Binder also shows that moderates in Congress are crucial: the fewer the moderates, the bigger the hurdle to enacting significant legislation, a finding consistent with research by Princeton professor Nolan McCarty on legislative productivity during periods of congressional polarization. The decline of moderates, furthermore, is changing the nature of governing even during periods of unified government. Until the past 20 years or so, even when one party controlled the White House and the Congress, major pieces of legislation were traditionally bipartisan. The 1965 legislation creating Medicare is one example; Democrats controlled the White House, House, and Senate, but almost half the Republicans in Congress also voted for the legislation. The 1993 budget deal and the 2010 health-care reform act, by contrast, are likely to represent the new model: major legislation enacted on a partisan basis. Both bills passed without a single Republican vote.
Most of the policy-making and punditry world still yearns for the days of the Medicare deal, but they are largely if not entirely gone. When I became director of the Office of Management and Budget, I was lucky because I had just come from being director of the Congressional Budget Office -- and since the CBO is a nonpartisan agency, I had good relationships with many Republicans. At my confirmation hearing, one of those Republicans said that even his Republican friends seemed to like me -- and then asked how long I thought this feeling of bipartisanship would last in my new job. He was prescient: Many of those relationships quickly frayed. Indeed, although a few agencies like CBO operate on a nonpartisan basis, there are now remarkably few people in Washington who work well with members of both parties, a reflection of the polarized environment.
The broader lesson for presidents is clear: Unless they have sufficient votes to legislate only with members of their own party, they're much less likely to enact major legislation -- that is, without surrendering key policy objectives to the opposite party.
WHERE HAVE THE MODERATES GONE?
So what's happened to the moderates? As the data assembled on Voteview.com (a website maintained by political scientists Keith Poole and Howard Rosenthal that contains data on congressional voting patterns) demonstrates, the most conservative Democrats in Congress 50 years ago often voted together with the most liberal Republicans. That common ground was dwindling by the 1980s; today, it is almost non-existent. Much of the shift has occurred because Republicans have become substantially more conservative.
A popular explanation among pundits for the rise of polarization and the decline in moderates is gerrymandering -- that districts have been redrawn to make them safe for one party or the other, allowing more partisan representatives. But it's mostly wrong. One study by Sean Theriault of the University of Texas at Austin showed that only a tenth to a fifth of the rise in polarization since the 1970s can be attributed to redistricting (other estimates are even lower). Nor is polarization just an inside-the-Beltway phenomenon, as some have suggested. From 1996 to 2008, most state legislatures also experienced striking increases in polarization, according to data assembled by McCarty and Boris Shor of the University of Chicago. If anything, over that period, most state legislatures polarized even more rapidly than Congress did.
Indeed, the polarization of our elected officials partially reflects the growing polarization of the public. We are increasingly surrounding ourselves physically and virtually with like-minded people, who then reinforce our biases and drive us further apart. In The Big Sort, for example, Bill Bishop documents increased residential segregation by political party. Over the past several decades, we have voluntarily separated ourselves into Republican and Democratic neighborhoods. Americans are also increasingly choosing to live near people in their own income bracket. Sean Reardon and Kendra Bischoff of Stanford University found that nearly two-thirds of American families lived in middle-income neighborhoods in 1970; by 2007, only 44 percent did. Since income is strongly related to voting patterns, this phenomenon may help explain the rise in residential segregation by political party.
The residential segregation by party, in turn, is reinforced by a splintered media market. Research suggests that Americans only tune in or log on to a small share of the media choices available to them, and they often pick the ones that fit their beliefs. The consequences are far-reaching. As Cass Sunstein emphasized in his book Going to Extremes, "When people talk to like-minded others, they tend to amplify their preexisting views, and to do so in a way that reduces their internal diversity."
GOVERNING IN THE AGE OF NO MODERATES
One obvious victim of this rise in polarization is the centrist legislating that has been the norm for most of postwar history. So what can be done? We can take on the problem on three different levels: encouraging less polarization in the population, increasing the number of moderates in Congress, and finding ways to govern effectively given a smaller number of moderates in Congress.
The first level involves dampening the degree of polarization in the population itself. To the extent that polarization is being driven by increased income inequality, one pathway to constraining it is to pursue policies (such as a more progressive tax code) that narrow after-tax income gaps. Another pathway is to highlight the importance of diversified interactions, both physically and virtually, to avoid the extremist tendencies that occur with self-reinforcing views. In speeches on college campuses, for example, I've been trying to communicate to students the importance of reading and listening to arguments from those with wildly different political views -- because doing so will not only better inform their own thinking, but may also provide a broader social benefit by mitigating polarization. We should also remember that polarization has occurred in long waves in the past. Epochal shifts, such as world wars and the Great Depression, have brought Americans together and attenuated schism. It is possible, though not pleasant, to imagine similar catastrophic events that would offset the intensifying polarization among the public and in Congress.
The second level focuses on the gap between elites and the population. Whatever the degree of polarization in the population, what can be done to dampen the signal by the time it reaches Congress? Changes in districting laws could help, even if much more modestly than the punditry believes, to mitigate the loss of moderates in Congress compared to the underlying population. The greater degree of polarization in Congress than the population also suggests an opportunity: focus on electing moderates to Congress from districts where polarization is least severe. It is not in the interests of either party to do so, which means the effort would need to be driven primarily by non-party leaders, including from the business community. Indeed, efforts to elect centrist candidates such as Americans Elect would do much better to focus on Congress rather than trying in vain to elect a third-party candidate as president.
The third and final level addresses the problem of how to govern assuming a given level of polarization and a general absence of moderates in Congress. Within this category, two approaches suggest themselves. The default approach is for both parties to aim for partisan dominance and unified government, and then to legislate on a partisan basis. This partisan legislative approach represents an understandable and potentially effective response to polarization (at least from the perspective of legislative success). If you can't govern in the middle, might as well govern -- even if it is from one side.
Placing our bets on a partisan legislative model has substantial downsides, however. For one thing, political dominance across the White House, House of Representatives, and Senate is a relatively rare phenomenon, so the approach will often be frustrated in practice by some degree of divided government. Furthermore, even when political dominance occurs, the model represents a fundamental shift from the bipartisan norm that has dominated American policy-making during the postwar era. Finally, the model is inherently unstable, as the 1993 budget deal and the 2009 and 2010 stimulus and health-care examples suggest, since it is likely to create a backlash so strong as to eliminate the political dominance that allowed it. That backlash, in turn, risks making the underlying polarization even more extreme.