26S Swing-o-Meter

The Swing-o-meter – a tool to estimate the impact that a given shift in the popular vote would have on the overall composition of parliament – is surely...

The Swing-o-meter – a tool to estimate the impact that a given shift in the popular vote would have on the overall composition of parliament – is surely one of Britain’s outstanding contributions to democracy.

Unfortunately, I don’t have the BBC’s graphics department at my disposal, so no pretty pictures. What I do have is this nifty Excel spreadsheet that pretty much replicates the functionality of a BBC-style Swing-o-meter. You just type in a single number number – the swing – and the Swing-o-meter models the effects of that swing, state by state and circuit by circuit, on Venezuela’s parliamentary elections scheduled for September 26th this year. 

The Swing-o-Meter allows you to estimate, for instance, that for Hiram Gaviria to lose his seat in Maracay, the government would need to get a 7.2% swing over and above its result in 2007. In such a scenario, chavismo would end up with 114 seats in the Assembly to our 51…seeing Gaviria lose his race would be pretty scant consolation if things went that badly wrong. 

That’s just one application, though: there are dozens.

Try it: it’s fun!

[Or try: El Swingómetro en Español.]

I also spent some time tracking down the names of all the Government (PSUV) and Opposition (MUD) candidates and typed them into the spreadsheet, just for reference. 

As usual, the election will be fought largely in a relative handful of places. In the U.S. they’d call them battlegrounds, in England "key marginals." I think of them as Tough Seats:

If the results of the 2007 referendum are replicated exactly, the government will take 26 out of these 36 Tough Seats, and walk off with a comfortable majority in the National Assembly. A 5% swing for the opposition in these key races would get us 27 of the 36 seats, and the majority of the A.N. 

Now some technical stuff: A Venezuelan Swing-o-meter is a little bit more complicated than a British one, because Venezuela uses a mixed electoral system. 110 seats are decided by constituencies, 52 by state-wide lists, and three by indigenous voters. To add a further wrinkle, not all the constituencies are single-member: some circuits elect two or even three members to the National Assembly. It is possible, though unlikely, for a multiple-member constituency to split its seats between the government and the opposition.

I went to some trouble to model the D’Hondt proportional representation method for deciding the List deputies before realizing it probably wasn’t worth the trouble:

  1. In the 21 states that elect two List deputies, one goes to the government and one to the opposition in all but the craziest scenarios.
  2. The three largest urban states (Carabobo, Miranda, Zulia), plus the Distrito Capital, elect three list deputies each. In all four cases, the opposition will almost certainly win two out of the three.

As for the other kinks:

  1. My Swing-o-meter assumes that whichever side gets most votes in a given multiple-member constituency wins all the seats in that constituency.
  2. I didn’t bother to do the Indigenous Representatives at all: I just assume all three will be chavistas. 

As I’ve discussed before, I chose the 2007 Referendum on Constitutional Reform as a baseline because it captures what a 50-50% electorate looks like in Venezuela – and it’s only if the overall vote distribution is close that any of this will matter. 

As a final note, please beware: the Swing-o-meter models a linear swing nationwide. If you type in 3%, it tells you what happens if the opposition gets a uniform 3% boost in every circuit. It doesn’t take into account any of the local factors – such as the local economy, the relative attractiveness of the contending candidates, the differential effects of power cuts and crimes, etc. etc. etc – that explain why shifts are never uniform in the real world. The Swing-o-Meter is a tool; it’s not magic. 

Is that clear?

Good. Now, download the thing already! 

Or else: la versión del Swingómetro en Español.