Monday, September 26, 2016

Debate-O-Rama 2016

Today's the day! Tonight, Hillary Clinton and Donald Trump face off in probably the most anticipated presidential debate in political history. If you miss it, (a) are you living under a rock? and (b) there will be two more chances to watch those two clash in October (plus a vice-presidential debate).

But true fans of the political soap opera need more than just four fixes every four years. There are thousands of elections taking place this November, from the presidential on down to the state and local level, and most of the people running in them will also pontificate and bicker—on public-access TV, radio shows, League of Women Voters events—at debates of their own.

Unlike with the presidential debates, though, there is no master commission scheduling and regulating these downballot fora. With candidates agreeing to and canceling debate appearances on a case-by-case, TV-station-by-TV-station basis, it's incredibly hard to keep track of when non-presidential candidates are debating, even for prominent offices like U.S. Senate. What's a debate junkie to do?

My solution: crowdsource. I've started this Google spreadsheet with the aim of creating a full schedule of debates for senatorial, gubernatorial, and congressional campaigns nationwide; now all we have to do is compile the data. I've started things out by including all debate information I could readily find online, but I'm sure that plenty of debates are missing. That's where you come in. The document is fully editable by the public, so please, if you see a notice for an upcoming debate that isn't on the list, add it yourself. Then, when the time comes, sit back and marathon them all like Netflix is broken.

For your viewing pleasure, here is the full, ever-evolving list of 2016 political debates:

Tuesday, August 16, 2016

Curt Schilling's Political Career Is Doomed

In 2009, after the death of longtime Massachusetts Senator Ted Kennedy, Curt Schilling publicly toyed with the idea of running for the seat. Despite receiving plenty of encouragement from elites and social-media followers alike, the just-retired hurler declined to seek the Republican nomination, saying "it just did not make sense." Instead, the GOP turned to little-known State Senator Scott Brown to face Attorney General Martha Coakley. The rest is history.

In choosing not to run, Schilling probably made the worst decision of his still-non-existent political career. As Brown proved, conditions were perfect for a Republican upset in blue Massachusetts. The low-turnout environment of a special election allowed the most passionate, anti-Obama voters to have a disproportionate say in the outcome. The GOP was gifted a gaffe-prone opponent in Coakley. And the 2010 election took place before most of Schilling's controversial political statements. Like Barack Obama (to his benefit) and Chris Christie (to his detriment) before him, Schilling should have struck while the iron was hot.

Instead, a full eight years later, he has apparently suddenly decided he does want to run for office after all. After slyly hinting at a potential presidential run on Facebook last week, Schilling has now revealed that he is considering running for Senate in 2018 against Elizabeth Warren—the liberal hero who wrested Kennedy's seat away from Brown in 2012. "I would like to be one of the people responsible for getting Elizabeth Warren out of politics," Schilling said during an interview on WRKO. "She’s a nightmare. The left’s holding her up as the second coming of Hillary Clinton, Lord knows we don’t need the first."

But here's the reality: Schilling, a Republican, would stand no chance in such a campaign. Massachusetts remains a deep blue state; in top-of-the-ticket races since 2008, Democrats have beaten Republicans by an average of 55.6% to 41.7%. While Republicans like Brown and Governor Charlie Baker have beaten the odds and won statewide in recent years, Brown did so in a special election with just 48% turnout, while Baker did it by overperforming in Yankee Republican strongholds like Wellesley and Newton—moderate, affluent towns unlikely to respond to Schilling's brand of bombast.

And while Brown and Baker were running for open seats—and both against the polarizing Coakley—Schilling in 2018 would be facing an incumbent senator with $3.8 million in the bank and a 61/27% approval/disapproval rating. The last time Schilling was polled in Massachusetts, in 2009, he had a 29% favorability rating and a 39% unfavorability rating—again, before his recent controversial statements. To give a sense for where Schilling's popularity might stand today, in Rhode Island in 2013 his numbers were 9% favorable and 74% unfavorable. Of course, Rhode Islanders are probably better acquainted with Schilling's 38 Studios debacle—although, if Schilling does run in Massachusetts, Bay Staters are going to hear about it very quickly as well.

