## Maps of candidates from coconsideration

G. Elliot Morris posted this embedding of the current Democratic presidential candidates in R^2 on Twitter:

where the edge weights (and thus the embeddings) derive from YouGov data, which for each pair of candidates (i,j) tell you which proportion of voters who report they’re considering candidate i also tell you they’re considering candidate j.

Of course, this matrix is non-symmetric, which makes me wonder exactly how he derived distances from it. I also think his picture looks a little weird; Sanders and Bloomberg are quite ideologically distinct, and their coconsiderers few in number, but they end up neighbors in his embedding.

Here was my thought about how one might try to produce an embedding using the matrix above. Model voter ideology as a standard Gaussian f in R^2 (I know, I know…) and suppose each candidate is a point y in R^2. You can model propensity to consider y as a standard Gaussian centered at y, so that the number of voters who are considering candidate y is proportional to the integral

$\int f(x) f(y-x) dx$

and the voters who are considering candidate z to

$\int f(x) f(y-x) f(z-x) dx$

So the proportions in Morris’s table can be estimated by the ratio of the second integral to the first, which, if I computed it right (be very unsure about the constants) is

$(2/3) \exp(-(1/12) |y-2z|^2$.

(The reason this is doable in closed form is that the product of Gaussian probability density functions is just exp(-Q) for some other quadratic form, and we know how to integrate those.) In other words, the candidate y most likely to be considered by voters considering z is one who’s just like z but half as extreme. I think this is probably an artifact of the Gaussian I’m using, which doesn’t, for instance, really capture a scenario where there are multiple distinct clusters of voters; it posits a kind of center where ideological density is highest. Anyway, you can still try to find 8 points in R^2 making the function above approximate Morris’s numbers as closely as possible. I didn’t do this in a smart optimization way, I just initialized with random numbers and let it walk around randomly to improve the error until it stopped improving. I ended up here:

which agrees with Morris that Gabbard is way out there, that among the non-Gabbard candidates, Steyer and Klobuchar are hanging out there as vertices of the convex hull, and that Warren is reasonably central. But I think this picture more appropriately separates Bloomberg from Sanders.

How would you turn the coconsideration numbers into an R^2 embedding?

## Political coordinates test

A popular political quiz on the internet purports to place you on a Cartesian plane with “left-right” on one axis and “libertarian-communitarian” on the other, by presenting you with 36 assertions you’re suppposed to agree or disagree with. One of them is

“There are too many wasteful government programs.”

Well, of course there are! For this not to be the case, the government would have to be uniquely unwasteful among all large institutions. The quiz does not ask whether you agree that

“There are too many wasteful private enterprises.”

I would like to agree with both, but the test only allows me to agree with the first while remaining silent above the second, which makes me seem more of a free-market purist than I really am. Which questions you choose to ask affects which answers you’re able to get.

## If you build it they will come and exploit it

There’s no way to build a program for people in need that can’t be taken advantage of by unscrupulous people who aren’t in need. I have a friend, an attorney, who used to work cases involving people who defrauded the foster care system, taking state money for the care of children who didn’t exist, or who weren’t really in their care. It’s maddening. She eventually quit that job, partially because it was so dispiriting to come in daily contact with people being awful. But what can you do? You can’t build a fence strong enough to keep out all fraud without making the administrative burden impossibly high for the many honest people doing the hard, humane work of raising kids who need parents. There’s some optimal level of vigilance that leads to some optimal level of fraud and that optimal level of fraud isn’t zero.

I thought of my friend when I read this story, about developer Dan Gilbert getting an “opportunity zone” tax break officially intended for spurring development in impoverished areas:

Gilbert’s relationship with the White House helped him win his desired tax break, an email obtained by ProPublica suggests. In February 2018, as the selection process was underway, a top Michigan economic development official asked her colleague to call Quicken’s executive vice president for government affairs about opportunity zones.

