Disproportional Elections

Four years ago, the Republican National Committee was frustrated. The 2012 primary had dragged on for months. The states couldn’t make up their minds. And the party ended up with Mitt Romney, bruised by attack ads even before the Democrats got their first swing.

Republicans agreed this couldn’t happen again. To beat Hillary Clinton, they’d need a strong candidate fast, without all the intraparty bloodletting a protracted primary can bring.

So they built the 2016 primary calendar. It’s slimmer. It slaps down states trying to move elections earlier. And by emphasizing a delegate-allocation system that matches how Republicans actually voted, leaders reasoned the party would gracefully winnow through competition and pick a nominee with a minimum of fuss.

That theory now sports one giant Trump-shaped hole. The billionaire’s towering lead is bolstered by the very system that was meant to prevent a candidate like him from ever making headway.  It’ll go down as one of the biggest ironies in recent political history.

But another well-intentioned tenet of the GOP's reform—a strong preference for proportional delegate selection—has also come into question. Through a series of baffling rules that defy flowcharting, states like Texas and Alabama have effectively unhooked a candidate's popular vote from the number of convention delegates they receive, the real currency of the election. Because of these rules, a candidate could get 30 percent of the vote in several Super Tuesday states and walk away with only a quarter of its delegates—or, conversely, more than half.

Of the 11 states casting ballots this Tuesday, all but Virginia require candidates to hit a threshold of at least 10 percent before they receive delegates. (A contender needs 1,237 to win the nomination.)

Most states also award delegates based on how a candidate performs within the state’s congressional districts. Those counts are often also subject to a minimum cutoff.

If that wasn't complicated enough, several states add or remove a candidate's delegates based on how the rest of the field fares, giving extra points when competition is distant and dinging contenders when they're caught in a scrum.

If that sounds complicated, it is. The official guidance for this stuff sounds like a logic puzzle from the LSATs. Here are the rules from Texas:

At-large delegates: Proportional based on statewide vote with a 20% threshold. If only one candidate reaches the 20% threshold, then the second highest vote-getter is also proportionally allocated delegates. If a candidate receives >50% of statewide vote, candidate receives all [at-large] delegates.

Congressional delegates: Proportional based on congressional district vote with a 20% threshold. If only one candidate reaches 20%, candidate receives all three CD delegates. If two candidates reach 20%, highest vote-getter receives two delegates and second highest receives one. If no candidate reaches 20%, top three candidates receive one delegate each. If candidate receives >50% of congressional district vote, candidate receives all congressional district delegates.

Whoof.

This system could seriously skew results. The congressional-district delegates would seem to introduce geographic variability, punishing candidates who don’t enjoy widespread support.  And cutoffs—especially ones that consider how the surrounding field performs—could lock contenders out of a supposedly “proportional” race.

To test this, I built a quick computer simulation focusing on Texas. The script, starting with an expected statewide vote percentage for a given candidate, randomly assigned votes for the entire five-person field at the congressional-district level a few thousand times, using Texas’s guidelines to see where the delegates would fall.

This obviously isn't how real voting works, and the results tended to clump around statistical averages that might not come into play during the actual election. Even so, it showed how delegate math, a matter of simple of multiplication in most states, could become something approaching calculus in Texas.

First, a candidate's overall delegate count is heavily affected by whether their support is spread across the state or concentrated in a few congressional seats. Low-scoring candidates got a boost if their votes originated from a handful of districts, which allowed them to scoop up local bonuses; the effect was the opposite for statewide favorites, who lost out when their votes were unevenly distributed.

Secondly, if a candidate does have broad support, Texas effectively turns into a winner-take-all state. In several simulations, a front-runner with a 51 percent majority could scoop up all the delegates. Even if that vote was squeezed into a few districts, usually a bummer for front-runners, the leader still got 65 percent of the delegates.

But mostly, I was surprised by the sheer spread of possibilities. If Texas was truly a proportional state, a candidate with 33 percent of the vote could expect 51 delegates. According to the simulation, the probabilities actually stretch from the low 50s to above 100.

Texas is not alone in this; Alabama has virtually identical rules, and states like Georgia have clauses that switch the election into a winner-take-all contest if one candidate reaches a certain percentage.

This is tolerated by the Republican National Committee, which only requires that Super Tuesday states adhere to some level of proportional allocation. But it's hard to argue that the state rules don't alter the spirit of the plan. If the RNC's intention was to peg a candidate's power at the convention to the popular vote, then the states have effectively undone that grand plan. It'll make for an interesting Tuesday.