Massachusetts Republicans will probably enjoy favorable tailwinds in 2018, with an unpopular Hillary Clinton likely to be president and a popular Baker likely to be coasting to re-election. But Warren is a perfect fit for liberal Massachusetts, and the Trumpian Schilling decidedly is not. If he runs, Schilling will have no better luck getting elected to the U.S. Senate as he has had getting elected to the Baseball Hall of Fame. The deserving Schilling—who remains a phenomenal baseball player even as he is a terrible politician—should focus on getting elected in Cooperstown instead.

Friday, August 5, 2016

Is This Maine Ballot Measure the Solution to Political Polarization?

Imagine choosing our senators, governors, and congressmen the same way they choose who wins Best Picture at the Oscars. Maine thinks it’s worth a shot.

This November, a ballot measure known as Question 5 will ask Maine voters to totally overhaul the way the state decides its elections. If Question 5 passes, Maine would no longer follow the rest of the country in determining a winner by simple plurality; instead, it would be the first state to employ instant-runoff voting to ensure every winner receives a full majority. This alternative method of vote-counting is a favorite of election reformers nationwide, but they’ve never had a prize this big within their grasp. If it delivers as supporters promise, the initiative could swing the outcomes of elections—and even end the trend of political polarization.

Currently, Maine uses the “first-past-the-post” method of vote-counting that we are all familiar with: all a candidate must do to win the election is receive more votes than anybody else. However, in a multi-candidate field, this means that our next leaders can be elected with well short of a majority. (Think Donald Trump winning the Republican primary with 45% of the national popular vote.) As a result, a small niche of voters can carry a fringe candidate to victory even if a majority of the electorate objects—but splits their votes too diffusely to stop him or her.

Supporters of Maine's Question 5 intend to keep that from ever happening. Their solution is instant-runoff voting, which asks voters not simply to check just one name, but to rank the candidates in order of whom they prefer. Ballot-counters total up all the candidates’ first-place votes to see if any candidate has a majority. If not, the candidate with the least first-place votes is eliminated, and his or her votes are redistributed to the candidates ranked second on his or her ballots. This process continues until one candidate receives a majority. Advocates of instant-runoff voting (also known as ranked-choice voting) argue that this method precludes a scenario where two ideologically similar candidates split a majority of the vote and allow a third candidate, who is reviled by most voters, to win by plurality.

In head-to-head elections like Democrat-versus-Republican generals, instant-runoff voting works the same as our current winner-take-all system. Its strength is in handling multi-candidate mêlées, making it a favorite of minor parties and independents. The ballot proposal would allow voters to cast their “first” votes for third-party candidates without fear that they are throwing their vote away or playing spoiler—because they know their vote will safely be redistributed to their second-ranked choice. For example, an environmental activist could safely vote for the Green Party candidate without harming the chances of the Democrat; as long as the Democrat was ranked ahead of the Republican on his or her ballot, his or her vote would eventually count toward the Democrat after the Green candidate is eliminated.

This method is already used to elect legislators in Australia, presidents in India and Ireland, and mayors in San Francisco and in Maine’s largest city, Portland. The Maine legislature has considered and rejected instant-runoff voting in the past, but the ballot box finally gives it a realistic shot at passing. In 2006, a similar referendum in Minneapolis passed with 65% of the vote, and today’s municipal elections there employ instant-runoff voting. The system worked to the extreme, both for good and for bad, in the 2013 mayoral election, when it successfully winnowed down a 35-candidate field to declare City Councilor Betsey Hodges the winner—after 48 hours and 33 rounds of tabulation.

It's no coincidence that this alternative method is now catching on in Maine of all places. The state has a well-known independent streak. Its longtime Republican senators, Olympia Snowe and Susan Collins, were liberal mavericks to the rest of their party, and when Snowe retired in 2012, her replacement was an actual independent, Angus King. Overall, independent candidates are more numerous and are taken more seriously in Maine than in most other states.

