, and blacks; among women and men; among voters with college degrees and voters without; among evangelicals and non-Christians; among libertarians and populists.

It is possible that Republicans will regain popularity, just as they have lost it, across the board -- if, for example, continued economic trouble becomes associated with the Democrats. Certainly there is no point in trying to add tiny demographic groups to the Republican coalition: The party is too far down to get a majority that way.

Indeed, Republicans are so far out of power right now that they will probably have to do what they should always have been doing: figure out the main challenges to the national interest and how to meet them. But even the most well-considered agenda will fail to accomplish anything if it is impossible to imagine how a majority of the electorate could ever be moved to support it. And the truth is that Republicans are going to have to choose which voters they are most eager to court. Time and money are limited, after all, and the actions that tend to please one type of voter will displease another.

The distinguished political journalist Michael Barone recently wrote that one of the key choices facing Republicans is "whether to go after downscale or upscale voters." The former tend to be "cultural conservatives, and rural and small-town voters," and to love Alaska governor Sarah Palin. The latter tend to be socially liberal and, though Barone does not underscore the point, to disdain Palin. Barone's tentative conclusion is that "going upscale is the right move." He points out that young, high-income voters were more likely to support Obama than to support House Democrats, suggesting that Republicans can win them over.

We think that Barone's tentative judgment is incorrect: To the extent that Republicans have to choose among which group to find new voters, they should look first to "downscale" voters without college degrees. Instead of fretting about Greenwich, Conn., the party needs to focus on the increasingly racially diverse working-class neighborhoods of New Jersey, Minnesota, and Colorado.

Though the college-educated represent a large and growing share of the electorate -- 45 percent in 2008, over a third of whom have post-graduate degrees -- the country still has a non-college-educated majority. Whereas John McCain won non-college-educated whites (39 percent of the electorate and shrinking) by 58 percent to 40 percent, he lost non-college-educated non-whites (16 percent of the electorate and growing) by an overwhelming 83 percent to 16 percent. Given the historic nature of Barack Obama's candidacy and the uneven impact of the economic downturn, this shouldn't be too surprising. Latinos, for example, are concentrated in hard-hit sectors and regions, and a disproportionately large number of non-white homeowners were impacted by the subprime-mortgage crisis. There is good reason to believe that Republican support among non-white voters has bottomed out. More to the point, it is crucially important that it has bottomed out.

McCain won college-educated whites (35 percent of the electorate and growing) by only 51 percent to 47 percent, and improving that number is certainly important for future Republican candidates. But it might actually prove more difficult and more risky than more aggressive outreach to "downscale" non-white cultural conservatives.

To understand why the young, upscale voters Barone mentioned may be hard for Republicans to reach, consider why Democrats have done so well with them in the first place. Many of them are clustered in dense, populous, high-cost communities, whether big cities or their inner suburbs. Government plays a more pervasive role here than it does elsewhere, and voters are socialized into believing that this is a good thing. (It helps that the state and local tax deduction insulates voters in these regions from the costs of governmental profligacy.) Elected officials, regardless of partisan affiliation, are expected to take an active role in managing the conflicts and trade-offs that inevitably emerge over traffic congestion, school funding, policing, and economic development.

One commonly held view is that "upscale" voters are voting against their economic interests out of distaste for Republican social views. This view ignores economic geography. The Center for an Urban Future, a centrist think tank, recently found that a middle-class lifestyle that costs $50,000 in Houston costs $72,772 in Boston and $123,322 in Manhattan. Many of the young, high-income voters who have flocked to the Democrats have done so because they feel financially strapped, and are eager to offload the burdens of acquiring health insurance and affordable housing onto the federal government.

The polling evidence suggests that these voters are as liberal on economic issues as they are on social issues. Among young voters overall, the Pew Research Center found that 69 percent favor an expanded role for government. And the more affluent voters are, the more likely they are to be ideologically consistent, i.e., liberal or conservative on social and economic issues. Shifting left on social issues would endanger the party's alliance with culturally conservative voters, a key part of Republican success over the last three decades, without sufficing to win over the upscale young.

An alternative strategy would largely maintain the Republican party's social conservatism while moving to the center on economic issues. That shift on economic issues need not take the form of supporting higher taxes. It would, rather, mean placing less emphasis on tax cuts for high earners and more on tax cuts for people in the middle of the income spectrum. It would mean working harder to get the public to associate Republicans with free-market policies to make health care more affordable and secure for the middle class.

The base of the Republican party is not averse to this type of shift. Midway through the 2008 election, pollsters Glen Bolger (a Republican) and Stanley Greenberg (a Democrat) conducted a survey for National Public Radio that included a "blind taste test" of party policies on a range of issues. The parties' positions and themes were presented both with and without party labels, and voters were asked to judge the position on the merits. And time and again, like Coke employees favoring Pepsi in a blind sample, Republican respondents preferred the Democratic position on domestic questions. On taxes, for example, the Democratic position called for rolling back the Bush tax cuts and focusing solely on middle-class tax relief. The Republican position called for renewing the Bush tax agenda, coupled with cuts in wasteful spending. When the Republican position was labeled as such, it was supported by 66 percent of Republicans. But when it was not labeled as the GOP position, it was supported by only 38 percent of Republicans -- and a narrow majority of Republican voters actually preferred the Democratic line.

A shift to the center on economics would do more than bring the Republican party in line with the views and priorities of its base. It would also allow the party to fight more effectively for downscale voters who are culturally conservative but consider the GOP unresponsive to their economic concerns. Many of these voters are, of course, non-white. As the demographic composition of the country changes, the electoral value of Republican dominance among married white Christians is eroding. But adding married black and Latino and Asian cultural conservatives could revive the party. (There is, incidentally, absolutely no reason to think that young single people will keep voting Democratic in the same percentages as they become older married people.)

The downscale strategy would involve applying old conservative principles in creative new ways, while the upscale strategy would involve jettisoning a fair number of those principles altogether. The downscale strategy might even help Republicans both directly and indirectly with upscale voters. In a post-election analysis, Greenberg surveyed the relatively well-off suburbanites of Oakland County, Mich., who have been leaving the Republican party, and found that among their reasons for supporting Obama were his support for middle-class tax cuts and expanded health-insurance coverage. Would Republicans turn off these voters with an economic agenda geared toward the lower middle class? We know that a Democratic agenda thus targeted has not done so.

A Republican party that advanced downscale cultural conservatives' economic interests, meanwhile, would not need to lean so heavily on their cultural resentments to win their votes. Republicans' caricaturing of Democrats as effete and unpatriotic latte-sippers has reinforced the GOP's own reputation as anti-intellectual and philistine, and this reputation has harmed it in upscale precincts. An economic agenda more attractive to the country would reduce the party's reliance on cultural polarization.

Historical trends also favor the downscale strategy. The movement of affluent social liberals into the Democratic party and working-class social conservatives into the Republican party has been going on since at least 1964, when Rockefeller Republicans defected in large numbers to vote for LBJ. Going after lower-middle-class voters would build on the trend of 1966-2004 and require reversing only the trends of the last four years. The upscale strategy, on the other hand, would require reversing trends from 1964 through the present day. It's a taller order.

A downscale strategy would also serve the public interest. Many of the desires of affluent, socially liberal voters are worth attending to. Republicans should try, for example, to address their environmental concerns. But economic stagnation and family breakdown are taking a big toll on the lower middle class and the American dream. An economic and social agenda that helps lower-middle-class families get ahead -- by, for example, encouraging marriage and reducing the cost of raising children -- ought to be an urgent priority. It ought not be sacrificed for a misguided political strategy.