Why is US Politics So Polarized?

[lpetrich]

Better Choices by Robbie Robinette, who describes his qualifications:

I studied physics at U.T. Austin but left before graduating to work at IBM and then a series of startups, eventually founding RGM Advisors with my partners Mark Melton and Richard Gorelick in my living room in 2001.

My experience with chaos theory, machine learning, agent modeling, and artificial intelligence leads me to take a different approach to the political problems we face today. I have built this website in an attempt to make some relatively complex concepts clear and understandable to a broader audience.

Sort of like mine - lots of experience with programming and data analysis. I like doing that. I remember someone who once felt offended by my doing that a lot. I tried to explain to him that that's what baseball fans like to do - they like doing statistics on baseball players' and baseball teams' performance.

Now to RR's work. He starts off with US House of Representatives political polarization, using DW-NOMINATE ideology scores from voteview.com

Showing numbers starting in 1939, the House was polarized back then, but there was plenty of overlap between the two parties. But starting in the 1970's, the overlap decreased until it pretty much vanished in the 1990's, and after that, the two parties moved apart, leaving a gap between them about half the size of each one's spread.

That does not represent the voters very well, because they have a center-peaked distribution of ideology scores. But even there, the distribution has been flattening over the last few decades.

 

[lpetrich]

Looking in House districts, candidates are sometimes rather far off from their voters, at least as estimated by RR from the 2012 and 2016 Presidential elections. He could have used more data, like from earlier Presidential elections, but he chose not to use Congressional ones, from their lower turnout, and thus their poorer representation of the voters' preferences.

RR then did simulated elections, and he showed that the parties tend to become polarized. In partisan primaries, Democrat voters pull their candidates to the left and Republican voters to the right. So in the general election, centrist voters have to choose between candidates on one side or the other of them.

Then he proposes ranked-choice ballots and head-to-head contests - use the Condorcet (head-to-head) winner if one is present, otherwise use some disambiguation method like the Schulze beatpath method.


Then some simulations.

The actual House members vary widely in how representative they are.

But in the simulations, with partisan primaries, they are roughly the same. That's whether using FPTP (single choice, most votes in one round), instant runoff (ranked-choice ballot, drop lowest top-ranked candidate in each round), and head-to-head (Condorcet).

But with blanket (nonpartisan, jungle) primaries, the representation was better, though not as good as it could be, with instant runoff. But with head-to-head, it gets very good.

I think that I'd have to look at the simulations more closely, because the simulated candidates seem like several candidates with equal strength, and that can cause confusion.

It also doesn't match real-world experience with ranked-choice elections. In the San Francisco Bay, in Maine, in Minneapolis, IRV, and in Burlington VT, IRV vote counting almost always picks the head-to-head winner as its winner, and winners often have majorities. Sets of several nearly equal candidates are not very common.

 

[Harry Bosch]

Better Choices by Robbie Robinette, who describes his qualifications:

I studied physics at U.T. Austin but left before graduating to work at IBM and then a series of startups, eventually founding RGM Advisors with my partners Mark Melton and Richard Gorelick in my living room in 2001.

My experience with chaos theory, machine learning, agent modeling, and artificial intelligence leads me to take a different approach to the political problems we face today. I have built this website in an attempt to make some relatively complex concepts clear and understandable to a broader audience.

Sort of like mine - lots of experience with programming and data analysis. I like doing that. I remember someone who once felt offended by my doing that a lot. I tried to explain to him that that's what baseball fans like to do - they like doing statistics on baseball players' and baseball teams' performance.

Now to RR's work. He starts off with US House of Representatives political polarization, using DW-NOMINATE ideology scores from voteview.com

Showing numbers starting in 1939, the House was polarized back then, but there was plenty of overlap between the two parties. But starting in the 1970's, the overlap decreased until it pretty much vanished in the 1990's, and after that, the two parties moved apart, leaving a gap between them about half the size of each one's spread.

That does not represent the voters very well, because they have a center-peaked distribution of ideology scores. But even there, the distribution has been flattening over the last few decades.

Well, I'm sure that there are many reasons why it's so polarized today. A big issue obviously is trump. We've never had a leader of a majority party who cared so little for the other side and moderates. We've never had a president who cared so little about everyone except for his voters. We are clearly becoming the United States of republicans.

 

[lpetrich]

Nevertheless, this suggests an important reform: doing away with partisan primaries and having nonpartisan ones instead. Candidates may still associate themselves with parties, but it should be easier for district-relative centrists to make progress.

