locating members of Congress along a left-right spectrum. (They also identify a second dimension of politics—race —which has been crucial in the rise of movement conservatism. But let’s leave that aside for now.) The method, roughly speaking, works like this: Start with roll-call votes on a number of bills that bear on economic issues. First, make a preliminary ranking of these bills on a left-to-right political spectrum. Second, rank members of Congress from left to right based on how they voted on these bills. Third, use the ranking of legislators to refine the left-right ranking of the legislation, and repeat the process all over again. After a few rounds you’ve arrived at a consistent ranking of both bills and politicians along the left-right spectrum.[5] Poole, Rosenthal, and Nolan McCarty of Princeton University have applied this method to each Congress since the nineteenth century. What stands out from their results is just how modest the differences between Republicans and Democrats were in the fifties and sixties, compared with a huge gulf before the New Deal, and an even larger gap today.
Poole and Rosenthal measure the gap between the parties with an index of political polarization that, while highly informative, is difficult to summarize in an intuitive way. For my purposes it’s sufficient to look at two descriptive measures that behave very similarly to their index over time. One measure is what I’ll call “minority- party overlap”: the number of Democrats to the right of the leftmost Republican, when Republicans controlled Congress, or the number of Republicans to the left of the rightmost Democrat, when Democrats controlled Congress. The other measure is what I’ll call “minority-party crossover”: the number of members of the minority party who are actually on the other side of the political divide from their party—Democrats who are to the right of the median member of Congress, or Republicans to the left. In each measure more overlap indicates a less polarized political system, while the absence of overlap suggests that there isn’t a strong political center.
Table 2 shows these numbers for three Congresses: the 70th Congress, which sat in 1927–28 and 1928– 29; the 85th Congress, which sat in 1957 and 1958; and the 108th Congress, which sat in 2003 and 2004. The table shows that congressional partisanship was much less intense in the 1950s than it had been before the New Deal— or than it is today. In the 70th Congress, in which Republicans controlled the House of Representatives, there was hardly any minority party overlap: only two Democrats were to the right of the leftmost Republican. And there was no minority party crossover: all Democrats were left of center. The situation was even more extreme in the 108th Congress, which was also controlled by Republicans: Every Democrat was to the left of the leftmost Republican, and needless to say there was no crossover. In the 85th Congress, however, which was controlled by Democrats, there were many Republicans to the left of the rightmost Democrat (largely because there were a number of quite conservative Southern Democrats.) More amazingly, nine Republican members of the House were literally left of center—that is, voted to the left of the median Congressman. That’s a situation that would be inconceivable today. For one thing, a twenty-first century Republican who took a genuinely left-of-center position would never get through the primary process, because movement conservatives would make sure that he faced a lavishly funded challenger, and because Republican primary voters, skewed sharply to the right, would surely support that challenger. In the fifties, however, Republicans couldn’t afford to enforce ideological purity if they wanted to win elections. As a result, actual liberals like Nelson Rockefeller and Jacob Javits, who would have been summarily excommunicated today, remained party members in good standing.
| Table 2. Measures of Similarity Between the Parties | ||
|---|---|---|
| Minority Party Overlap | Minority Party Crossover | |
| 70th Congress, 1927–29 | 2 | 0 |
| 85th Congress, 1957–58 | 112 | 9 |
| 108th Congress, 2003–4 | 0 | 0 |
Source: www.library.unt.edu/govinfo/usfed/years.html.
The relative absence of difference between the parties’ positions on economic policy meant that voting behavior on the part of the public was very different from what it is today. In recent elections partisan voting has been very strongly correlated with income: The higher a voter’s income, the more likely he or she is to vote Republican. This presumably reflects voters’ understanding that a vote for a Republican is a vote for policies that favor the affluent as opposed to the poor and working class. But the relatively nonideological nature of the Republican Party in the fifties, reflected in the way its members voted in Congress, was also reflected in public perceptions. During the postwar boom, voters evidently saw very little difference between the parties on economic policy, at least when voting in presidential elections. Table 3 compares the average voting patterns of white voters grouped by income level in presidential elections between 1952 and 1972 on one side and 1976 and 2004 on the other. In the more recent period there was a strong relationship between higher income levels and voting Republican. During the period from 1952 to 1972, the era of bipartisan acceptance of the welfare state, however, there was hardly any relationship between income level and voting preference. The one presidential election in which there was a large voting difference by income level was 1964, the year in which Barry Goldwater—a true movement conservative, and the harbinger of things to come—seized the Republican nomination. Other surveys show that in the fifties and sixties there was remarkably little relationship between a voter’s income and his or her party registration: The upper third of the income distribution was only slightly more Republican than the middle or lower thirds.
| Table 3. Percentage of Whites Voting Democratic in Presidential Elections, by Income group | ||
|---|---|---|
| Percentage Voting Democratic in 1952–1972 | Percentage Voting Democratic in 1976–2004 | |
| Poorest Third | 46 | 51 |
| Middle Third | ||