As a result, election by plurality is a sore subject for many in the Pine Tree State. In Maine’s past 11 elections for governor, nine of the winners failed to win a majority of the vote thanks to the presence of independent candidates. In 2010 and 2014, many observers believe that centrist independent candidate Eliot Cutler drew liberal votes away from the Democratic candidate, leading twice to the election of Tea Party Republican Governor Paul LePage in a state that gave Barack Obama 56% of the vote in 2012.

In the 2010 election, LePage won the governorship with a narrow plurality of the vote (38.1%), defeating Cutler (36.4%) and Democrat Libby Mitchell (19.1%). Under instant-runoff voting, however, Cutler likely would have won. Mitchell would have been eliminated in an instant runoff, and the vast majority of her votes would almost certainly have gone to Cutler, her most ideologically similar candidate. Under a moderate Governor Cutler, the last four years in Maine would have looked very different.

This is a major selling point of the ballot measure to many Mainers—and a major sticking point for the state’s Democratic and Republican elite, who fear instant-runoff voting’s friendliness to independents will dilute their party’s clout. But historically this has not been the case. The system may benefit minor parties by helping them reach thresholds—in Maine, 5%—to gain ballot access and recognition as political parties, but it rarely hands them the election. The mayors of Minneapolis, San Francisco, and Portland are all Democrats, and fringe candidates have still been the first to be eliminated in instant runoffs. (Contrary to Cutler’s example, most independent candidates make their runs from either the far left or the far right.)

In fact, that may be instant-runoff supporters’ best argument: it stops polarizing candidates dead in their tracks. Because the winner has to be acceptable to over 50% of the electorate, instant-runoff voting rewards middle-of-the-road candidates who tack closest to the median voter. In a state where the legislature and governor are almost literally at war, and in a country where Congress and the White House’s inability to agree have ground governance to a halt, the potential of instant-runoff voting to stem political polarization is intriguing.

It has worked in Hollywood, of all places. Two years ago, American Sniper, the Iraq War film by conservative director Clint Eastwood, had many passionate supporters for Best Picture—quite possibly a plurality of Oscar voters in the fragmented eight-film field. But under the Academy’s instant-runoff method of voting, Birdman eventually won the award. Liberal Academy voters who didn’t rank American Sniper first probably listed it toward the bottom of their ballots, preventing it from picking up much support in the instant runoffs.

However, it’s easy to mischaracterize instant-runoff voting as always leading to the election of the consensus choice—the candidate acceptable to the broadest swath of voters. In reality, instant-runoff voting favors a combination of passion and consensus. Take a three-way election between a Democrat, a centrist independent, and a Republican. You would expect all Democratic voters to rank them first, second, and third, respectively, and all Republican voters to rank them third, second, and first, respectively. In that case, the independent candidate is every voter’s second choice—but he or she will be also the first eliminated in the instant runoffs.

This was essentially the scenario of Maine’s 2014 gubernatorial election, in which LePage was reelected with 48.2% of the vote. Democrat Mike Michaud earned 43.4%, and Cutler, despite having the best favorability ratings of the three candidates, pulled only 8.4%. According to exit polls, Cutler’s support would have roughly split evenly between LePage and Michaud in a two-way race. The outcome would not have changed: because more Republicans turned out to vote than Democrats, LePage took the election largely on his own strength.

This is the ugly truth to many Question 5 proponents who are supporting it for tactical reasons: instant-runoff voting does nothing more than reflect the electorate it is given. If one party is excited to vote and another stays home, no election-law change will make a difference. For Maine Democrats, Question 5 should not be considered a panacea, or a crutch to use to avoid the harder and more important work of organizing. In the end, Mainers should vote for Question 5 not for any perceived political side effects; history has shown they are unpredictable and in many cases nonexistent. Instead, Mainers should vote for Question 5 simply because they believe it is a fairer way to conduct elections.

