Don't Blame Howard Schultz... Blame “Unranked, Plurality, Winner-Takes-All” Voting!
If a majority of people prefer Maxwell House to Folgers, then can introducing Starbucks make Folgers the “Best Tasting” coffee?
OK, while I don’t know if anyone has a relative preference between Maxwell House and Folgers, I do know that Starbucks CEO Howard Schultz has been taking a lot of incoming fire for publicly considering running for President as an independent candidate. Democrats have been contending that a non-wackjob 3rd option, such as Schultz, could peel off enough of the anti-Trump vote in 2020 to swing the election in Trump’s favor.
To be fair, they are not necessarily incorrect, as even a fringe candidate such as Jill Stein was able to scrape together enough votes (with Russian help) in Michigan (1.07%) and Wisconsin (1.04%) to cover the margin of victory for Trump in those states. Award those votes to Hillary, along with the votes for Gary Johnson in Pennsylvania and Florida, and we wouldn’t be eagerly awaiting the latest Mueller news drop, since there would be no President Trump.
Let’s dissect the problem…
In math, if A > B, then introducing C should never cause B > A. And, despite behavioral economics showing that, with individual preferences, a choice C can often induce a person to flip relative preferences for A & B, if an individual's preference is “Never B” then introducing C should not cause the person to prefer B > A… (that is, unless the individual’s preference is also “Never A” — such as the feeling many voters felt in 2016)!
The matter gets more complicated when we aggregate individual preferences to create a “macro” preference. Kenneth Arrow’s “Impossibility Theorem” deals with this in a brilliant Nobel-Prize-winning way, but for simplicity, we can easily understand that, in our current state-by-state Presidential voting system, if there are exactly 3 candidates, a candidate could theoretically win a state with only 33.4% of the vote… even if that candidate is a distant 3rd choice for 66.6% of the voters. Throw in a few more candidates (e.g. Green Party, Libertarian, etc.) and you could have a state’s electoral votes going to a candidate whom over 2/3 of the voters would rank as “Never”. (Note, this is true in many primaries as well — for example, Donald Trump won all 50 of South Carolina’s delegates, despite only receiving 32.5% of the vote.)
And if we then consider that a candidate need only win 50%+1 available electors, then even if there are only 2 candidates running, one could theoretically win with only (roughly) 25% of the national popular vote, and even less if there is a viable 3rd candidate!
Is there a solution?
Why should a candidate whom a majority of people specifically want not to win be able to win?!
It is because of the current presidential election system with the following “features”: (1) unranked voting, (2) winner by plurality, (3) winner-takes-all electors. Let’s consider these one-by-one, along with the impact of adjusting these features:
1) "Unranked voting":
If among 2 choices, A & B, only 34% of people prefer B, and the other 66% prefer A, that’s an easy win for A. But now if C enters the race, and 50% of the people who prefer A now vote for C, then somehow, B has just won the election 34%-33%-33%! However, if we introduce a very simple ranked voting system, giving 3 pts for 1st choice, 2 pts for 2nd choice and 1 point for 3rd (if the voter chooses a 2nd & 3rd choice), then even if those voting for B don’t put a 2nd or 3rd choice, and all 66% of the others put a 1st & 2nd choice, then (using weighted percentages), A & B would each have 165% (33%x3 + 33%x2), while C would only have 102% (34%x3), putting C in 3rd place, rather than 1st.
With unranked voting, candidate B would have won the election 34%-33%-33%, because 34% is a “plurality” (i.e. highest %). However, if we required a “majority” (i.e. minimum 50% of unranked vote), then B would be forced into a runoff with A or C (whichever had the higher % - I know I have them equal in this example). And, since those who voted for A and C are both “Never B”-ers, B would lose the runoff 66%-34%. (Even with ranked voting, plurality could make a “last choice” for a majority end up the winner.)
Even if there are only 2 candidates, with the winner-takes-all electoral system, a candidate who wins a large majority of the population votes across the country can lose the overall election. (Note: Even though she “won" the popular vote in 2016, Hillary did not win a majority of the popular vote.) If electors are instead apportioned proportionally (or if we drop the electoral system completely), then this skewing of outcomes wouldn’t be as severe.
So, what about Howard Schultz?
Yes, if the sole goal is to not re-elect Trump, then, given our current system, with the above flaws, it seems that Schultz would only complicate that effort. But that’s forgetting that the pendulum can swing both ways. What happens if the Democrats nominate a candidate, A, who actually would lose a head-to-head election with B (i.e Trump), unless a 3rd candidate like Schultz peels off enough “Never A or B” votes to swing the electoral college to A? What if Schultz is actually the Democrats’ savior?!
Either way, there is something unsatisfying about hoping to win an election by being the "less-undesirable" alternative. Democrats who are currently attacking Howard Schultz sound like they’re hedging against the possibility they may nominate a “Never A” candidate. They should instead publicly embrace Schultz… say he’s welcome to run, but express confidence they will nominate a candidate who could easily win a contested 3-way race with Schultz and Trump, even with the current electoral system.
Howard Schultz is a smart businessman. He’s using game theory here. He knows that, given the current “Unranked, Plurality, Winner-Takes-All” Voting system, he can only have a viable shot to win if he’s the Option C to a “Never A nor B” election. So, let him explore a run as an independent during the primaries… then have him drop his plans once he sees that the Democrats have nominated a candidate A for whom not only A > B and A > C, but for whom A > (B + C).
While I said I don’t know if anyone prefers Maxwell House over Folgers or vice-versa… I DO know that “Peets' Coffee" is a vastly better cup of coffee than either. And if people are given a choice between Peets and Folgers, then introducing Starbucks will not make Folgers the “Best Tasting” coffee… even if it’s an unranked, plurality, winner-takes-all taste test!