Archive for February, 2014

I hate (the name) FPTP

February 19, 2014

[This post has nothing to do with politics in Taiwan.  If you only care about that, you can skip this one.  This is a nerdy political science terminology rant.]


I don’t like First Past the Post (FPP*).  For one thing, it’s a lousy electoral system.  However, that’s a topic for another day.  What I’ll be hating in this post is the name itself.

(*The acronym is also an irritant.  Most people write FPTP even though we drop the T for almost every other use of “the.”  So you can have the inconvenience of writing a four-letter acronym (and remember, the whole point of an acronym is brevity), or you can use a three-letter acronym and look like you are unaware of the rest of the world.  Ugh.)

FPP rests on an analogy: elections are like a horse race.  Just as the first horse to run the whole track and pass the post at the finish line is the winner, so it is in plurality elections with only one seat at stake.  Just as the fastest horse wins, so does the candidate with the most votes.  Easy.

Except that single member plurality (SMP) elections are nothing like a horse race.  In a nutshell, there is no fixed post.  In a horse race, the finish line is set out clearly before the horses start running.  Every jockey knows that the race will be exactly one mile (for example).  The competition is to go exactly one mile faster than everyone else.  In an election, no one knows where the finish line will be before the voting starts.  The candidates may know how many eligible voters there are, but they cannot say (with any certainty) how many votes they will need to win.

What would a true FPP election look like?  Imagine a mayoral race in a city with 100,000 eligible voters.  Where is the post?  In a FPP horse race, the post is always attainable.  (Have you ever seen a horse race where all the horses dropped dead from exhaustion before reaching the finish line?)  So we have to imagine what the lower range of turnouts is and then we have to account for an unexpectedly large number of candidates running.  So let’s assume that we are really, really sure that at least 50,000 voters will turn out, and we can’t imagine the race becoming so fractionalized that no one could get 20% of the vote.  That implies that we should set the post at no higher than 10,000 votes.  (Remember, that’s a maximum; it would often be lower.)  So the electoral commission advertises the election, saying the first candidate to get to 10,000 votes will be named mayor.  Several candidates register, thinking they can get 10,000 votes.  On election day, the race is to get your votes out as early as possible.  Votes would have to be counted as soon as they are cast (so much for secrecy), and somehow all the precincts would have to be synced up so that they all reported simultaneously.  If voting starts at 8:00am, it might be over by 8:30.  One imagines that campaigns would become expert in lining up voters before the polls opened so that their votes would be cast before opponents’ supporters had a chance to weigh in.  They would also practice voting as fast or as slow as possible so that the people in line would or would not have a chance to vote, depending on past voting history at that precinct.

This race looks nothing like a SMP election.  There is no reason to expect there will only be two candidates.  On the one hand, lots of horse owners think their horse can run a mile really fast, and, in a city with 100,000 eligible voters, several candidates will think they have 10,000 supporters.  If the race is to get 10,000 supporters to the polls as early in the morning as possible, the true FPP election is not a contest over the median voter.  Rather, you will want to mobilize a highly energized but relatively small fraction of the electorate.  (Wait, that sounds more like PR with large M!?!)  On the other hand, there is no reason to aggregate.  In an election, two candidates with overlapping support might split their vote and both lose.  In a horse race, no horse runs faster or slower because a similar horse is also in the race.  If one horse drops out, its time isn’t added or subtracted from any other horse’s time.  Most horse races have 8 or 12 entrants; most SMP races have far fewer.

I could go further.  For example, a lower or higher turnout would be akin to jockeys being able to move the post closer to or further away from the starting line.  However, I think the basic point is clear by now.  SMP elections are not like horse races since there is no fixed post.  Consequently, there is no way to be the first past the post.  The analogy and the terminology are fatally flawed.  If you think it through to the logical conclusions, a “true FPP” election would actually look more like the polar opposite of a SMP election.