Sunday, December 29, 2019

Setting the Record Straight on "Sanders Voters Elected Trump!"

As part of a 2017 exchange on Medium, I wrote a brief post that dealt with Clintonite claims that Bernie Sanders voters were responsible for Donald Trump's victory over Hillary Clinton in the 2016 presidential contest. I've recently learned (after seeing two people cite it in as many days on Twitter) that the post has traveled. It was just an informal Medium post, nothing fancy, and the subject is, like so many Clintonite claims, absurd, but I made at least one error in it, more data on the subject has become available since it was written, I have some time on my hands and one of the recurring themes of this blog as it has developed has been refuting the absurd claims of the Clinton cult and the larger Clintonite right. An upgrade is perhaps in order.

The Clintonite charge breaks down along three variants of increasing severity. The first is that Sanders supporters voted for Trump in large enough numbers that they put Trump over the top. A more expansive variant includes Sanders supporters who voted for 3rd-party candidates instead of Clinton. A still more expansive variant also ropes in those Sanders supporters who just sat out the general election. The ludicrousness of all of these to anyone with any understanding of the political system in the U.S. seems so obvious it shouldn't have to be explained. That it so often must be is a reflection of the political moment in which we live but this writer confesses he feels diminished by the exercise.

Here are some obvious points, explained pedantically:

1) At their most basic, politics involve building coalitions to accomplish particular goals. A politician who presents himself as a candidate for an office like President of the United States offers a program to the public and tries to build a sufficient enough coalition to be elected.

2) It is ultimately the responsibility of the politician to build that coalition.

3) When the politician fails to do so, one can examine and debate what factors went into this, and should--that is, after all, the nature of democracy. What one can't do is absolve the failed politician of any real responsibility, placing him in a privileged position by trying to shift the burden of his failure on to voters. Well, one can try but that's a juvenile and abjectly pointless exercise. Rooted in a fundamental misunderstanding of--or an infantile indifference toward--how these things work, the momentary emotional satisfaction that may accompany it is empty; it certainly isn't going to impress anyone who wasn't impressed by the politician in the first place. The only practical effect of that kind of finger-pointing is to breed resentments and thus to stand as an obstacle to building the next coalition. In a democracy, even a problematic liberal one like the U.S., the politician doesn't get to be the default. Politicians sometimes fail voters; voters don't fail politicians.

4) No politician is entitled to anyone's vote. If a politician wants someone's vote, it is incumbent upon that politician to earn that vote. If he fails to do so, that's on him.

That last point is alone sufficient to bring this subject to a close; if one doesn't proceed from the absurd and unsupportable assumption that Hillary Clinton was entitled to the votes of people who didn't wish to give them to her, there's nothing to discuss here.

Unfortunately, the narcissistic Clinton has nurtured a rather rabid cult following whose views reflect her own sense of entitlement.[1] It has transformed her insistence that she's never really responsible for anything into something akin to an article of faith. In the last few years, the cultists have come up with a lot of ways to mindlessly hate Bernie Sanders--the man who committed the unpardonable offense of challenging their Cult Queen's coronation--along with his supporters and progressives in general.

Their bad faith in this particular matter is transparent. It's there in, among other places, the data they, themselves, cite.

The factoid they most constantly reference in this respect is the same now as it was in 2017 when I wrote that earlier piece: 12% of Sanders voters who participated in the general election went on to vote for Trump. The figure comes from Brian Schaffner of the Cooperative Congressional Election Study (CCES). Writing in the Washington Post, John Sides, research director of the Democracy Fund Voter Study Group, discussed a fairly significant limitation of Schaffner's work:

"Schaffner examined only voters whose turnout in the primary and general election could be validated using voter file data. This excludes people who said they voted but actually did not--although it also excludes people who voted in caucuses or party-run primaries, for which validated turnout data are not as readily available."

This cripples the headline conclusion the Clinton cult tries to slap on it. Fourteen states held caucuses or party-run primaries. This methodology excludes them, leaving a significant chunk of the U.S. entirely unrepresented in the figures. Moreover, those excluded tended to be strong Sanders states; Sanders won all but 2 of them, often by large margins.

