Thursday, April 24, 2014

Are Comity and Bipartisanship Inconsistent?

Much has been made of the marked increase in partisanship in the United States (at large/in Congress/among the elites) over the past 15 years or so.

The commonly presented evidence for this is is the difference between the estimated ideal points of Democratic and Republican Congressmen and Senators. [0]

This, in many ways, is good evidence.  For example, it is based on an explicable theory of voting, it provides a neat (aka, unidimensional) measure of polarization within either chamber of Congress, and it is based on a behavior that has (some) teeth in it: individual roll call votes in Congress.

But before we walk away, shaking our heads and fists slowly but with a sultry sense of superiority, let's ask ourselves...what's at the heart of that measure of polarization, again?

Well, unsurprisingly, roll call vote scaling is based on roll call votes.  In Congress (and in many, but not all legislatures), roll call votes are not required on all matters.  This point has been well-documented by scholars (inter alia, here and here).

I wanted to point out something related to, but separate from, the point that not every legislative matter is subject to a (recorded) roll call vote.  Pushing on the point that roll call votes are costly in both direct (i.e., time) and indirect (i.e., position-taking) ways, there is reason to believe that large majorities---and particularly those that cross party lines---will tend to be "under-recorded" in the observed sample.

To make the presentation as clear and succinct as possible, note that most recorded roll call votes are preceded by a formal or informal unrecorded, but public, vote (e.g., a voice vote). 

1. Roll Calls Take Time.  To see the first (direct cost) line, consider the following situation: suppose that the voice vote reveals that 80% of the chamber favors passage of the measure/motion. (The logic would be similar is 80% favored defeat of the measure/motion.) Leaving everything else aside, if time is money (and it is), then the incentive to take the time to "record" this vote through a roll call, ceteris paribus, is lower than if the voice vote revealed (say) something like 55% putatively in favor of the measure/motion.

Thus, large majorities are less likely to be recorded in the roll call record.[1]

In line with the title of this post, if the parties are cooperative at least at the basest of efficiency levels, then large bipartisan majorities are less likely to be recorded, because recording a roll call vote takes time...and time is money.

The Americanists out there will note that this argument binds more quickly upon the Senate, given how it records its roll call votes.

2. Agreeing To Disagree...In Private.  To see the second (position-taking) point, suppose that the voice vote reveals a nontrivial number of members of each party voted "out of step" with (say) their leadership.  That is, suppose that 20 of the 233 currently sitting Republicans in the House of Representatives voted against a measure, with 213 voting in favor it, and 20 sitting Democrats voted for the measure, while 179 voted against it.  The vote totals for the margin are identical to the partisan breakdown (233 in favor, 199 against), but 20 members of each party are "exposed" as breaking with their leadership, on each side, so that 40 members are (arguably) more vulnerable if these votes are exposed.[2]

To keep it short, suppose that the party leaderships are cooperative and decide that they will agree to agree not to call for a roll call vote when they each have roughly equal numbers of defectors.  Putting aside the details, this can be supported in equilibrium through a variety of ways, including the very plausible idea that each leadership prefers the devil they know (i.e., they'd prefer to have a seasoned maverick who is marginally tougher to control in their ranks than the untested and presumptively unreliable upstart challenger).

Putting #1 and #2---the direct costs and the cooperative leaderships arguments---together amplifies the possibility of a conflict between comity and estimated polarization. In other words, it is entirely plausible, given the marginal incentives to both use cheap ("voice") votes and preserve the appearance of intra-party unity, that bipartisanship (in terms of voice votes crossing party lines/mingling party members) would lead to a censoring of roll call votes,[3] and this censoring would be---for independent reasons---weighted in favor of recording only roll call votes that were (1) close and (2) partisan.

Thus, if the party leaderships did cooperate---a central foundation of interparty comity---the roll call vote record would look (in both layman's and "scaling" terms) more polarized.  In other words...

to the degree that the 2 parties pursue greater comity through agreeing to not record consensual and/or "mutually party crossing/splitting" votes, the resulting roll call voting record would look increasingly polarized/partisan.

Is this what has driven the estimated increase in polarization over the past few decades?  Well, the point here is that it's hard to tell.  Sometimes all the love in the world will just accentuate the subjective nastiness of our fights.

