Maybe Third Parties Really Did Decide The Election
One bit of conventional wisdom that floated in the ether during the presidential campaign, and which did have some validity, was that support for Third Party candidates was inflated early on in the race and would eventually wane as “polarization” kicked in — the binary choice between Hillary and Trump would become more acute, and voters would quit their flirtation with fantasy candidates. This turned out to be especially true for Gary Johnson, who was hovering around 10% in the polling averages through the summer, but will end up with around 3.3% of the popular vote — not bad for the Libertarian Party, but nowhere near what some had thought he might get. Jill Stein will end up with around 1% of the popular vote, which is a slight decrease relative to where she’d polled over the summer, but not as dramatic a decrease as Johnson.
It’s simply a numerical fact that if approximately 25% of Jill Stein voters in Michigan had opted for Hillary, Hillary would have won the state. That doesn’t mean Jill Stein is to blame for Hillary’s loss, or that the requisite of percentage of Stein voters would necessarily have gone for Hillary in the first place — it’s just a statement of (hypothetical) fact.
One can come up with similar counter-factual scenarios all over the place:
Again, none of this is to assign direct blame to anyone for anything, it’s more to emphasize that the race was always highly volatile — to a far greater degree than pundits popularly supposed.
At present count, 5.84% of all voters went for a Third Party candidate or “other,” i.e. write-ins or blank votes. In 2012 that same percentage was 1.85%, representing a 216% increase in the percentage of voters who chose one of these options. This dynamic alone should’ve tipped off the pundits that there was an element of volatility in the electorate that certainly was nowhere near comparable to 2012, and thus they should be extra careful about making any definitive pronouncements.
At present count, 7,958,649 people voted for an “other” option compared with 2,384,728 in 2012, representing an increase of 234% over four years. That’s huge. Despite popular wisdom, Third Parties did end up playing a major role in deciding the election, which came down to narrow margins in a handful of states.
That this was a strong possibility was evidenced all along by the data, but pundits, convinced that it was sufficient to sit around mindlessly refreshing the New York Times “Upshot” forecast eighteen times a day, ignored such data. I wrote in August:
My personal, rough “fix” on the pundit consensus — which I’ve been doing since Trump was the clear Republican nominee, after the South Carolina primary — is to tack on an additional 20% to Trump’s odds relative to the pundit consensus. So, if the pundit consensus holds that Trump has 10% odds to win the election, my personal adjustment would be to make his odds 30%. That doesn’t necessarily mean Trump is the favorite to win at any given time, it just means his odds are consistently higher than what is popularly assumed, at least according to this metric I’ve devised for myself.
The reasons for my devising of this metric are manifold. First and foremost, the state of the race is highly volatile. Though we should be careful to not dwell excessively on any individual poll, general trends and patterns do indicate that an inordinately great number of voters are undecided or stating their preference for a third party candidate. A poll released yesterday by PPP found that 21% of voters either said they were undecided, or preferred one of three third party candidates.
If between a fifth and a quarter of the electorate presently states that they do not intend to vote for either of the two major party nominees, to me that suggests the race is highly “volatile” and thus susceptible to potential unforeseen shifts. One impetus for such a shift could be the first debate, in late September. We recall that the first Romney-Obama debate did end up having a substantial impact on the polling averages, even though Romney still ended up losing the election. So that “shock” from last cycle could be seen as a precursor to potential “shocks” between now and November.
There’s more to say on the import of Third Parties in particular, but this should also hammer home a more overarching point — past events are not necessarily predictive of future events, especially when you have such a limited sample size as is the case with presidential elections. Jill Stein’s vote total diminished somewhat relative to her polling average, but not by that much. Gary Johnson’s support collapsed, per the same metric. Maybe this made a tangible difference in Michigan, Wisconsin, and Pennsylvania. Who knows?