In the states CCES examined, 74.3% of Sanders voters went on to vote for Clinton. To get that figure, it's worth noting, CCES includes those who didn't participate in the general, so Clintonites who cite it to "prove" Sanders supporters cost Clinton the election are working from the most extreme version of Clinton entitlement mania, one that holds that Clinton was entitled to the votes of even those who didn't vote at all.

What was the actual Sanders-to-Trump crossover? There's data on that.

The RAND Corporation conducted a study in which they tracked the same nationwide sample-group, surveying them half a dozen times throughout the campaign process. It found that 6% of Sanders primary voters subsequently cast their general-election ballots for Trump.

This dovetails with an ABC News/Washington Post poll conducted only days before the election which found that only 8% of Sanders primary supporters planned to vote for Trump or had already done so, with an additional 8% distributed among third-party candidates:

Neither RAND nor ABC News verified these voters' participation in both the primary and general but the latter used a representative sample, one that didn't exclude a big portion of the U.S. and, warts and all, probably gives a better reading of where things stood.[2]

That great big blue stack in that graphic points toward the inherent absurdity of the Clintonites' claims. The outcome of all of these surveys may differ in details--and the differing figures are interesting and suggest one should exercise caution in the use of them--but all agree that Sanders supporters overwhelmingly went on to vote for Hillary Clinton.

Some quick back-of-the-envelope estimates can give an indication of how the election would have gone if this wasn't the case. As Clintonites are so fond of noting, Clinton drew nearly 3 million more votes than Trump. There's no complete count of primary voters/caucusgoers but by the partial list that is available, Sanders drew over 13 million votes.

CCES' partial data suggests 74.3% of Sanders voters went on to vote for Clinton, which would mean over 9.6 million Clinton votes--74.3% of 13 million--came from Sanders supporters. Without them, Trump would have a 6.6 million vote advantage over Clinton.

The ABC News/Washington Post poll reported that 82% of Sanders voters went on to vote for Clinton, which would mean over 10.6 million Clinton votes came from Sanders supporters. Without them, Trump is 7.6 million to the better.

This isn't something that should have to be said: If Sanders voters hadn't voted for Clinton, she would have lost. Badly. Not just the national popular vote either. To the extent that those percentages--74.6% and 82%--can be applied to the number of known Sanders voters in individual states, Clinton would have lost all the states she lost anyway but by larger margins, and would have also lost New Hampshire, New Mexico and Minnesota by even the more expansive figure, Virginia by the less expansive one, and probably some others my quick scan has missed. Trump would have won the electoral college in a huge landslide.

Of course, those percentages wouldn't necessarily apply to individual states. Something like them would though, and those quick, sloppy figures are an illustration of the effect this would have.

By contrast, Clinton cultists have, for years, chosen to portray the relatively small number of Sanders supporters who didn't vote for Clinton as representative of Sanders voters, of Sanders himself and of progressives in general. It's an utterly arbitrary--and, by the numbers, indefensible--enterprise.

It's also the case that Sanders was the energizing hope-and-change candidate in the 2016 race, who brought into the process many new people who had never voted or had given up on voting but who came back to support him. Among other things, Sanders drew an incredibly enthusiastic youth following, which gave him a significantly larger number of young voters than both Clinton and Trump combined. Most of those went on to vote for Clinton; her general-election vote-count is, because of Sanders' presence in the race, padded with x number of votes she probably wouldn't have gotten absent Sanders' presence. If there are any precise estimates on how many, I've never seen them but the available information makes clear that Sanders added to, not deducted from, Clinton.

Returning to the back of that envelope drives the stake further through the heart of the Clintonite narrative. Consider:

All told, 136,669,276 people voted in the 2016 presidential election. Looking at the exit polling, the Democratic contingent equaled 37% of that, or 50,567,632. Of that, the defection rate for Democrats was 9%...

...which comes to 4,551,086 self-identified Democratic voters who chose to cast their ballots for Trump, rather than Clinton. If one includes Dem defections to third-party candidates, the number rises to 5,562,439.5.