“They worked with the White House on it and want to be sure we are coordinated,” wrote the official, Christine Roeder, in an email with the subject line “Quicken.”

The exact role of the White House is not clear. But less than two weeks after the email was written, the Trump administration revised its list of census tracts that were eligible for the tax break. New to the list? One of the downtown Detroit tracts dominated by Gilbert that had not previously been included. And the area made the cut even though it did not meet the poverty requirements of the program. The Gilbert opportunity zone is one of a handful around the country that were included despite not meeting the eligibility criteria, according to an analysis by ProPublica.

Maybe there’s no way to design a program like this without billionaires with phalanxes of lawyers and friends in high places being able to sop up some of the money. Even before the “opportunity zones,” Jared Kushner was able to game a similar program by drawing a gerrymandered “low-income district” that snaked its way through Jersey City to include the site of his luxury skyscraper and also some poor neighborhoods miles away. But I have to believe the optimal enforcement level is higher and the optimal malfeasance level lower than what we have now.

Ward 49 and Ward 66 are two big voting precincts in Madison. Ward 65, where I vote, has 2.819 registered voters. Ward 49, in the campus area with tons of undergrad-heavy high rises, has 3,505.

In the 2018 governor’s race, Ward 65 went for Tony Evers 2190-179. Ward 49 also went big for Evers, though not as dominantly: he won there 1985-591.

Now look at the April 2019 Supreme Court election. Ward 65 went strongly for the more liberal candidate, Lisa Neubauer, voting for her by a 1631-103 margin. Ward 49 also liked Neubauer but the margin was 531-101. 25% more voters but about a third as many votes. Evers narrowly won his election. Neubauer narrowly lost hers. Young voters sitting out downballot elections is pretty important.

## A lot of affluent liberals

Binyamin Appelbaum wrote an article in the New York Times about my native county, Montgomery County, Maryland, and this is what he tweeted about it:

A lot of affluent liberals in Montgomery County, outside Washington, D.C., fiercely opposed a plan to build a little more affordable housing. “Affordable is not what people move here for,” one of them told me.

The plan in question was approved unanimously by the County Council, all nine of whom were Democrats, but, as Appelbaum reported, not everyone in progressive Montgomery County was happy about it. He quoted several residents making remarks that made them look like, well, uptight snobs:

Ellen Paul, 59, said in-law suites were bad enough: “It’s changing suburbia to allow two homes on each lot. You’ll have strangers walking by your house all the time now.”

“That’s where the backyard trailers are going to go,” said Dale Barnhard, one of the more than 1,500 people who signed a petition opposing the “dramatic” change in rules.

or worse:

One county resident, Katherine C. Gugulis, wrote a protest letter in The Washington Post that concluded, “Just because others flee crime-ridden and poverty-stricken areas doesn’t mean Montgomery County has to be turned into a slum to accommodate them.”

I was interested in these affluent liberals and wanted to learn about them. A few minutes of Googling later, here’s what I found out. Katherine Gangulis is a Republican appointed official. Ellen Paul, according to her public LinkedIn profile, is a former staff assistant to a Republican member of the House of Representatives, and her most recent listed activity was public relations for a Republican candidate for Montgomery County Board of Education in 2014. Dale Barnhard doesn’t have any political career, as far as I know, but he wrote a letter to the Washington Post last year complaining about their biased coverage of Donald Trump. Hessie Harris, who worries aloud in Appelbaum’s piece about “flophouses” and literally utters the words “There goes the neighborhood,” is listed by the FEC as contributing thousands of dollars a year to Americans for Legal Immigration; that’s a PAC which describes its mission as “our fight AGAINST the costly and deadly illegal immigration & illegal immigrant invasion of America.”

These people aren’t liberals!

I don’t doubt there are liberals in Montgomery County who oppose relaxation of zoning. (I grew up there, and I live in Madison, WI: liberal NIMBYism is not a foreign idea to me.) But why weren’t any of those people in Appelbaum’s article? Why run a piece featuring a bunch of conservatives protesting a decision by an all-Democratic county council and bill it as a portrait of progressivism failing to live up to its ideals?

