The latest word out of the Senate is that if Republican Minority Leader Mitch McConnell doesn’t accede to changes in the filibuster rules over the next few days, Harry Reid will invoke the so-called “nuclear option” and change the rules with 51 votes. The most likely outcome would be outlawing the filibuster on motions to proceed, thereby forcing senators to take to the floor to filibuster bills, “Mr. Smith”-style.
This move, and the overwhelming progressive enthusiam for it, are head-scratchers. Over the short-to-medium term (and no one can really see beyond that), the filibuster probably helps the Democrats more than it helps the Republicans. Before going any further, let me make clear that the following argument is couched purely in terms of political advantage and ability to move the agenda. I think there's a lot to be said for what we might call the small-"c" conservative arguments for the filibuster: requiring 60 votes creates the need for some sort of consensus before legislation moves through, and the chances of a destabilizing period of time where parties trade majorities and implement wildly divergent agendas willy-nilly are greatly diminished. In that sense, the filibuster helps the entire country, and both parties should be pleased with it.
That said, let’s analyze the politics of the filibuster, beginning with the following observation: The filibuster doesn’t really matter unless you control the House of Representatives, Senate, and presidency -- what we might call the “trifecta.” Even if Reid were to lower the number of votes needed to move legislation through the Senate to 20 votes, it still wouldn’t significantly advance the Democratic cause in Congress, because the Republican House acts as an effective filibuster. Similarly, when Republicans hold the presidency, a veto would stop any legislation.
Sure, a Democratic Senate sans-filibuster could pass legislation that might make a Republican House or president uncomfortable, and would be able to prevent its most vulnerable members from casting difficult votes (Blanche Lincoln, for example, would not have had to cast the deciding vote for Obamacare, and Democrats might still hold her seat). And to the extent that Republicans might try to block judges or cabinet appointees, the filibuster matters. But that is a two-way street (Lincoln Chafee might still be around had he not had to vote for parts of President Bush’s agenda), and in terms of advancing a progressive agenda, without the trifecta, the filibuster is largely superfluous.
So here’s the problem for Democrats: Republican trifectas are more likely, all other things being equal, than Democratic trifectas, at least in the near future. Let’s assume for the sake of argument that Democrats have a structural demographic edge in the presidency (I don’t agree, but let’s assume). Even the most rabid defenders of what we might call the Emerging Democratic Majority thesis concede that Republicans will still win the presidency if there is a recession, unpopular war, or other national upheaval.
Those contingencies happen with some regularity. So setting aside the fact that in the long run, presidential races sort out much the same way as do coin flips (the parties have each won the popular vote 20 times since the Republican Party was founded), let’s assume that in the short-to-medium term Democrats will win the presidency two-thirds of the time.
The House is another matter entirely. While it isn’t impossible for Democrats to retake the House (even in 2014), it is difficult. This is a subject worthy of a separate article, but given the current redistricting lines, and given how the Democrats’ coalition has sorted out into tightly packed geographic constituencies (urban liberals, minorities crammed into minority-majority districts), it makes a switch unlikely except in wave years.
Take 2012 as an example. With an electorate that featured one of the most favorable demographic tilts toward Democrats in recent memory, with a president winning by a decent margin, and with Democrats even winning the popular vote for the House by a point, Republicans won the third-largest number of seats they have enjoyed after an election since the 1920s.
The fact that we assume Democrats will dominate presidential elections works against them here as well. Since we began regularly holding our House elections in even years (around the Civil War), a party has captured control of the House in an election held while that party also held the presidency exactly once: 1948.
So let’s assign Democrats a 20 percent chance of winning the House. That leaves us with the Senate. The Senate is a natural GOP gerrymander. Consider: Mitt Romney lost by 3.8 points, but still carried 24 states. John McCain lost by 7.3 points, but carried 22 states. On the other hand, when John Kerry lost by only 2.5 points, he carried just 19 states. Twenty-seven states currently have Republican PVIs, meaning that in a completely neutral environment, we’d expect them to vote for a Republican.
Over the long term, this translates into an advantage for Republicans, since presidential results are fairly predictive of where Senate races shake out. Obviously there are shortcomings with this approach: Democrats run better in West Virginia than recent presidential results would suggest, while Republicans run better in Maine. Poor candidate choice (think Alexi Giannoulias in Illinois, Todd Akin in Missouri) can disproportionately hurt the parties, while unfavorable environments combined with overexposure (a la Republicans in 2008, Democrats in 1980) can result in disaster. But over time these contingencies should cancel each other out, and Republicans should tend to win control.