Put another way: sometimes we care about how nasty the nastiest fight is (they've been pretty nasty lately),[4] and other times we care about how often we fight at all.  How we, as a nation, have moved on that second dimension is less clear.  Why do I think about Diff'rent Strokes right now?

Before I leave, let me say how happy and honored I am to be part of the Mischiefs.  My putative role here (since I usually just prattle on about math and politics) is a bit odd and concomitantly fun. If you have ideas for something related to party politics that you'd like me to pose a mathematical/strategic/contrarian take on, email me at

[0] Of course, other evidence includes casual observation, Facebook, twitter, and the nightly news.
[1] To keep the presentation "as gripping as possible," I will omit the nonetheless technically relevant point that sufficiently lop-sided roll call votes are (understandably/correctly) dropped from scaling analyses for technical reasons.  Including this consideration would merely strengthen my conclusions.
[2] This raises a point about "pairing" that I have thought about before, but not done anything with.  Perhaps somebody else already has.  If you're reading this and interested in this...first, get some sleep, and then, if you wake up still interested, email me.
[3] To be quick about it, note at this point that the second point does depend on the party leaderships fearing "credible" intra-party challenges.  Think Blue Dogs and Tea Parties.
[4] Or, "A `Miss Jackson' Theory of Congressional Politics."

Wednesday, April 23, 2014

Was Paul Ryan A New Kind of V.P. Candidate?

The Ryan budget is in the news once again, attracting questions about the issues that will shape the 2014 midterms. Worth mentioning, however, is that only in the last 20 years – since the Contract with America in 1994 - have Congressional campaigns consistently focused on national party agendas.  Because of this nationalization of campaigns, Ryan, the chair of the House Budget committee, can establish the terms by which voters will evaluate Republican candidates in November.

Others have suggested that his place on the losing presidential ticket in 2012 has boosted Ryan’s stature nationally and within the Republican Party. By itself, this isn’t remarkable. Other losing veep candidates, from the formerly obscure Sarah Palin to Joe Lieberman (who was running for his third Senate term when he was nominated) have probably enjoyed a higher national profile than they otherwise would have.

In contrast with most other modern VP picks, Ryan joined the ticket in the middle of his Congressional career, and returned to Congress as the chair of a prominent committee. The typical model for modern running mates has been to choose someone with little or no national experience - Palin, Spiro Agnew, Geraldine Ferraro, John Edwards - or someone closer to the end of his career. Among modern Republicans, running mates who have retired from Congress have made several appearances – George H.W. Bush, Dick Cheney, Jack Kemp.

In 2012, there was little doubt that if the Romney-Ryan ticket lost, the Wisconsin Republican would return to Congress (his district has typically returned him to Washington with comfortable margins). Appearing in the second slot is a risky proposition for a mid-career politician. A loss could derail a career, especially if the running mate is perceived to have contributed to the loss. Winning could be even worse for a politician with his or her own presidential aspirations - only one sitting VP has been elected president since Martin Van Buren (Bush). As a result, research on the "veepstakes"  suggests that it’s not uncommon for those who are asked to say no. We've progressed past Daniel Webster's declaration that, "I do not propose to be buried until I am dead." A few unsuccessful vp candidates, like Edmund Muskie and Bob Dole, went on to have strong careers - though neither became president.

It’s possible that the current focus on national messaging in the Republican Party – building a brand around deficit and budget politics as well as opposition to the ACA – has changed this calculus somewhat. After his appearance on the presidential ticket, Ryan is even better poised to be the face of the party and to shape its national message. If appearing on an unsuccessful ticket can boost a political profile, then we may start to see more ambitious, mid-career politicians in the second slot. 

Monday, April 21, 2014

Are Democrats Pervs? Some Problems with a State-Level Analysis of Individual-Level Behavior

Pornography aggregator Pornhub put out a report recently looking at porn use by state and noted that blue states seemed to be watching more than red states. The data presentation really cried out for a scatterplot, though, that examined presidential vote shares rather than just red/blue status. Chris Ingraham at the Washington Post made one (via Dylan Matthews). Here's my version:
That's actually a pretty strong relationship. Porn pageviews explain 16 percent of the variance in state level presidential vote shares. Each per capita pageview is associated with a two-tenths of a percentage point increase in a state's Democratic vote share, and this is statistically significant at the p≤.01 level. The one big outlier is Kansas, a pretty conservative state that nonetheless watches huge amounts of porn. (The R-squared nearly doubles when you remove Kansas.)