This exit polling doesn't really allow for a similar assessment of defections among Democratic-leaning independents but we have other data that touches that question. A much-remarked-upon phenomenon of 2016 is how many people who had previously cast their ballots for Barack Obama switched to Trump. Sabato's Crystal Ball, run by the University of Virginia Center for Politics, cites three surveys that examined this:

--The American National Election Study suggests that 8.4 million Obama voters backed Trump.

--The CCES suggests 6.7 million Obama-Trump voters.

--The University of Virginia Center for Politics suggests 9.2 million Obama-Trump voters.

Faced with so unappealing a candidate as Hillary Clinton, many Democrats simply stayed home in 2016. The drop in turnout among black voters, the most reliably Democratic voting constituency, was widely noted; black turnout saw its first decline in 20 years, falling to its lowest level since 2004. More broadly, 4.4 million people who voted for Obama in 2012 sat out 2016.

Moving out even further, 45% of the potential electorate--over 90 million people--declined to go to the polls. If "non-voter" had been a political candidate, he would have beaten everyone else and taken the presidency by an electoral college landslide:

The state-by-state numbers for the above chart:

The Pew Research Center studied the 2016 electorate. One of their findings about the non-voters:
"Among members of the panel who were categorized as nonvoters, 37% expressed a preference for Hillary Clinton, 30% for Donald Trump and 9% for Gary Johnson or Jill Stein; 14% preferred another candidate or declined to express a preference. Party affiliation among nonvoters skewed even more Democratic than did candidate preferences. Democrats and Democratic-leaning independents made up a 55% majority of nonvoters; about four-in-ten (41%) nonvoters were Republicans and Republican leaners."
Nothing about the Clinton candidacy inspired these people to show up.

All of these numbers come with qualifiers. A crossover rate of 6-10% for both Dems voting Repub and Repubs voting Dem is normal in a presidential race. The surveys regarding Obama-Trump voters show some statistical variance with the election results as a consequence of voters inaccurately remembering the candidate for whom they cast their ballot years earlier (a matter discussed on the Crystal Ball site). Turnout in American elections is typically terrible. It also stands to reason that there's plenty of crossover between these groups--self-identified Dems who voted for Trump with Obama voters who voted for Trump and so on. That isn't particularly important here though.

What matters is the one thing all of these groups have in common: all of them outnumber--vastly outnumber--the Sanders-"not Clinton" faction.

The partial data from CCES suggests 25.7% of Sanders supporters were "no" on Clinton, which, with 13 million Sanders voters, would come to 3.47 million. By the ABC News data, 18% didn't back Clinton, which is 2.34 million. There is, no doubt, plenty of overlap between Sanders-"not Clinton" voters and the groups identified here but all of these groups are much larger than the Sanders-"not Clinton" contingent and any one of them could just a credibly be said to be responsible for Clinton's loss. Singling out Sanders voters for blame is, in short, entirely arbitrary--a bit of demagoguery offered only because it serves the emotional needs of the Clintonite right. As the CCES data on Sanders voters includes the relatively small number of them who cast a ballot for Sanders in the primaries than sat out the general, Clintonites citing that data are, as noted earlier, making a claim that Clinton was entitled to the votes of even those who didn't vote. Planting a Clinton flag on that group while not doing so on the rest of these 90-million+ voters who didn't participate in the general lets them emotionally blame Sanders for the outcome but it is, like the rest of this, an entirely arbitrary exercise.[3]

Some Clintonites single out Michigan, Wisconsin and Pennsylvania--longstanding blue states that Trump flipped by narrow margins to clinch the election--and point out that Trump's margin of victory in those states is less than the number of Sanders voters who switched to Trump. A well-traveled graphic of this genre:

But the numbers in the graph make rubbish of the Clintonite conclusions; if the overwhelming majority of Sanders supporters hadn't voted for Clinton, she would have lost all three of them by hundreds of thousands more votes than she did. Sanders had defeated Clinton in two of the three (Wisconsin and Michigan). All of the dynamics I've outlined above--the Dem-to-Repub defections, the Obama-to-Trump voters, the drop in general Democratic turnout, the general low turnout--apply to these races as well. NBC News created an interactive map that shows the previously Obama-voting counties across the U.S. that flipped to Trump. In Wisconsin, there were 20 of them, all but one double-digit flips, most of them massive:

Clinton didn't flip any Wisconsin counties. She did manage to flip one county in Pennsylvania--Chester County, by 10.2%--but all of the other flips were Obama-to-Trump:

In Michigan, Clinton flipped no counties, but 11 Obama counties voted for Trump:

Michigan also saw 75,335 voters show up then only vote on downballot races, skipping the presidential contest. There were 7 times as many of these "undervoters" as Trump's popular-vote victory in the state.[see note #3 below] In all three states, Trump managed this by a combination of Obama-to-Trump voter defections and lower Dem turnout. Though turnout was slightly up in Michigan--0.1% over 2012--turnout in Democratic-leaning areas declined, often sharply, while going up in Trump-friendly rural counties. In Wisconsin, turnout was generally down--the lowest in 20 years--but again, this wasn't distributed evenly: it was lower in strongly Democratic areas. In Pennsylvania, turnout was slightly higher than in 2012 (by just under 3%) but "counties that vote Democratic largely turned out below their 2008 levels, while Republican-leaning counties generally saw higher turnout levels--some more so than either of the past two elections."

If all of us who, over the last few years, have taken part in public affairs discussions on the internet were given a dollar for every time we've seen some Clintonite say Bernie Sanders "isn't a Democrat," we'd be quite wealthy indeed. Wiser heads inside a party would encourage those outside it to join up, not trash them for doing so--that's how the coalition-building of politics works--but wisdom isn't a trait one readily associates with Clintonites. In 2016, Sanders was building a Democratic coalition that included many Republican-minded voters. "Perhaps the most important feature of Sanders-Trump voters is this," writes John Sides in that Washington Post article referenced earlier: "They weren’t really Democrats to begin with":
"Of course, we know that many Sanders voters did not readily identify with the Democratic Party as of 2016, and Schaffner found that Sanders-Trump voters were even less likely to identify as Democrats. Sanders-Trump voters didn’t much approve of Obama either.... [I]t doesn’t appear that many [Sanders-Trump voters] were predisposed to support Clinton in the first place."
Sides, Schaffner and other commentators have tried to posit attitudes toward race as the characteristic distinguishing Sanders-Trump voters from Sanders-Clinton voters, with the former holding retrograde opinions, but that's a dubious undertaking. Conservatives tend to have opinions on race that are some degree of backwards; that's well-established by polling. Sanders doesn't offer anything that would appeal to the racist sentiments of racists,[4] but he has long made a project of reaching out to Republicans with his populist message and most of the headline policies he placed at the center of his campaign were very broadly popular. If those policies attract conservatives, the conservatives' attitudes toward race come with them.[5] Clintonites who use these Sanders-Trump voters against Sanders' larger base of supporters and Sanders himself are, in effect, condemning Sanders for offering a program that attracted broad popular support--the very thing every broad-based political movement wants to do and must do to win.

I tacked a note on to the end of my earlier Medium article that dealt with a factoid often employed by Sanders supporters: that in 2008, 24% of Clinton primary voters went on to vote for John McCain over Barack Obama in the general election, a significantly larger percentage than Sanders supporters who voted for Trump. I erroneously attributed that to the Cooperative Congressional Election Study, which, given the dubious eye I'd cast toward CCES's methodology, suggested it was a dubious conclusion. It was actually the product of the 2008 Cooperative Campaign Analysis Project (CCAP), which, like the CCES polling, was conducted by YouGov. CCAP interviewed the same respondents multiple times throughout the 2008 campaign and found that 24% of Clinton supporters went on to support McCain. That, it seems, is a conclusion not as questionable as I'd suggested. I'd also noted what CNN had reported in 2008, that
"Exit polling also showed that Democrats who supported Sen. Hillary Clinton during the primaries overwhelming voted for Obama in the general election, 84 percent to 15 percent for McCain."
Since that earlier article, I've come across a paper by a trio of researchers under the aegis of the Associated Press-Yahoo News 2008 Election Study. It was published in the Fall 2010 issues of Public Opinion Quarterly and on this question, it concludes:
"Ultimately 25 percent of these Clinton primary voters cast a ballot for McCain in the general election. Just five percent of them supported neither of the major party candidates--either staying home or voting for a third-party candidate..."
Make of all of this what one will--as before, the variation in them suggests one should approach them with caution--but this writer isn't a big fan of the kind of whataboutism that generally accompanies Sanders supporters' use of such numbers. When Hillary Clinton's followers parrot her incessant crowing about how she worked so hard to elect Obama and convince her supporters to vote for him but how Sanders never did the same for her (though he did three times as many appearances on her behalf as she did for Obama), torch them for their hypocrisy by throwing these numbers in their face all you like but beyond perhaps providing some context for the Clintonites claims, the number of Clinton supporters who went McCain isn't really relevant to the effort to blame Sanders for Clinton's loss. Theirs is a stupid emotional attack; meet it with reason.



