Bloggings by boz

I see numbers being debated in the Honduran media and online. This is a good post to put up before the election, so nobody on either side can accuse me after of setting the bar once I know the result.

Here are the turnout numbers for the past few election cycles:

1993: 2.73 million voters, 1.78m total votes, 1.71m valid votes
1997: 2.90 million voters, 2.10m total votes, 1.98m valid votes
2001: 3.45 million voters, 2.29m total votes, 2.18m valid votes
2005: 3.98 million voters, 2.19m total votes, 2.00m valid votes
2009: 4.6? million voters, ?.??m total votes, ?.??m valid votes.

The Honduran election tribunal adds registered voters to their census every four years and the 4.6 number continues/accelerates the trend. The 4.6 million number seems off to me and even Honduran election officials have questioned the number, although that may be their attempt to lower expectations.

What we see from the last three election cycles is that the number of registered voters increases, but the number of actual voters doesn't increase at the same rate. So let's ignore turnout percentage for a moment and look at total votes. Fortunately, you don't have to be a statistics junky to eyeball what we should expect in this election. There should be about 2.25 million total votes and about 2.1 million valid votes in 2009 if raw number trends continue as they have for the past decade. Now looking at turnout, if you adjust for increased voter registration, there should be about 2.5 million total votes and 2.3 million valid votes. I'll leave it to the reader to decide whether adjusting for increased registration makes sense and how far off they want to claim I am.

I think both total votes and valid votes are worth considering, but valid votes are more important. There are two ways to boycott an election. One is to not show up. However, in situations where voting is mandatory or there are various institutional pressures to vote, another way is to cast a blank or null ballot, which will reduce valid votes. So I recommend anyone looking at this problem take both numbers into consideration and use valid votes as the main number.

Once we know the number of voters and valid votes, there will be two questions. What would be a statistically significant change in voter behavior? And what would be an effective boycott that raises real questions about the legitimacy of the election? The statistically significant question is interesting, but less important to the actual politics of the situation. From a political strategy point of view, the question is what number will constitute an undeniable victory or loss for the boycott movement.

Here's the number Zelaya has to beat: 1.5 million valid votes.  Below 1.5 million, there is no doubt that his boycott is a success and no amount of spin that can change that fact for the de facto government. That will mean at least half a million people, 25% of the expected electorate (and possibly more) chose to actively protest by not voting when they would have otherwise. That number would create a legitimacy problem for the next president and almost certainly force a change in action by the de facto government before the inauguration.

Here's the number the de facto government has to beat: 2.4 million valid votes. At 2.4 million, they'll cross the 50% mark in turnout, beat 2005's turnout and be able to proclaim it the highest raw number turnout ever in Honduras. At that point, the boycott is statistically irrelevant and they get to say more people voted than voted when Zelaya was elected in 2005.

Here's a scenario that's likely to happen. The valid vote total number is going to come in somewhere between 1.7 and 2.3 million valid votes. The Zelaya supporters are going to proclaim their boycott a massive success, saying that over 50% of voters didn't show up, the lowest turnout in recent history. The election supporters are going to proclaim the boycott a failure, saying that Zelaya's supporters consisted of less than 10% of voters, maybe 300,000 people who would have voted who didn't show up. Both sides will disagree over the interpretation of the results the same way they have disagreed through this whole process and nobody changes their mind or feels politically pressured to change their actions. There is over a half million vote gap in which the results are going to be ambiguous and contested by the two sides.

I'm sure I will hear complaints about setting the bar too low or too high on all those numbers. Everyone is trying to set expectations right now. I would encourage people to put out their own numbers before the election.

The unfortunate fact for the resistance is that bar for Zelaya to succeed in the boycott is higher than the bar for the de facto government to muddle through the election. Boycotts are very hard to pull off effectively from the point of political strategy, particularly in a country where only a slim majority of the population voted in the last election. Boycotts in general, are usually ineffective at making their point and often weaken the side that chooses not to participate (see Venezuela 2005 for a perfect example). Can Zelaya's boycott be the exception? We'll start to find out this weekend.