Monday, July 18, 2016

How Many Fans Does Each MLB Team Have?: The 2016 Baseball Census

At 2.7 rWAR, Brandon Belt was not the most deserving National League Final Vote candidate. Michael Saunders was even last in rWAR among the five American League candidates. Yet Belt and Saunders were the ones elected to the All-Star Game last week—not because of their immense value, but because of their immense fan bases.

No one can possibly know exactly how many fans there are of the San Francisco Giants or the Toronto Blue Jays. The best we have are rankings of MLB teams by popularity, and even those disagree sometimes. To solve that problem, a few years ago, I developed a method of quantifying how many fans each MLB team has—a baseball census, if you will. This allows us to answer the question of not only which team has the largest fan base, but how large it is.

Whenever they poll a state, our friends over at Public Policy Polling (PPP) ask about more than just politics: barbecue, giant meteors, and, yes, baseball. Using PPP's state-by-state breakdowns of baseball fandom and multiplying by the actual Census Bureau's latest population estimates for each state, we get the U.S. population of each fan base:

Team Fans Team Fans
New York Yankees 23,789,450 Kansas City Royals 4,824,354
Atlanta Braves 20,307,037 Baltimore Orioles 4,602,667
Boston Red Sox 20,106,387 Colorado Rockies 4,530,586
Chicago Cubs 18,425,056 Minnesota Twins 4,517,874
San Francisco Giants 11,095,094 Cleveland Indians 4,331,531
Texas Rangers 10,275,941 Pittsburgh Pirates 4,169,171
St. Louis Cardinals 8,730,179 Arizona Diamondbacks 4,056,240
Los Angeles Dodgers 8,137,226 Milwaukee Brewers 3,970,562
Detroit Tigers 7,589,747 Oakland Athletics 3,738,013
New York Mets 7,237,208 San Diego Padres 3,329,816
Houston Astros 6,646,472 Chicago White Sox 3,035,525
Los Angeles Angels 6,437,770 Washington Nationals 2,804,064
Seattle Mariners 5,931,089 Tampa Bay Rays 2,765,617
Philadelphia Phillies 5,806,971 Miami Marlins 2,525,967
Cincinnati Reds 5,102,696 Toronto Blue Jays* 211,264

*Blue Jays fans are drastically undercounted because PPP does not poll in Canada.

Since last year, PPP has asked about baseball in three additional states, including big ones like New York and Maryland, giving this year's numbers their best accuracy yet. As a result, the New York Yankees have seized the title of America's favorite team, with an estimated 23,789,450 fans. Also moving up in my rankings, for the same reason, are the Mets (now at 7,237,208 fans) and Orioles (4,602,667). The Royals, meanwhile, have gained about a million fans since last year—pretty much all due to bandwagoning fans in updated polls from Missouri and Iowa. The Braves, Red Sox, and Cubs join the Yankees in baseball's Big Four; no other team comes within seven million fans of them.

Of course, the data are still incomplete. Our census/PPP's polls now cover 38 states, or 87.5% of the U.S. population—but that means 12 states (plus the District of Columbia) and one-eighth of Americans aren't accounted for. Here's a map of which states are still missing:

The Red Sox look like they could still be underrepresented; three New England states are missing. The Braves will also undoubtedly gain when Alabama and Tennessee are eventually added. But Yankees fans in New Jersey and Cubs fans in Indiana are also undercounted. While the raw numbers of my top four teams should be a little higher, we can be fairly confident that the increases will be proportional.

In the name of full disclosure, other factors may be throwing off the numbers as well. PPP's baseball questions are worded in an opt-in manner, so the 78% of Americans who tell them they have a favorite baseball team is much higher than the 42% of Americans who are baseball fans. Therefore, this may inflate fandom across the board. However, the nature of polling also limits PPP to asking about just eight or so teams per state. Not asking every state about teams that could conceivably have national fanbases, like the Dodgers or Tigers, could be undercounting them as well. The bottom line: with the data at our disposal, the above numbers are the most accurate enumerations we have.