Of course, this probably isn't explaining anything real. For one thing, there's a big potential ecological inference problem here. We're making assumptions about individual level behavior by examining data aggregated at the state level. We really don't know if this would apply at the individual level (that is, whether people who watch more porn are really more likely to vote Democratic).

Second, chances are that even if there is an individual relationship here, it's not a direct one. Porn usage may correlate with something else that also correlates with partisan voting patterns. This could be poverty, internet speed and availability, age, marriage rates, etc. That is, people's interest in porn probably doesn't directly translate into their preference for one of the parties.

Nonetheless, fun graph.

Update: Dylan Matthews figured out what was the matter with Kansas. Apparently the geographic data are based on IP addresses, and when an IP address can't be determined for anywhere in the country, it defaults to the center of the country, which is roughly Wichita.

Why Are Conservatives Obsessed with Reagan? He Coincided with Ideological Party Takeover Like No Democratic President Did

Jonathan Chait finds a great quote from conservative pundit Jennifer Rubin, who proclaims that Rand Paul's unbending opposition to war with Iran "definitively de-Reaganizes him." Why do conservatives spend so much time talking about Ronald Reagan? Why is it so important for Republican politicians (especially those running for president) to behave as Reagan supposedly would? I think the key to understanding this is in Hans Noel's new book. Noel shows that modern liberalism and conservatism were developed by pundits in the early to mid twentieth century and gradually took over the two parties.

The Reagan personality cult comes from the fact that the ideological take-over at the presidential level happened differently in the two parties. For Democrats, the party's successive presidents gradually adopted more aspects of modern liberalism's tenants. Truman was the first to endorse racial liberalism and the Kennedy/Johnson administration solidified and extended that embrace, while it wasn't until Clinton that a sitting Democratic president clearly endorsed abortion rights, and Clinton and Obama took successively more liberal positions on LGBT rights. But on the Republican side, there is a clear break at the presidential level pre and post Reagan. Eisenhower, Nixon, and Ford all deviated from modern conservatism in major ways. All three acquiesced to a substantial welfare state and high taxes, while Ford was a strong supporter of the Equal Rights Amendment and moderate to liberal on abortion policy. Then all of a sudden, there is a Republican president in Reagan who fully embraces modern conservatism as constructed by William F. Buckley and other mid-century thinkers. Even Reagan himself had deviated prior to 1980 on abortion, but by the time he won the Republican nomination, he embraced the complete modern conservative issue bundle that included low taxes, social conservatism, and aggressive foreign policy. While there are still factional fights within the GOP after 1980, no Republican president (or presidential nominee) has seriously challenged that bundle.

Emergent insurgents, and why the new RNC rules matter

If the Republican National Committee has its way, we will see a different – shortened and streamlined - nomination process in 2016. The GOP will shorten its primary season, move its convention up, and hold fewer debates. A fourth rule change caught a few eyes in the media recently. According to Forbes, the RNC has changed its rules about what it takes for candidates to be considered at the convention at all. In 2012, the rule was that candidate had to have a plurality of the delegates from five states in order to be put forth for nomination. Depending on whom you ask, Ron Paul either met this threshold or came close to it in 2012. In 2016, the minimum requirements for a candidate to be considered at the convention will be higher. Instead of a plurality of delegates from five states, a candidate will need to have a majority of the delegates from eight states in order to have his or her name put before the convention. <br/>

It’s certainly possible that alteration in the rules will have little or no immediate practical impact on the party’s 2016 nomination. But the impact of changes like these can go beyond a single nomination season. First, they send a signal about what elites in the party organization think about the challenges they face. Someone thought it was important enough to make the change. Access to the convention floor has symbolic importance, and symbols matter in politics, even if their effects are difficult to measure. Furthermore, rules have the capacity to shape political conflict.

Party leaders change the rules in order to mediate among the party’s factions. This might make plenty of sense if we think of factions as fixed and based on fairly simple ideological labels, or as candidate-centered and ephemeral. Instead party factions are malleable, yet rooted in durable political ideas.
Right now, the different factions within the Republican Party are somewhat unclear. For example, Ron Paul was tried to position himself the Tea Party “favorite” in 2012 (with mixed success) and also identifies as a libertarian. But libertarians and Tea Partiers aren’t the same, although there appears to be some overlap. Nor are Tea Partiers necessarily the same as the party’s evangelical “base.” And what does “establishment” mean anyway?