Here’s my theory. I don’t think Appelbaum purposely gathered quotes from Montgomery County’s small but nonzero Republican population for his piece. I think he had a story already in mind, a story of rich liberals who profess a commitment to affordable housing but really have a lot of contempt for the kind of person who lives there, and who would certainly under no circumstances stand for such people residing in Potomac or the nice parts of Bethesda, you know, the Whitman part. Those people might say their opposition to density had to do with something other than snobbery. But their words would show how they truly felt about the poorly-to-mediocrely-heeled.

And he got the quotes he wanted, the quotes that supported this story. Good, salty quotes. But the people who said those things weren’t self-styled progressives. They were Republicans.

Maybe there’s a reason for that!

Tagged , ,

## Madison Feb 2019 primary: quick and dirty clustering

Madison had a primary election last night for mayor and for several seats on the City Council and School Board. Turnout was high, as it seems to always be in Dane County lately. The Dane County Clerk has all the results in handy csv form, so you can just download things and start having some fun! There were four major candidates for mayor, so each ward in the city can be mapped to a point in R^4 by the vote share it gave to each of those; except of course this is really R^3 because the vote shares sum to 1. It’s easier to see R^2 than R^3 so you can use PCA to project yourself down to a nice map of wards:

This works pretty well! The main axis of variation (horizontal here) is Soglin vote, which is higher on the left and lower on the right; this vector is negatively weighted on Rhodes-Conway and Shukla but doesn’t pay much attention to Cheeks. The vertical axis mostly ignores Shukla and represents Cheeks taking votes from Rhodes-Conway at the top, and losing votes to Rhodes-Conway at the bottom. You can see a nice cluster of Isthmus and Near West wards in the lower right; Rhodes-Conway did really well there. 57 and 48 are off by themselves in the upper right corner; those are student wards, distinguished in the vote count by Grumpy Old Incumbent Paul Soglin getting next to no votes. And I mean “next to no” in the literal sense; he got one vote in each of those wards!

You can also do some off-the-shelf k-means clustering of those vectors in R^4 and you get meaningful results there. Essentially arbitrarily I broke the wards into 5 clusters and got:

[28, 29, 30, 32, 39, 40, 41, 42, 44, 45, 51, 52, 53, 62, 63, 64, 65, 81, 82, 105]

(east side Isthmus and near West)
[3, 4, 5, 7, 9, 10, 11, 17, 18, 22, 23, 24, 26, 38, 75, 88, 89, 90, 94, 96, 106, 107, 110, 111]

(far east and far west)

[15, 43, 46, 47, 48, 49, 50, 55, 56, 57, 58, 59, 60, 61, 66, 68, 69]

(campus and south Park)

[2, 12, 13, 14, 16, 21, 31, 33, 34, 35, 36, 37, 67, 80, 83, 84, 85, 86, 87, 93, 108, 109]

(west side, Hill Farms, north side, east of Monona)

[1, 6, 8, 19, 20, 25, 70, 71, 72, 73, 74, 76, 77, 78, 79, 91, 92, 95, 97, 98, 99, 100, 101, 102, 103, 104]

(southwest Madison and south of Monona)

Now what would be interesting is to go back and compare this with the ward-by-ward results of the gubernatorial primary last August! But I have other stuff to do today. Here’s some code so I remember it; this stuff is all simple and I have made no attempt to make the analysis robust.

Update:  I did the comparison with the August primary; interestingly, I didn’t see very many strong relationships. Soglin-for-mayor wards were typically also Soglin-for-governor wards. Wards that were strong for Kelda Helen Roys were also strong for Raj Shukla and weak for Soglin, but there wasn’t a strong relationship between Roys vote and Rhodes-Conway vote. On the other hand, Rhodes-Conway’s good wards also tended to be good ones for… Mike McCabe??

