Rick Perlstein urges Democrats to set out resolutely on the path of Barry Goldwater, on the off chance that it will lead to the promised land of Franklin Roosevelt. I see three big problems with that game plan.

First, the idea that any policy platform concocted by pundits and public intellectuals can provide the blueprint for a “dominant political party” is far-fetched. Dominant parties are, as it happens, quite a rare thing in American political history. Since the 1830s there has been only one instance of a party winning three consecutive presidential elections by as much as ten percentage points. (That party was the Republicans in the 1920s.) FDR’s New Deal coalition, which Perlstein takes as his model of what a dominant political party can be, was a product of voters’ responses to New Deal policies in action–certainly not a prospective endorsement of anyone’s long-term plan. (FDR ran on a balanced budget platform in 1932, precisely the sort of political trimming Perlstein is at pains to castigate in today’s Democrats.) And Ronald Reagan was elected in 1980 in spite of the principled conservatism he inherited from Goldwater, not because of it.

Second, Perlstein’s insistence that short-term popularity is detrimental to long-term political success is unsupported by any evidence that I can see. He notes that a list of ordinary people’s most pressing concerns “reads like the score for an Old Democrat symphony”: health care, earnings, education, economic inequality. So why are voters not singing along? In Perlstein’s view, because solving those problems “simply takes a while in the conception, in the execution, and, not least, in the political promotion.” But isn’t that all the more reason to make a start, however small, on actual solutions?

Perlstein’s prime example of counterproductive short-term thinking is Bill Clinton and Dick Morris’s “triangulating” in the wake of Clinton’s 1994 midterm rebuff. But he sidesteps the fact that Clinton’s electoral problems in 1994 stemmed in significant part from exactly the sort of go-for-broke strategy that Perlstein himself seems to long for. Of course the political fallout was “not for proposing health care, but for losing on health care,” as Perlstein puts it. But the fact that Clinton lost on health care is hardly unrelated to the fact that he insisted on proposing a reform package ambitious enough to cement Democratic loyalties for years to come–a package too ambitious even to make it to the floor of a Democratic Congress. That is what happens to jumbo dreams. In retrospect, how can any Democrat not wish that Clinton had been willing to settle for half a loaf? Or fail to see that doing so would have helped, not hurt, the party’s long-term prospects? Or imagine that a principled defeat in 1996 would have left the party or the country any better off?

Third, Perlstein’s notion that the Democratic Party “must give voters something to identify with” flies in the face of a good deal of evidence about trends in party identification, partisan attitudes, and voting behavior. Voting patterns have become increasingly partisan over the past two decades. At the presidential level, partisanship is a significantly better predictor of vote choice now than it has been at any point in the past half century. Nor is this simply a reflection of Republican inroads. In the 2000 presidential election, Democratic identifiers were more numerous than Republican identifiers and more loyal in their voting behavior.

Over the same period, partisan attitudes have become significantly stronger and more polarized. The proportion of ordinary citizens who see important differences between the two parties is at an all-time high, as is the proportion who correctly place the Democrats to the left of the Republicans on an ideological scale. People are increasingly likely to like one party and dislike the other; and they have more to say about the parties’ good and bad points than at any time since the 1960s. Perlstein’s complaint that “it’s hard to identify with a party when you don’t know what it stands for” reflects a different political world from the one most Americans are living in today.

The pattern of open-ended comments about the parties’ good and bad points also contradicts Perlstein’s characterizations of the Republicans as a party of clear, principled long-term stands and the Democrats as “the party of . . . nothing at all.” For example, in National Election Study surveys since 1992, people have had more things to say about the Democratic Party than about the Republican Party–and more good things to say about the Democratic Party than about the Republican Party. Moreover, the average number of good things people had to say about the Democratic Party increased by more than 20 percent during Bill Clinton’s eight years as president–precisely the period in which Perlstein sees Clinton, Morris, and the DLC diluting the Democratic brand.

Perlstein’s more specific concern that the Democratic Party is in danger of losing its historical identity as the party of the working class seems especially odd in light of the continuing economic polarization of the two parties’ supporting coalitions over the period he purports to be describing. Nolan McCarty and his colleagues have shown that the correlation between income and party identification has strengthened markedly over the past half century. In the 1950s, Republican identifiers were almost as common in the bottom quintile of the income distribution as in the top quintile; but in recent years the top quintile has been more than twice as Republican as the bottom quintile. Put another way, between the 1950s and the 1990s, the Democrats’ advantage over the Republicans in party identification declined by 13 percentage points in the upper third of the income distribution, by 19 percentage points in the middle third of the income distribution, and not at all in the bottom third of the income distribution. It should be obvious from this pattern that the erosion of the New Deal coalition over the past half century has little to do with the working class and much to do with defections among the middle and upper-middle classes. The problem is not the core but the periphery.

So what is to be done? For what it is worth, here is my game plan. First, win. Second, govern. Third, win again. Fourth, keep at it.

Click here to return to the New Democracy Forum “How Can the Democrats Win?

Originally published in the summer 2004 issue of Boston Review.