I’ve made this point before, but there have only been 58 presidential elections in the entire history of the United States, 18 since the end of World War II, 6 since the internet came into wide use (I’m using 1996 as the marker of that) and arguably one or two since social media became a dominant feature of American life. Thinking that you can extrapolate from prior elections exactly what will happen in future ones with such a limited sample size, and with such massive shifts at play (technological, social, etc.) is just absurd. I’ve always felt this was the case with regard to the 2016 election, even back in the banal days of 2014 when everyone was throwing their hands up in the air and resigning themselves to the inevitable coronation of Hillary. No: nothing is inevitable in politics, especially as control of the media is increasingly being wrested from a tiny cadre of powerful elites.
I’m going to quote here a post from my personal website, dated November 13, 2014:
Now, yes, it’s sometimes useful to refer back to previous cycles for insight into how future cycles might shake out, but there’s also an undercurrent of fatalism to these objections. “Hillary has a 70% favorability rating right now,” these interlocutors might shriek, “And no candidate with such high favorability ratings has ever failed to win the party’s nomination for president!”
One aspect of why this reasoning is fallacious: it’s contingent on a dreadfully small sample size. The “modern era” of presidential primaries is regarded as beginning in 1972, when a host of reforms changed the way nominees were selected to include more democratic participation. Therefore, a grand total of 11 cycles comprise the sample size which these nay-sayers insist is supposed to be so predictive of future election outcomes. (And that includes years in which primaries were uncontested, such as the 2012 Democratic presidential primary or the 2004 Republican presidential primary.)
Further, there’s little reason to believe that the 1988 Democratic presidential primary, for instance, bears any useful lessons for how the 2016 Democratic presidential primary might unfold. These two cycles will have occurred in vastly different social/economic/political contexts, with vastly different sets of variables at work — ranging from profound changes in media, to campaign finance, to US foreign policy. I mean, the list of massive differences between American life in 1988 and 2016 is just going to be endless.
Of course Hillary did win the Democratic nomination — I never predicted otherwise — but I think the overall thrust of the post holds up rather well. Very few foresaw that an elderly gadfly socialist senator from Vermont would end up receiving 46% of pledged delegates in the Democratic Primary contest, but he did, and that in large part owed to technological shifts that upended past assumptions about how these things would work. We still have very, very little good data on how best to contextualize the effect of social media on presidential-level politics. It doesn’t matter much with 80-year-old voters, but it could be the single-greatest opinion-shaping factor for 20-year-old voters. That’s a huge area of uncertainty that will need to be sussed out going forward.
American politics in general is a lot more volatile and changeable than people often believe. I can recall being immersed in Occupy Wall Street, and thinking — “Holy ****, is this really happening?” It was really a heady few weeks for those of us in the epicenter of it all, and it was not previously fathomable. One reason it was not previously fathomable is because the potency of social media as an organizing tool had not been totally comprehended to that point, and maybe still isn’t.
Virtually our entire Media-Pundit Industrial Complex spent literally 3+ years talking about the inevitability of Hillary Clinton; that’s what led to her getting anointed by Democratic Party bosses, who didn’t feel it was necessary to have a fully competitive primary, and just wanted to get the nuisance out of the way as painlessly as possible. And then that’s why she spent half of her time during the general election at big-dollar fundraisers rather than actually making a case to the American people, because she felt it was already in the bag — all the Smarty Pants Pundits had decreed such, so why bother going to Toledo for a rally when you can hobnob with Harvey Weinstein in The Hamptons?
In order to accurately gauge what’s going on around you, you need more than just a smattering of parlor-game predictions and folk wisdom (which is the full extent of what so many pundits have on offer). You need a combination of numerical and experiential data. (By the way, I’m demanding that people use the term “experiential data” rather than “anecdotal evidence.” Of course first-hand reporting is going to be ultimately “anecdotal” — there is no way to assemble a statistically-representative sample for the purposes of in-person reportage.)
Pundits and their compadres rely on lazy assumptions and premises that have no basis in fact, but get continually reinforced by colleagues they deem smart and savvy, and then they all act bewildered when these end up being proven false. The assumptions are lazy in part because they don’t indicate any awareness of the huge cultural shifts afoot, which are ever-present, and which need to be incorporated into any holistic political analysis.