Tagged , ,

## Cities, Villages, Towns, and Scott Walker

Speaker of the Wisconsin Assembly Robin Vos, still smarting from Scott Walker’s loss in his re-election bid, said “Evers win was due to Dane County and the City of Milwaukee.”  It’s typical GOP politics here to split off Madison and Milwaukee like this, as if liberalism in Wisconsin is a pair of dark blue inkstains on an otherwise conservative shirt.

Not so.  There are liberals all over your shirt, Mr. Vos.

You can find tons of interesting data about Wisconsin elections in Excel spreadsheets at the Wisconsin Elections Commission page.  This already gives you the ability to do some quick and dirty analysis of where Evers’ victory was won.  In Wisconsin, every municipality is either a City, a Village, or a Town, in roughly decreasing order of urbanization.  So it’s easy to separate out Wisconsin into three parts, the Cities, the Villages, and the Towns.  This is what you get:

CITIES:  Walker 542148 (40%), Evers 808145 (60%)

VILLAGES: Walker 257858 (55%), Evers 208596 (45%)

TOWNS:  Walker 495074 (62%), Evers 307566 (38%).

That’s a pretty clear story.  Evers won in the cities, Walker won by a bit in the villages and by a lot in the most rural segment of the state, the towns.

But wait — Madison and Milwaukee are cities!  Is that all we’re seeing in this data, a distinction between Madison and Milwaukee on the one hand and real Wisconsin, Republican Wisconsin, Robin Vos’s Wisconsin, on the other?  Nope.  Take out the cities of Milwaukee and Madison from the city total and Evers still gets 521265 votes to Walker’s 477447, drawing 52% of the vote to Walker’s 48%.  There are decent-sized cities all over the state, and Evers won almost all of them.  Evers won Green Bay, he won Sheboygan, he won Appleton, he won Wausau.  Evers won Chippewa Falls and Viroqua and Oshkosh and Neenah and Fort Atkinson and Rhinelander and Beloit.  He won all over the place, wherever Wisconsinites congregate in any fair number.

Craig Gilbert has a much deeper dive into this data in the Journal-Sentinel.  The shift away from Scott Walker wasn’t just in the biggest cities; it was pretty uniform over localities with population 10,000 or more.  Update: An even deeper dive by John Johnson at Marquette, which brings in data from presidential elections too.

There’s a general feeling that the urban-rural split is new, a manifestation of Trumpian anti-city feeling.  Let’s look back at the 2010 election between Scott Walker and Tom Barrett, an election Walker won by 7 points.  2010, when Donald Trump was just Jeff Probst in a tie.  But the urban-rural split is still there:

CITIES: Walker 505213 (46%), Barrett 603905 (54%)

VILLAGES:  Walker 207243 (59%), Barrett 143297 (41%)

TOWNS:  Walker 416485 (62%) , Barrett 257101 (38%)

Here’s the thing, though.  You can see that Walker actually didn’t do any worse in the towns in 2018 than he did in 2010.  But his support dropped off a lot in the villages and the cities.  And if you take Madison and Milwaukee out of the 2010 totals, Walker won the remainder of Wisconsin’s cities 53-47, which is actually a bit ahead of his overall 2010 statewide margin.

I don’t think Donald Trump has made Wisconsin politics very different.  I think it’s still a state that calls its own tune, and a state where either a Democrat or a Republican can win big — if they have something to say that makes sense all across the state, as Walker and Ron Johnson used to, as Evers and Tammy Baldwin do now.

## Wisconsin post-election post

A few thoughts.