Factions and alliances can shift in a party, based on perceptions of shared goals, identities, and interests. Labels like “establishment” and “insurgent” may not have intrinsic, fixed meaning, but they carry symbolic importance and have the potential to mobilize people based on their own conceptions about their needs and identities.

Based on what’s happened during their years out of (presidential and Senate) power, the Republican Party appears to have the ingredients for a major divide into insurgent and establishment factions. This isn’t the same thing as moderates vs. conservatives or social vs. economic conservatives. The moderate-conservative division probably accurately captures what was happening in 1952 or 1980, but the politics of the twenty-first century suggest something new. The elections of 2006 and 2008 were devastating for moderate and liberal New England Republican types, many of whom were replaced by Democrats. As the party has become more uniformly conservative, the infighting hasn’t necessarily declined.

Going back to the 2012 convention, it seems intuitively clear that the difference between Tea Party affiliated libertarian Ron Paul and Mitt Romney is one of kind, not degree. More broadly, political scientists also find evidence for a difference between Tea Party supporters and non-Tea Party conservatives. These differences have historical and philosophical roots. There are three basic tenets of conservatism: a preference for limited government; the idea that civilized societies have “moral orders” or social hierarchies; and a more process-based preference for slower change and elite-driven leadership. Modern conservative parties have mixed and matched these in different ways, leading to a variety of interpretations about what it means to be conservative. The moral order and limited government tenets inform the goals and views of the Tea Party/insurgent faction. But you won’t find much about incremental change and elite-driven governance.

The result is that there are qualitatively different ideas within the Republican Party that reflect not just different views on taxes or abortion, but on how America politics should work. The conflict over Ron Paul delegates at the 2012 convention illustrates how the insurgent faction operates under a different set of informal rules. Rick Santorum and Mike Huckabee (in 2008) were both arguably more conservative than their party’s eventually nominees (certainly, socially). But both bowed to party pressure to drop out of the race and step back at the convention, (eventually) throwing support behind the party ticket. The insurgent faction does not appear to be bound by these rules. The choice to defy party leaders wasn’t incidental; it reflects the group’s ideas about politics. These ideas date back to ideological traditions that were historically part of the Democratic Party – populism, suspicion of elites and institutions, and a desire to be left alone. In contrast, establishment Republicans draw their core ideology from traditions within their own party – neo-liberal economic policies, a strong pro-business orientation, and, historically, an orientation towards elites.

Insurgent movements in American parties are nothing new. Theodore Roosevelt’s campaign to wrestle the nomination away from William Howard Taft in 1912 used the newly instituted primary contests to challenge the power of party elites. In the late 1960s and early 1970s, liberal insurgents employed similar tactics, contesting the power of elites to control the party’s nomination processes and its policy positions. In both movements, the seeds of contemporary insurgent politics are apparent. Both staked their claims on the division between party elites and the will of the people.

But there’s an important difference. Although there were certainly policy motivations, the TR/Bull Moose movement was ultimately a personalistic one, unable to outlast Roosevelt’s candidacy. The liberal insurgency was about policy and ideology - specifically, the Democratic Party establishment’s support for the Vietnam War. Insurgency was a means to another kind of political end. Among Tea Party conservatives, insurgency itself appears to be a core value. In other words, this faction within the party seeks to challenge established power (as seen in efforts to “primary” Senate Minority Leader Mitch McConnell), and places a high value on ideological purity rather than compromise. A higher bar for inclusion in the nomination process could rally others within the party to their cause.

For adherents to this brand of conservative politics, the new rules send a signal that the party leadership isn’t very interested in their voices or concerns. It also has the potential to crystallize the insurgent-establishment division within the party, with implications for other dimensions of conflict – social issues, foreign policy, economics. While the RNC was likely trying to avoid major, visible conflict within the party in 2016, the rule may have the opposite effect. Both anti-establishment insurgents and establishment types have resonant claims about what it means to be conservative. A rule designed to shut one faction out could become a rallying point instead of a restriction.