So we think we know who MLB's most popular teams are. But what's the least popular team in baseball? Although the Blue Jays have only about 211,264 American fans, they clearly enjoy more popularity than that in Canada. Instead, it appears that MLB's two least popular teams both reside in Florida, a state that has always polled better for out-of-state teams like the Yankees and Braves than for the Marlins and Rays. Here in 2016, the Marlins appear to have the "advantage" in the race to the bottom, a switch from last year, when PPP's 2015 poll of Florida indicated that the state had slightly more Marlins fans than Rays fans. And unlike for the Nationals (DC is missing) and White Sox (Indiana), there is no state missing from our dataset that seems likely to give those two fanbases a boost. Like between the four at the top, it's a close race for last, but there can be no question it's between those two teams.

Sunday, June 26, 2016

When Will the Vice-Presidential Picks Be Announced?

If you take a leisurely stroll around the internet these days, it's impossible not to trip over some veepstakes speculation—even some going so far as to predict whom Hillary Clinton and Donald Trump will pick. But there's an equally important question that largely goes unaddressed—when will they be picked?

It can be useful to know when to expect these major announcements. Reporters need to mobilize at a moment's notice to cover the unveiling; rival campaigns need to react quickly with a well-placed counterattack and get started on opposition research. As for the rest of us? We just need to know when we should be monitoring Twitter and tracking airplanes.

As usual, if you want to predict future behavior, look at past behavior. To that end, I've compiled data on the announcement dates of all vice-presidential reveals since 1972. Here they are:

Historically, VP announcements have always happened between July 6 and August 29. Democrats' median announcement date is July 13; Republicans' is August 11. But that's a flawed way to look at it. The announcement dates have varied so wildly because the main event of the summer—each party's convention—is held at a different time each year. We learn a lot more by looking at announcement dates relative to the convention's timing.

Historically, the vice-presidential candidate (not to mention, going even farther back in time, the presidential candidate) has been chosen and announced at the convention itself. This was true in 1972, 1976, 1980, and for Republicans in 1988. Recently, though—starting with an innovation by Walter Mondale in 1984, when he picked Geraldine Ferraro four days before the convention in an historic stunt—parties have decided to spread out their big news events and announce the bottom of the ticket in advance. This trend makes it unlikely that we'll have to wait until either convention to know who the running mates will be—despite Trump's public flirtation with a primetime convention-week reveal.

However, the winner of the veepstakes is still frequently announced just before the convention. The vice-presidential pick was unveiled the week before the convention in 1984 by Democrats, in 1988 by Democrats (six days before the DNC started), in 1992 by Democrats (four days before), in 1996 by Republicans (three days before the RNC), in 2000 by Democrats (seven days before), in 2000 by Republicans (seven days before), in 2008 by Democrats (two days before), and in 2008 by Republicans (three days before). The median BC (Before Convention) announcement date is four days before for Democrats and three days before for Republicans. One week before each convention would thus seem like a pretty likely time to expect an announcement.

But is there a new trend developing in veep selection timing? Twice now, and both in recent history, the running mates have been announced well before the convention: in 2004 by Democrats (20 days before the DNC started) and in 2012 by Republicans (16 days before the RNC started). Of course, this is a small sample size; we probably can't know whether these are just exceptions or the next logical step in the progression of moving the VP selection farther and farther away from the conventions. But we can make a couple observations.

First, no running mate has ever been tapped more than three weeks before the start of a convention. That means it would be unprecedented for the Republican vice-presidential nominee to be announced this year before June 28 or the Democratic nominee to be announced before July 5. So I'd say you can safely stay off the grid until then without fear of missing anything.

Second (and, again, small-sample-size caveats apply), both of those campaigns ended up losing in November. In 2012, Mitt Romney's early selection of Paul Ryan was widely seen as an attempt to change the narrative of his campaign at a time when Romney wasn't exactly having his best week. This could suggest that Trump, whom everyone agrees is losing so far in 2016, could try the same tactic: seize control of the news cycle away from his poor fundraising and controversial statements by announcing his vice-presidential candidate early.