[1] And, of course, that sense of entitlement only goes one way. Jill Stein, the lefty Green party candidate, isn't, in the Clinton cult formulation, entitled to the votes she earned. Her 1,457,216 voters are merely treated as stolen Clinton votes. This entitlement mentality takes no account of the rest of the race. Gary Johnson, the former Republican governor of New Mexico, ran as the Libertarian party candidate and earned 4,489,341 votes. Evan McMullin, another Republican (though critical of Trump) drew 731,991. Constitution party candidate Darrell Castle managed 203,069. These are all right-wing parties and, combined, drew many times the vote of Stein. Are these just stolen Trump votes? If one applies to them the same logic as that of the Clinton cult, Trump would be entitled to their votes and with them would have not only won the presidency but also the popular vote.

[2] RAND uses a somewhat unique methodology, and I haven't gone into it here.

[3] Still another big group--thought not as big--is the "undervotes." In December 2016, the Washington Post undertook an effort to determine how many people went to the polls but only voted for downballot races, declining to cast a vote in the presidential contest. The Post was able to compile information from 33 states and Washington D.C.. Only a partial count but it found that 1.7 million people in those states fit that description (compared to only 754,000 people in those same states in 2012). It's likely that this is largely a consequence of having, as the party nominees, two spectacularly unpopular candidates--the two most unpopular major-party candidates in the history of polling. "In several states," wrote the Post's Philip Bump, "the number of people who didn't vote was near or greater than the eventual margin of victory":

"Notice that all of those states are ones that would count as 'swings,' writes Bump. Without a more detailed survey to provide estimates, there's no way to know whether these were Democratic or Republican voters but they could have swung the election. As Bump notes, "had one-seventh of those Michigan voters who decided not to vote in the presidential race cast a ballot for Hillary Clinton, she would have won the state."

[4] The same can't be said for Clinton's appalling 2008 campaign. Clinton didn't have any populist social-democratic policies to attract a broad base of support. Faced with a black opponent, she instead went hard on the race- and Muslim-baiting:
"Clinton’s claim that Obama's support among 'working, hard-working Americans, white Americans' was weakening and 'how whites in both states who had not completed college' were supporting her; her criticism of Obama over Reverend Jeremiah Wright; her insistence that Obama both 'denounce' and 'reject' Louis Farrakhan’s endorsement; and her doe-eyed claim that there was nothing to suggest Obama was a Muslim, 'as far as I know.'"
Ryan Cooper of The Week covers more of this, writing that as "Obama began to slowly pull ahead, the Clinton camp resorted to increasingly blatant race- and Muslim-baiting," including the infamous incident when the Clinton campaign circulated an image of Obama in traditional Somali garb--the beginning of a decade of right-wing efforts to Other-ize Obama. "A Clinton flack then went on MSNBC and argued that Obama should not be ashamed to appear in 'his native clothing, in the clothing of his country.'... The capstone came in May, when Hillary Clinton started openly boasting about her superior support from white voters." This worked, to an extent. Exit-polling showed Clinton getting a boost of support from those who said the race of the candidates was important to them.

[5] It's also likely that some of Sanders' conservative supporters were just casting anti-Clinton votes or were otherwise unserious about voting for Sanders, though that's probably not as widespread as some seem to suspect.

No comments:

Post a Comment