What might the core issue positions of a such a faction look like? That's a post for another day...

Friday, April 18, 2014

Will Jeb Bush Get Stopped at the Border?

A lot of the punditry commenting on Jeb Bush’s flirtation with a presidential run has focused on his views on immigration.  But where do Republicans actually stand on immigration?  Nate Silver argues that Republicans actually have fairly centrist views on immigration.  The U.S. Chamber of Commerce, traditionally a stalwart of the GOP, has pushed hard for comprehensive immigration legislation  Large donors, who seem eager for Jeb to enter the race, generally back immigration reform.  Of all the leading Republican presidential possibilities, only Ted Cruz wholeheartedly embraces a hard line on immigration – and I doubt strongly that he will be the nominee.  Being “soft-liners” on immigration didn’t keep George W. Bush or John McCain from winning their party’s presidential nomination.

On the other hand, House Republicans have seemed pretty opposed to a “path to citizenship.”  (Or would they rather just not vote on the issue?)  The 2012 Republican platform denounces “amnesty,” calls for stricter workplace enforcement, and does not mention any form of a DREAM Act.  Marco Rubio’s presidential prospects took a beating last year when he took the lead on immigration legislation, similar to what happened to John McCain in 2007.  Even if Republicans overall hold moderate views on immigration, perhaps those who hold more restrictionist views care about the issue more strongly.  (There’s enough variation in polling on immigration to make one suspect that many citizens don’t have real views about it).   Polls show that Republicans – especially conservative Republicans – have more hardline opinions relative to other voters.  To put it another way, while most Republicans may not be immigration “hawks,” most immigration “hawks” are Republicans. 

But what constitutes an immigration “hawk?”  There we run into a problem.  If I want to know who is “pro-life,” I can ask the National Right to Life Committee.  If I want to know who is “pro-gun,” I can ask the National Rifle Association.  But there’s no large, respected organization that backs a more restrictionist approach to immigration.   The two best known groups both have extremist baggage, leading most politicians to avoid being associated with them.  The nation’s most vocal immigration “hawks” have been radio talk show hosts, who have a professional interest in being as provocative and strident as possible.  There are no set standards for deciding who has acceptable views on immigration.  

The immigration issue could also hurt Jeb Bush if it keeps him from winning support from key constituencies.  He’s unlikely to win support from Tea Party activists, but he does have a warm relationship with social conservatives.  Many evangelical leaders back immigration reform, but it’s not clear that their followers share those views.

So I have trouble reading the politics of the GOP on immigration.  I’m guessing that, while it is not irrelevant, it is not a litmus-test issue like abortion or gun control.  But I just don’t know. 

The Affordable Care Act's Near-Death Experiences Revisited

With President Obama announcing yesterday that enrollments through the Affordable Care Act have now topped 8 million (well over the CBO's projection of 7 million) and that the "overall age mix... is virtually the same as the age mix was in Massachusetts," it is a nice time to re-up my post on "The Affordable Care Act's Remarkable 4 Near-Death Experiences" that I did for my personal blog. In it, I speculate about the connection between this law's near disasters and recoveries, oppositional media coverage of that, and polarized mass opinions and beliefs about the law. Here is my concluding paragraph:
Finally, the way opposition media responded to these near death experiences is important. In essentially every circumstance where the progress of the ACA was under serious threat, Republican political elites and major conservative media didn't just alert their supporters that this was a major possible turning point. Rather, at basically every point I highlighted, they declared the ACA definitively dead. It's amazing how many premature touchdown dances the opponents of a single law have performed. The conservative media's reaction to the ACA's near death experiences helps us better understand public opinion on health reform. Political scientists know that one of the strongest determinants of the public's preferences is the rhetoric of politicians and elite media figures from one's own party. When a law has been over-and-over declared dead by elites you trust, yet it still somehow persists, it is less surprising that you think something illegitimate is going on, it was enacted by devious underhanded tactics, or that it still is a disaster and will never be implemented. The intersection of the ACA's near death experiences and the conservative media's coverage of them has surely enhanced polarization and conspiracy theorizing on this issue among the mass public.
To reiterate, I think neither the roller-coaster history of the law to this point, nor polarized opposition media are solely to blame for polarized opinions and beliefs about the law. Rather these two forces interact to create our remarkably persistent mass-level polarization on this topic.