If I were a betting man, I might view this as the most likely of all the possible outcomes. If precedent pretty much mandates that Trump is not going to wait until the convention, it makes strategic sense for him to announce sooner rather than later. I'll make my prediction that Trump will announce his VP as soon as this coming week: on Friday, July 1, heading into the holiday weekend, so people can chew it over along with their hot dogs. (Since 2008, VPs have all been announced heading into a weekend, although over the long term, there is no discernible preference in the data for any specific day of the week.)

As for Clinton, the safe call would be that she will make her pick the week before the convention—except this year, the RNC and DNC occur on consecutive weeks, so that would be directly in the middle of the RNC. While it might be tempting to steal some of Trump's thunder, the Clinton campaign would never announce during the RNC for fear that her pick, and the message he or she sends, would be drowned out in the news cycle. So when can she fit the announcement in?

Again, history holds the answer. Only twice since 1972 were the conventions on consecutive weeks: in 2008 and in 2012. In 2012, Democrats (whose convention was later) had no vice-presidential candidate to announce; everyone knew Joe Biden would be renominated. But in 2008, the Republican convention was second, but John McCain announced Sarah Palin the week before (and therefore the same week as the DNC) anyway—he just waited until Friday, when the DNC was over. This is a politically shrewd move; immediately after your opponent's convention is over, swoop in and steal their media cycle, hopefully putting a swift end to any residual coverage and honeymoon period they were enjoying.

This year, the RNC ends on Thursday, July 21, and the DNC doesn't begin until Monday, July 25. I think there's a strong chance Clinton will pull a McCain and reveal her running mate the Friday in between: July 22. Now the only question remains—who will it be?

Tuesday, May 24, 2016

Trump's Lack of Havoc Downballot

Now that Donald Trump is, like we all predicted, the presumptive Republican nominee for president, we can close the book on his impact on the Republican primary. But while he did a great number of remarkable things over the first half of this campaign, wreaking havoc in downballot races wasn't one of them.

This year's primaries saw unprecedented levels of interest: Trump attracted an enthusiastic following who faithfully rushed to the polls for him, while those whom Trump terrified were equally determined to participate and snuff out his campaign. As a result, turnout has increased to near-record levels—nearly doubling up 2012 levels in some states. With so many new voters participating in Republican primaries, the potential was there for a major departure from past results that so often favored uneventful establishment victories. But in the handful of states where the Republican Party held its presidential and congressional/downballot primaries on the same day, there's little evidence that the anti-establishment wave that obviously took hold at the top of the ticket trickled down into other races.

The first question I had was whether all these new presidential-primary voters would even bother voting in downballot races. After all, so many of them seemed laser-focused on the presidency—either getting Trump elected to it or stopping him from ever getting close. But while turnout in downballot primaries certainly was lower than in the presidential primaries, it wasn't that different from the same such gap that always exists.

Between the Iowa caucus on February 1 and Trump's clinching of the nomination on May 3, nine states held downballot primaries concurrently with presidential ones. I looked at downballot Republican primaries that were contested in both 2012 and 2016 and compared turnout between the two years. The increase in Republican downballot turnout from 2012 to 2016 was typically just a little less than the increase in Republican presidential turnout. For instance, in Alabama, 139.5% of the number of 2012 Republican presidential voters showed up to vote for president on the GOP side in 2016, while an average of 132.9% of the number of 2012 voters in various Republican downballot primaries showed up to vote in those same Republican primaries again in 2016. Looking at the data another way, an average 75.6% of Republican presidential voters dropped down to vote in the selected downballot races in Alabama in 2016; an average 78.9% did so in 2012, a modest drop of 3.3 percentage points. In summary, yes, a few of the "new" 2016 Republican primary voters skipped downballot races—but most of them didn't.

What's more, there wasn't a lot of consistency from downballot race to downballot race. The table below goes beyond the aforementioned averages and gives you each race's turnout changes; this makes it abundantly clear that the presidential turnout was only a secondary influence on downballot turnout. It turns out that these races stand on their own merits; when they're competitive, turnout goes up.

