Chait's main arguments 1 largely share a belief that Democrats simply need to elect better Democrats who are less beholden to parochial concerns or less attracted to filthy lucre. Add a dose of Republicanesque ruthlessness, and you would have a successful Democratic coalition.
The problem is that such analysis largely ignores or at best pays lip service to the structural and historic reasons for the Democrats' unruliness. In an evenly-divided nation with our political structure, it is harder to elect a center-left coalition than a center-right coalition. Because of this, and because of the way the parties developed, progressives/liberals tend to have a harder time governing than conservatives.
Structural barriers for liberals/progressives
Parochialism is embedded in our constitutional system. America does not have proportional representation, where many members from different parties can be elected from a single district. Instead, we have winner-take-all single member districts. In other words, you don't vote for "Democrat" or "Republican," you vote for Ben Nelson or Pete Ricketts. The party label is important, but often is secondary to the candidates and their uniquely-tailored positions.
This isn't mandated, we just inherited it from England. Indeed until Congress fairly recently mandated single-member districts, many states elected their members in statewide "at-large" elections (though these seats tended not to be allocated proportionately). Because we have geographic, single-member representation, Democrats "feel it is not just a right but a duty to slavishly attend to the interests of their home-state businesses" (as Chait puts it). Our system of government encourages - even requires - such behavior.
There's an additional complication. Because of this representation system, a Democrat running in Nebraska will be competing for a very different set of voters than a Democrat running in New York. He will have to adjust his voting habits accordingly. This means, in turn, that in order to win that Nebraska Senate seat, Democrats generally must elect a Senator who accepts some portion of the Republican agenda.
Because America remains a deeply centrist country - a recent study by the left-leaning Center for American Progress puts the average American's ideology at 209 on a 400-point scale, just a tick to the left of center - to gain any large majority, the parties will have to venture into ideologically hostile territory.
This would affect both Republicans and Democrats equally (witness the voting pattern of Senator Lincoln Chafee). The problem for Democrats comes from the nature of the districting schemes the chambers use. Senate districts are pretty much set in stone. Chait has a valid point that he does not emphasize enough when he says the small-state bias of the Senate presently favors Republicans. This bias - which is an integral part of our founding and Constitution - is a structural barrier to Democrats so long as Republicans continue to dominate rural areas, which tend to be conservative. To win a filibuster-proof majority, Democrats have to win at least 14 Senate seats in Republican-leaning states ("Republican-leaning" is presently defined as a state or district where Obama's margin was less than his national average of 7.2%). Republicans only have to win 6 Senate seats in Democratic-leaning states to win a similar majority.
The current House has built-in Republican biases as well. Democrats waste a lot of votes when redistricting. Democrats, joined by some Republicans (for different reasons) decided that it is important to maximize minority representation in Congress. The result is that the most reliable Democratic voters are clumped together into heavily Democratic districts. Because of this there are only about a dozen districts that lean Republican by twenty points or more, but there are over fifty districts that lean Democratic by twenty points or more.
This gives Democrats some stability, and a larger floor beneath which they cannot fall (especially in the South). But by clustering Democratic voters, Democratic representation is diluted, meaning that there are fewer seats that lean to the Democrats overall. That means that there will usually be more Democrats who answer to a conservative-leaning constituency than vice-versa (and paradoxically, the larger the Democratic majority, the more members answer to rightward-leaning voters).
