Honduras, 2013 election: Will electoral reform, a la Uruguay, be adopted?

The second-place presidential candidate in the 24 November Honduran election, Xiomara Castro, is claiming fraud. The official results show a widened lead for the winner, Juan Orlando Hernandez, compared to what was reported the day after the election: now at 36.8% to 28.8%.

Vote compilations released by Castro’s Libre party show substantial discrepancies compared to the official results. However,  the claimed irregularities amount to only about half the official margin. Thus even if accurate, they would not change the result.

Honduras elects its president by plurality. One might wonder if the traditional parties–Hernandez’s National Party and the weakened Liberals–might now work towards the implementation of some form of two-round system for future presidential elections.

Had a runoff been required in this year’s election, there is little doubt that Hernandez would have won. In a second round, he would have received the bulk of the Liberal vote, which was 20.3%, making for a likely comfortable majority. Even the votes of the other debutant party in this election, Salvador Nasralla’s Anti-Corrpution Party, on 13.5%, would put Castro  clearly short of half the vote–even if we assume that Castro’s “real” vote was more like 33%.

The situation is somewhat analogous to Uruguay in the 1990s. Like Honduras, Uruguay is one of the Latin American countries in which two “traditional” parties survived, and dominated, the longest. Uruguay likewise elected its president by plurality (actually an “open list” whereby the winner was the plurality candidate within the party that had the overall plurality, a system also used in Honduras on a one-off basis in 1985). With the threat of a rising new party on the left, the traditional parties moved to a two-round majority system (with primaries to set each party’s nominee). At first it worked as a means to block the upstart, with the traditional parties combining to win a runoff in 1999 against the Broad Front candidate who had a plurality (40% to 33%) in the first round. Then, in 2004, the Broad Front was able to win a first-round majority and elect the president. (Another Broad Front candidate was elected in 2009, this time requiring two rounds.)

Such a trajectory may be the best Honduras can hope for. Of course, Honduras is not Uruguay, in many respects. But the old two-party system now has broken down, and new parties are knocking on the door. The adoption of a runoff for presidential elections may be necessary not only for the power-preserving interests of the traditional parties, but also to prevent a future regime-threatening crisis of legitimacy–assuming, that is, that the country’s democracy gets past the current, likely spurious, claims of the loser in this year’s plurality contest.

13 thoughts on “Honduras, 2013 election: Will electoral reform, a la Uruguay, be adopted?

  1. I am still waiting for the electoral commission to produce digestible congressional results. The way they are currently organized–and I use that term loosely–makes it extremely inconvenient to get a sense of the result.

  2. Thanks MSS, hope you’re enjoying the holiday 🙂 I’ve updated the data with the most recent count, and also done some more processing on it.

    Here’s another CSV file with information on which candidates are elected, and with summary lines: https://github.com/vasi/honduras-election-2013/blob/master/elect.csv

    Here’s a textual summary of the results: https://github.com/vasi/honduras-election-2013/blob/master/formatted.txt

    Here’s how far along the counting has gotten: https://github.com/vasi/honduras-election-2013/blob/master/polls.csv

  3. How would I save the .csv file to import into Stata (or Excel, etc.)?

    From a very quick perusal, one thing stood out: it appears that Libre candidates have less variance within a district in their preference-vote totals. As I understand the Honduran system, a voter may vote for as many candidates as there are seats in the district. In such a system, we might expect parties that have a strong label to have less variance in their candidates’ vote shares, and it looks as if Libre was such a party. The traditional parties, on the other hand, may be more dependent on the popularity of specific candidates (which would produce variance in their vote totals).

    It would be easy enough to calculate some simple statistics–if I could get the file into Stata.

  4. You can click the “Raw” button to get just the CSV data. You can also use this link to download a ZIP file of the repository, including all the CSV files: https://github.com/vasi/honduras-election-2013/archive/master.zip . I have no idea what sorts of formats Stata likes, if CSV is no good, I can generate something else instead.

    Interesting note about the voting system, I was wondering why even smaller parties run a full slate in a 23-seat district! One of my issues with open-list PR has always been the difficulty of registering a negative vote, and this system would seem to solve that problem.

    • Thanks. I will try this. I did not think of the “raw” button. But the zip archive will do the job, and, yes, .csv is fine for Stata.

      Latvia is the only OLPR system I know of in which voters can make either positive or negative votes. Honduras system does not really allow that, although of course one can vote for every candidate except those one dislikes.

    • Is “elect.csv” the same as “candidate.csv” other than a code for whether the candidate was elected or not? Or is there some other difference?

  5. It also has some summary lines near the bottom:

    – Lines starting with ‘party’ are the aggregate votes and seats for a party in a department.
    – Lines starting with ‘dept’ are the total votes and seats of a department.
    – Lines starting with ‘total’ are the total votes and seats nationally, per-party or as a whole.

  6. Also, the electoral commission’s website doesn’t indicate which candidates are elected, so my scripts figured that out. I’m adding the candidate votes in a dept. to get the party vote, then using largest-remainder with a Hare quota to find the number of seats S for each party, and finally picking the top S candidates in that party by votes. I think that’s the actual system in use, but correct me if I’m wrong. Maybe votes per-candidate can’t be simply added to yield the vote per-party, because of the multiple vote thing?

    It’s also possible there’s a bug in my code, so feel free to do spot checks. Currently I don’t handle special-cases at all: Ties are resolved by alphabetical order. If a party wins more seats than it has candidates, which is extremely unlikely in this system, the script would probably produce strange results.

    • Of course, why would an electoral commission want to indicate who was elected?

      I actually am not sure of the rule, though I think it is what you describe. I don’t think there is a party-vote option, given that the results do not seem to show any party votes. So unless they apply some sort of formula to weight the candidate votes, I assume a party vote is just the sum of candidate votes.

      Thanks again!

    • I will need to look at this more carefully (and will make a separate post when I have), but preliminary conclusion: yes, the standard deviation of candidate vote shares within a district is significantly lower for Libre than for National. Liberals (older but now smaller traditional party) have even greater standard deviation.

      That’s without any controls for party strength in a department or for the department’s district magnitude, so caveats apply.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.