For example, turnout in TX-15 shot up in 2016 (202.0% of 2012 levels) to the point where more people actually voted in the Republican primary for Congress than in the Republican presidential race in the district (+34.6% more, to be precise); this is because the district is an open-seat race in 2016. Likewise, in IL-11, a competitive GOP primary for a Democratic-held swing district caused turnout to almost double (194.0%) that of 2012, when incumbent Judy Biggert was the unquestioned nominee. This was such a significantly bigger increase than the presidential race that voters in IL-11 were probably motivated primarily by the race for Congress, not the Oval Office!

On the other side of the coin, Republican primary turnout in Indiana's US Senate race rose more modestly (+48.6%) from 2012 levels than did the presidential race (+74.4%), but this is more about reversion to the mean; the 2012 Senate primary in Indiana between Richard Lugar and Richard Mourdock attracted huge interest—even more than the presidential race that year (104.1% of the number of presidential voters cast a ballot for Senate). This year's contest, while an open seat, was more muted.

Here are the numbers for every race:

There are even exceptions in non-obvious places. For instance, 2016 turnout in all five contested GOP primaries for constitutional office in North Carolina surpassed 2012 levels by more than presidential turnout did. And, despite a snoozer of a Republican primary in which Senator Rob Portman faced only token opposition, this year's US Senate primary in Ohio saw more interest relative to the presidential race than in 2012. At times, the turnout differences can seem almost random. With such random variation, we can be pretty sure that there wasn't an unusual degree of downballot dropoff in this year's primaries—and we know unambiguously that voter participation in downballot primaries was higher than in 2012. With only one exception, turnout increased over 2012 levels by at least 18% in every race with enough data to analyze. (The exception is the Republican primary for MD-08, which has a simple explanation: the Democratic primary for Chris Van Hollen's old seat was extraordinarily competitive in 2016 and dominated the airwaves for weeks before the election, so unsurprisingly a lot of voters developed stronger opinions about that party's primary and pulled Democratic ballots instead.)

So Trumpites (and anti-Trumpites—whoever this new army of voters includes) are participating in lower-level elections. Are they having any effect? To find out, I took another look at the same elections analyzed above, except this time at the actual results. In those races with an incumbent running for re-election and/or a clear establishment-versus-insurgent tone, I found that the establishment candidate did better in 2016 than 2012 14 times and did worse 17 times. Overall, there is a very frail argument to be made that establishment candidates did fare worse this year, but their average drop—just 1.6 percentage points—is statistically negligible.

The numbers broken down by individual elections show yet again that the difference is dependent on what race you look at—which suggests that it's not the presidential campaign, but rather intrinsic specificities about each election, that drives results. Some incumbents faced very close calls this year, such as Kevin Brady's bare 53.4% majority in TX-08 or Bill Shuster's 1,227-vote win in PA-09 (not included in the chart because Shuster had no 2012 challenger). But the lack of competitive races surrounding them suggest that the culprit was those candidates' weaknesses rather than a pervasive anti-establishment wave. And some of the biggest swings from 2012 can be attributed to unique circumstances; for instance, TX-19 became an open seat this year, which is always going to lead to a more competitive primary.

Overall, even if their performance dipped a little from 2012, incumbent/establishment types still averaged a healthy 64.5% of the vote in the 31 elections—more than enough for a comfortable victory. So even if the influx of new voters made a difference at the margin, they rarely affected the final outcomes. For example, potentially potent challenges fizzled in both Alabama, where Senator Richard Shelby won 64.9%, and North Carolina, where Senator Richard Burr beat libertarian hero Greg Brannon 61.4% to 25.2%. In fact, no statewide or congressional Republican incumbent has yet lost their primary bid. Most have simply gone unopposed. The bottom line is that, for whatever reason, primary challenges just haven't gained an unusual amount of traction this year.