Thus, in the 107th Congress (2000-2002) there were only 26 Republicans in Democratic-leaning seats, whereas in the 110th Congress (2006-2008), there were 53 Democrats in Republican-leaning seats. Republicans were able to obtain majorities roughly equal to those enjoyed by Democrats in the 110th Congress, without becoming nearly as exposed. This encouraged more "parliamentary" style voting among Republicans, especially when Bush's approval ratings were high, because almost all Republicans were representing "pro-Bush" districts. 2
This plays out as we would expect. A review of Poole-Rosenthal ideological ranking of Senators for the 110th Congress confirms that Democrats who represent conservative states have the most conservative voting records. Only two of the 16 Senators from Republican leaning states were in the left half of the Democratic caucus, and one of those two, Robert Byrd, had an anomalous session; he typically has one of the more conservative voting records for Senate Democrats. Of the 20 most conservative Democrats, 11 were from states McCain would eventually carry while two more (Bill Nelson and Evan Bayh) were from a states Obama would only carry narrowly. It's also not surprising that of the 40 most conservative Democrats in the 110th Congress, 20 of them were elected in the party's landslide year of 2006 - which pushed them far into Republican territory, and into the majority. The flip is also true - the more liberal districts elect Republicans who cluster toward the left of their party.
In short Ben Nelson and Kent Conrad don't slavishly attend to their parochial interests and compile conservative voting records because they are bad Democratic Senators. They are fulfilling the role the Constitution envisioned for them. The problem for Democrats is that (a) to get a super-majority they have to win seats in places like Nebraska and North Dakota, but (b) being a Democratic Senator from Nebraska or North Dakota necessarily means that you cannot embrace fully the national Democratic platform - at least if you want to remain a Democratic Senator for long.
Historic problems for liberals/progressives
The Democrats' problems aren't limited to Carter and Clinton. They have historically had a less ideologically unified coalition than the Republicans. This is why the agenda of the Democratic party has largely remained both unchanged and unfulfilled since Roosevelt's 1941 Message to Congress, notwithstanding the general Democratic dominance of the period.
Consider the following chart. If you want to know the methodology behind this chart, read this footnote3, but the bottom line is that when the line is at zero, the parties have equally heterogeneous caucuses. When the line dips below zero, the Republicans have a more diverse caucus, and when the line goes above zero, the Democrats have a more diverse caucus. As you can see, the relatively heterdox nature of the Democratic party is a historical feature.
From the beginning, the Democrats were a diverse bunch. Much is made of the "Solid South," which formed the base for the Democrats from 1832 through the 1950s. But the founders of the Democratic Party knew that this base wasn't enough to sustain victory. So they formed a coalition with machines in large Northern cities, in particular New York City. This gave Democrats another solid base of support, that allowed them to compete in the largest Northern state. Since 1836, New York County has been about as solid as the South has been; Democrats have only lost it four times: In 1848, 1896, 1920, and 1924. Even in 1836, New York County cast more votes than the average Southern state.
This marriage of big city machines and southern conservatives continued through the years, and by 1932, New York City was casting over half as many votes as the entire South combined. FDR's addition of progressives to the Democratic coalition in 1932, LBJ's and McGovern's addition of cultural leftists in the 1960s, and the gradual addition of black voters created a true patchwork party. As the chart above shows, by the 1950s, the Democrats were barely a party, and during the Carter and early Carter years they still had substantial internal divisions, even as they were shedding conservative Southern Democrats.
Republicans, on the other hand, have generally had a more ideologically coherent party, focused on a friendly attitude toward banking and business, encouragement of stable immigration and assimilation, and an embrace of religious social movements. There was a nasty split between progressives and conservatives from 1912-1932, but that has long since resolved itself with the expulsion of liberal Republicans (Specter, Collins, Snowe, and even Lincoln Chafee are not nearly as liberal as Jacob Javits, Clifford Case or Charles Percy once were). There are still tensions between Republican supply-siders and deficit hawks, and libertarians and SoCons, but these differences are comparatively minor compared to the differences Democrats endured between Representatives like Adam Clayton Powell and William Colmer.
Conclusion
The good news for Democrats is that this flexibility allows them to compete in just about every district in the country. This is part of why they have chalked up wins in odd places like LA-06 and MS-01. They have an easier time winning majorities than Republicans. The bad news, as Chait observes, is that this flexibility makes it harder for them to use their majorities to advance an ideological agenda.