Nonpartisan primaries are roughly comparable to having multiple rounds in sports events, where the top performers in each round advance to the next round.

An example of where nonpartisan primaries can be helpful is in Nancy Pelosi's district, CA-12, encompassing most of San Francisco. CA and WA have recently adopted top-two elections, a nonpartisan primary followed by the top two competing in the general election. This is widely used elsewhere in the world, where it is called a two-ballot or two-round system.

Most of the time in CA and WA, the two that advance are a Democrat and a Republican. But in CA-12 this year, it's two Democrats who are advancing: NP herself and a challenger on her left, Shahid Buttar. This is SB's second run; he had challenged her in 2018, placing third. This makes the race much more interesting than with yet another Republican.


There are lots of ways in which this work can be extended, like giving candidates different strengths that fit the statistics of real-world primary challengers. Also considering multiseat districts.

 

[Politesse]

An example of where nonpartisan primaries can be helpful is in Nancy Pelosi's district, CA-12, encompassing most of San Francisco. CA and WA have recently adopted top-two elections, a nonpartisan primary followed by the top two competing in the general election. This is widely used elsewhere in the world, where it is called a two-ballot or two-round system.

Most of the time in CA and WA, the two that advance are a Democrat and a Republican. But in CA-12 this year, it's two Democrats who are advancing: NP herself and a challenger on her left, Shahid Buttar. This is SB's second run; he had challenged her in 2018, placing third. This makes the race much more interesting than with yet another Republican.

This has been occurring with increasing frequency since 2014. It's very strange to me, as a native Californian, to see the state turn as genuinely Blue as everyone else in the country always seemed to think it was, rather than the decisively purple state that I actually grew up in.

Buttar doesn't have a chance. Pelosi is as popular as a career politician can hope to be, and Buttar is fighting allegations that he has been sexually harassing the women on his staff.

 

[Tigers!]

Part of the problem (looking at Australia but assume applies in USA, UK etc.) is that many suitable candidates are repelled by the thought of having their every utterance, action, thought etc. scrutinised, attacked by the web warriors of all stripes. Their family's followed, criticised, ostracised etc.

Why would you subject yourselves to that level of infantile, juvenile and puerile attention? When the death, rapine threats begin it is time to wonder into what morass have we descended?

 

[prideandfall]

because the US is the only country, that i know of at least, which such a significant cultural pairing of secularism and religious fanaticism.
most countries seem to be largely secular where religion exists but isn't a major cultural or political factor, or are basically theocratic states where secular forces are negligible.

since religion causes politics to be based on fantasy thinking and delusional paranoia while secularism runs the gamut across all possible foibles of humanity both positive and negative (and in the US tends towards liberalism and humanism) there is a pretty inevitable split in politics as those opposing forces butt against each other.

 

[ronburgundy]

I think the Civil Rights movement and reactionary Southern Strategy has a lot to do with it. It's is in large part what led to such stark divisions between the parties. There used to be racists more evenly spread between both parties, with racism not being a central dividing factor. Racists and non-racists within each party formed alliances based on shared interests in other aspects of society. But then the GOP made a conscious effort to court the votes of white supremacists and thereby win national elections via winning the south where such sentiments were so strong and still overtly expressed. Simultaneously, the Dems bought into being the party of equality. Racist conservatives left the party in droves. In all of the 1800's only 14 Dems who held state or national offices switched to the GOP. From 1900 to 1960, only 17 switched. Then in the single decade of the 1960's a sea-change shift occurred and 33 switched. That is a 12 fold increase over the average per decade number in the prior 6 decades. Not only did the GOP become the party of white supremacy, but basically a surrogate of the Confederacy.

Then as the cities became more and more ethnically diverse and may whites "fled" to suburbs and more rural areas, you had a greater and greater association between the politics of race and clashes in other aspects of culture and interests tied to rural vs urban life.

And then as particular regions become known for their political leanings, you have people who begin taking that into account in their decisions on where to move and live, combined with the greater than ever % of people who don't remain in their native area. That leads to a kind of migratory sorting by party affiliation. While job availability and pressing practical concerns are the major determinants of such moving decisions, there is room for cultural and political values to play a role, and there is research that people are migrating to areas more similar to their ideology.

Proximity heavily determines relationships, so now fewer and fewer people have any close work or personal relationships with those affiliated with the other party.

 

[Politesse]

Desegregation ended up replacing one form of racial segregation with another.