• Crazy that, once again, a statewide election in Wisconsin is decided when a county clerk, late into the night, reveals a stash of not-yet-counted ballots.
• Evers winning and Baldwin cruising while national media darling Randy Bryce got soundly beaten by boring former UW Regent Bryan Steil in a not-that-Republican district is further evidence for the ham sandwich theory of Wisconsin politics.  Bryce ran about a point behind Tony Evers in Racine, his own home county!
• A lot of people asking “How can there be so many Baldwin-Walker voters?  It makes no sense!”  I think it makes sense.  In a state not experiencing any visible crisis, incumbency is an advantage.  The last Marquette poll had “state on the right track / state going the wrong direction” at 55/40.  Roughly speaking, if incumbency is a 5% boost and the state’s “mood” was +5% Democratic, you’d get a Democratic incumbent winning by 10 and a Republican incumbent locked in a tie, which is pretty much what happened.  There are plenty of people whose votes aren’t strongly ideological.
• With all the focus on the governor’s race, hardly anyone was watching the state legislative elections.  They didn’t go well for Democrats, who lost Caleb Frostman’s Senate seat and may lose seats in the Assembly too, while Democrats elsewhere were making pretty big gains in state legislatures.  Some of that, probably most of it, is our ridiculously gerrymandered Assembly map.  I’ll write more about that when I know more of the final numbers in these races.  But there’s another thing going on, too — I think Wisconsin Democrats have figured out that turnout efforts in Dane and Milwaukee are the most efficient way to turn effort into votes.  Turnout for this election was high statewide but in Dane County it was crazy; Tony Evers got more votes in the county than Hillary Clinton did in a presidential year!  And it’s not cause Dane County doesn’t love Hillary Clinton.  But that focus doesn’t shift the Assembly or Senate map.
• In 2014, the total votes for Attorney General were 95% of the governor votes.  This time it was 99%.  I read that as:  more voters voting party-line, fewer voting only for governor and leaving the AG line blank because all they know is the party.  But maybe I’m reading too much into it.

## Wisconsin pre-election post

I see no reason to doubt the polls that show a very close race between Tony Evers and Scott Walker for governor, but a healthy lead for incumbent Senator Tammy Baldwin over her challenger, Leah Vukmir.

In the current hyperpartisan environment, why are these two races so different?  For one thing, Baldwin and Walker are incumbents, and people in Wisconsin seem to mostly feel things are OK here (55% said the state was “on the right track” in the latest Marquette poll.)

But there’s something else.  I think in Wisconsin we like our politicians, well, not too salty.  Scott Walker is a bland guy.  He famously eats two ham sandwiches in a paper bag for lunch every day, and the thing is, I don’t think that’s an affectation, Walker really is a guy who doesn’t mind eating the same thing for lunch every day.  Tony Evers is bland, too, and he fought off seven spicier opponents in a wild primary, winning just about every county in the state.  When Tammy Baldwin announced for Senate people thought there was no way a movement liberal and out lesbian from Madison could win a statewide race.  She won it easily.  Because she is is a movement liberal out lesbian from Madison who in every way comes off as the super-nice mom at the PTA meeting who you go ahead and let make all the decisions because it kind of seems like she’s got this.

Leah Vukmir, by contrast, is a cookie-cutter Fox News Republican who wants to bring a meaner, harder-edged style to Wisconsin politics.  I don’t think it’s gonna work.  I think Wisconsinites, both Democrats and Republicans, prefer the ham sandwich in the paper bag.

## Unitarians, legislative districting, and fairness

I gave a talk about gerrymandering at the Prairie Unitarian Universalist society.  As usual, I showed how you can pretty easily district a state with a 60-40 partisan split to give the popular majority party 60% of the seats, 40% of the seats, or 100% of the seats.  After I do that, I ask the audience which map they consider the most fair and which they consider the least fair.  Usually, people rate the proportional representation map the fairest, and the map where the popular minority controls the legislature the least fair.

But not this time!  The Unitarians, almost to a one, thought the districting where the popular majority controlled all the seats was the least fair.  I take from this that the Unitarian ethos rates “the minority rules over the majority” as a lesser evil than “the minority is given no voice at all.”