To try to quantify this a little bit better, let’s do something a little bit outside the box and run a rudimentary Monte Carlo simulation for Senate races. This is an old statistical technique (popularized recently by Nate Silver at FiveThirtyEight), which more or less allows a researcher to generate hypothetical results from a set of variables.
So we might say that a Democrat in Colorado can be expected to win 53 percent of the vote on average, and will be within five points of that 95 percent of the time, while a Republican in Idaho will be within five points of 63 percent of the vote 19 times out of 20. Monte Carlo simulations will generate random results within those confines.
If you run 100 simulations, the Democrat should, on average, get 53 percent of the vote in Colorado, and will almost always be between 48 percent and 58 percent, while the Republican will almost always be between 58 percent and 68 percent in Idaho. If you run 1,000 simulations, you would be able to get a pretty good estimate of how often Republicans would win both seats in that hypothetical two-seat Senate (about 27 percent of the time).
The trick here is estimating what percentage of the vote a state will tend to cast for a Republican, all other things being equal. Presidential performance gives a decent base line. The rest -- retirements and incumbency, political environment, exposure, candidate quality -- matter in the short term, but should cancel out over time.
Of course, President Obama’s numbers likely overstate Democratic strength somewhat in Illinois, and understate it in West Virginia. To help correct for that, we’ll take the average of the past five presidential elections in each state.
But a state that gave the Democrats, say, 62 percent of the vote on average wouldn’t elect a Republican 62 percent of the time; that number would approach zero. Likewise, a state that gave Democrats 45 percent of the vote on average probably wouldn’t result in a Democratic win more than one time in four. To correct for this, a simple linear transformation is applied to these averages, effectively making a state that gave Democrats 60 percent a sure-fire win for that party, while the opposite is true for states that gave Democrats less than 40 percent of the vote.
Finally, I took the standard deviation of these five elections for each state. The basic idea is that a state that gave Democrats 52 percent of the vote in all five elections shouldn’t be expected to move much off of that 52 percent average, but a state like West Virginia, where results have ranged between 36 percent and 58 percent of the vote for Democrats, should be allowed to move substantially off the median.
Using these as our metric, and running our simulation 2,000 times, we find Republicans winning control of the Senate 74 percent of the time, with 51 seats on average. This is encouraging, since they have won control of the Senate after six of the past 10 elections, and have won 50 seats on average during those same elections.
Over the course of 2,000 elections, things would probably end up pretty close to what our simulation predicts; for that matter, if 2008 hadn’t been such an unusually bad environment for Republicans (and if Ted Stevens hadn’t been convicted of a felony on the eve of the election), our simulation would likely be spot-on. And even if this is a touch too optimistic for Republicans, our House and presidential estimates are probably a touch too optimistic for Democrats.
Already we can see the problem for the Democrats -- control of just the Senate is very difficult for them to maintain over the long haul. At the same time, Republicans should never achieve a filibuster-proof majority; the best they do in our simulation is 59 seats, once every 1,000 elections.
Moreover, if our estimates are right (Republicans win the presidency 33 percent of the time, the Senate 74 percent of the time, and the House 80 percent of the time), Republicans should win the trifecta 20 percent of the time, versus just 3 percent of the time for Democrats. And again, Republicans held the trifecta after three of the last 10 elections, while Democrats won it in just one (held in extraordinarily favorable circumstances). In truth, the odds are probably a little better for both parties to win the trifecta, since these elections aren't completely independent variables, but that affects both parties and does little to affect the underlying observation about the relative likelihood of Republican trifectas vis-a-vis Democratic ones.
Weakening the filibuster was a perfectly sensible goal for liberals back in the 1970s, when Democrats arguably had natural Senate and House majorities, and their main goal was to diminish the ability of Southern Democrats to cross the liberal leadership. But it makes much less sense today. Yes, Democrats can point to some cherished action items that were lost in the first half of Obama’s term (the public option, card check).
At the same time, however, the filibuster greatly restrained Republicans’ ability to implement their agenda during the Bush years. Without it, they probably would have passed tort reform, ANWR drilling, Social Security privatization, school vouchers, made the Bush tax cuts permanent, and further diminished unions’ ability to organize. Republicans might have passed immigration reform, President Bush almost certainly would have placed Miguel Estrada on the Supreme Court, and Hispanics might be a more reliably Republican voting group. In short, in terms of policy changes, the filibuster has probably inhibited Republicans more over the past decade than Democrats.
Overall, that’s what makes the move such a head-scratcher. There’s no immediate benefit for Democrats, little short- or medium-term benefit, and some potentially catastrophic downsides for them during that time frame. Unless their coalition changes in the coming decade or two, Democrats are probably more likely to rue these changes than to celebrate them.