The one election you could point to where anti-establishment anger prevailed was in OH-08: John Boehner's old seat. In a highly symbolic triumph, businessman Warren Davidson won the primary here with the support of Boehner's old nemeses like the Club for Growth. But the ironic thing for our exercise is that Ohio is the one state where GOP establishment forces prevailed in the presidential primary! John Kasich won the state with 46.8% of the vote, so almost a full majority of the electorate was inclined toward his brand of moderate, reasonable Republicanism. By contrast, in states where Trump won (including Alabama, Mississippi, Illinois, North Carolina, and Indiana), the GOP establishment downballot escaped relatively unscathed.

But, interestingly, the state worst for the establishment overall was Texas, which was carried by Ted Cruz at the presidential level. Cruz isn't Trump—we don't think of him as infusing millions of new voters into the primary process—but he is certainly anti-establishment, having risen to power as a Tea Party hero. And we tend to think of his supporters as savvy, grassroots-oriented conservatives—the kind who organize at local conventions to gain every last delegate edge. Perhaps those supporters are also more likely than Trump's (who are, stereotypically at least, mainly just obsessed with him as an individual figure) to study downballot races and vote for Tea Party candidates there as well. It's an intriguing theory, but unfortunately it is impossible to prove.

In the end, the preponderance of evidence suggests that, while some of 2016's new Republican primary voters are undervoting in downballot races, most are participating. And while perhaps slightly more citizens are voting against establishment candidates, most are still toeing the party line. Overall, other than the greater turnout and interest in this year's elections across the board, downballot Republican primaries appear to be business as usual in this supposed "year of the outsider."

I have a few theories as to why. First, it may suggest that the myth of the type of voter Trump is turning out is overblown. The Donald has claimed to be turning out millions of disaffected people who haven't voted in decades as part of his effort to give a voice to whole swaths of America that have been left behind. But a POLITICO investigation found that Trump's voters aren't that new; rather, they're regular general-election voters who are just participating in primaries for the first time. That suggests a much milder shakeup than Trump himself would predict; after all, incumbents still get re-elected every November at rates around 90%.

Second, even if being part of Trump's new-voter army means you're anti-establishment, it does not necessarily mean you're especially conservative. Much has been made of the fact that Trump and his followers are not actually as far right as, say, Ted Cruz; instead, they are more likely to self-identify as moderate, and many were former Democrats. On the other hand, most downballot insurgent candidates remain conservative Tea Party types—not necessarily Trump-esque populists. Maybe Trump voters aren't shaking up those downballot races because they consider Tea Partiers a different breed. (This goes back to our theory of why insurgent candidates did better in states won by Tea Partier Cruz.) It could be that Donald Trumps just haven't shown up on the congressional level yet. But if this year is an indication, maybe that's what we should look out for in 2018—a new wave of populist primary challengers. If so, maybe that big anti-incumbent wave could still materialize after all.

Thursday, May 19, 2016

How Far is Baseball from Your State Capital?

I was bored last night and started a little game on Twitter: which MLB teams were closest and farthest from a state capital? Enough people nerded out with me over this that I figured I'd turn it into a blog post. Below are the distances from each MLB team to the nearest state capital, measured as the crow flies from each team's home ballpark to the closest state capitol building using this tool.

Following the trend of state capitals tending not to be states' biggest cities, only four MLB teams are based in state capitals: the Boston Red Sox, the Atlanta Braves, the Colorado Rockies, and the Arizona Diamondbacks. (The Toronto Blue Jays are also based in a provincial capital, while the Washington Nationals are obviously based in the nation's capital.) Of these, the Braves are based the closest to the seat of power, which is less than a mile up Hank Aaron Dr. (which becomes Capitol Ave. in Atlanta). When the Braves move to Cobb County, the Nats will, appropriately, be closest to a capitol building—the U.S. Capitol. If you just want to count state capitals, Coors Field will be closest.

The farthest MLB team from any state capital is the Marlins: 406 miles from Tallahassee. If you are just wondering about the farthest team from its own state government, though, the Padres take the cake—Petco is 472 miles from Sacramento.