But the better news is that these divisions in the Democratic party are dying out, and the party is much more ideologically cohesive than it was even ten years ago. Ike Skelton is conservative, but he is no John Rankin. Obama will probably have the most liberal Congress since 1964 or even 1932, as witnessed by the relative ease with which his budget package eventually passed both Houses of Congress.
But Obama still has a number of members from conservative districts who will have to watch their backs with their constituents. Until Democrats find a way to address the structural problems described above - and are willing to shed some members of their coalition in the process - they will continue to have a difficult time governing. In the meantime Obama will have to set an agenda that can be sold in some Republican-leaning states and districts, or risk having that agenda derailed.
[1] I should say that overall, Chait overstates his case. Bush was re-elected in 2004 with the largest congressional majorities of any Republican President since 1928, yet his second term domestic agenda was a failed one. Compare this with his first term, when negotiation with the Democrats was viewed as necessary. Bush's tax cuts are controversial today, but were favored by a 56 to 25 margin when Bush took office. Bush proposed $1.6 trillion in tax cuts, then accepted taking out about $300 billion from his initial tax cuts and adding some tax cuts favored by Democrats, and saw his tax cuts pass by a 58-33 margin in the Senate. His pre 9/11 summer was pre-occupied with pushing through No Child Left Behind (87-10 in the Senate) and negotiating a bill to allow patients to sue HMOs. In other words, Bush was successful in moving his agenda through at first because he actually did reach out to the other side.
In other words, it isn't just that Republicans have trouble governing with large majorities, its that large majorities give way to hubris, and cause Presidents to place their agenda further to the left or right than the public will accept. It doesn't affect Democrats any more than Republicans; its just that Republicans haven't had many instances of large majorities.
[2] It is even worse for Democrats in the present Congress. Today, Democrats have elected 66 Democrats in districts that lean toward Republicans, while only 15 Republicans represent districts that lean toward the Democrats.
[3] I began with Poole-Rosenthal first dimension DW-NOMINATE scores. Books have been written on this methodology, but the bottom line is that these scores are derived by looking at all legislators' votes, as opposed to a select few (as interest groups like ACU and ADA), to get a holistic view of the legislator. Votes on social issues, especially racial issues, are placed on the second dimension, and hence are not part of this analysis. Legislators are then assigned a score roughly from -1 (most liberal) to 1 (most conservative).
I divided the legislators (Senators and Representatives) for each Congress and sorted them by party. I then took the standard deviation for each party, which gives us a rough sketch of how ideologically diverse a Party's scores are for a given caucus. Then I subtracted the Democratic Party's standard deviation for a given Congress from the Republican Party's standard deviation, to see how much more ideologically diverse the caucuses are. Because this does not take into account social issues, it probably actually understates the Democrats' diversity, at least until the 1990s.
For those interested, there are two other measurements I looked at. First, I examined the number of "line crossers." What that means is that I looked at the number of Democrats in a Congress who had positive numbers (i.e. more conservative than liberal) and the number of Republicans who had negative numbers (more liberal than conservative). I subtracted them. The chart shows that, as we would expect, Democrats typically had more members with conservative records than Republicans had members with liberal records. This lends strength to the "always had a more diverse caucus" theory.
Finally, I measured the distance between the extremes of the parties and subtracted them. In other words, I took the average scores for the twenty most conservative Democrats and the twenty most liberal Democrats, and subtracted them. That gave me an idea how far apart the fringes of the Democratic party were. A high number here meant that there were a lot of conservative members on one side, and a lot of liberal members on the other. I did the same for Republicans, and then subtracted them. Again, a number above the line meant that the Democrats' fringes were further apart from each other than were the Republicans' fringes; the numbers here aren't as "clean" as we saw in the first two measurements, but the Democrats do tend to be more polarized at the extremes than the Republicans for much of the twentieth century. This would, in turn make it harder for Democrats to get their program through.
