Salvadoran assembly election, 2018

El Salvador has its nonconcurrent assembly election on 4 March. (Also, municipal elections on the same day.) There are not very many pure presidential systems that have assembly and presidential elections nonconcurrent; this is one of many reasons why El Salvador fascinates me.

Another is the enduring relatively rigid division of the party system between the ex-guerrilla FMLN and the right-wing ARENA parties. Yet another is the free-list (panachage) system adopted in 2015. It makes for pretty big ballots!

El Salvador has presidential elections on a five-year cycle and assembly on a three-year cycle. Since the current regime was established in the early 1980s (in the midst of the civil war), elections have occurred in the same year twice. In 1994 they were concurrent. In 2009, the right-wing bloc in the assembly moved their own election up ahead of the presidential, making for a counter-honeymoon election.

In the most recent assembly election (2015), ARENA won 32 seats to the FMLN’s 31 (the assembly size is 84). In the most recent presidential election (2014), the FMLN candidate, Salvador Sánchez Cerén, won the two-candidate runoff with 50.1%. As I said, it is a closely divided party system! One factor that makes it less rigid than it once was is the emergence of GANA, which split from ARENA after the election of the first FMLN president (Mauricio Funes) in 2009. The emergence of GANA can be seen as an example of “presidentialization” of the party system, in that it was a splinter from the former presidential party to support the new president from the other side of the political divide. GANA remains broadly “conservative” but works with the FMLN in the assembly. It won 11 seats in 2015.

Polls* suggest that ARENA is in the lead. So does the electoral cycle. This election is quite late in the president’s term. The FMLN won 37.2% of the vote in 2015, somewhat down on its first-round presidential vote a year earlier (39.0%). That’s a ratio of assembly to presidential vote of 0.949. The formula presented in Votes from Seats suggests that an assembly election with 20% of the time elapsed between presidential elections would tend, on average, to feature a ratio of 1.055. (See my discussion of the formula and application to France, 2017.) So the party did only very slightly worse than we would expect, based only on a pattern documented from a wide range of elections in presidential systems. This election is at 80% elapsed time between presidential elections, which ought to produce a ratio of 0.620, or a vote of 24.2%. I suspect the Salvadoran party system is still too rigid for the FMLN to fall that far, but anything below 30% would not be a surprise, given the timing of the election. [UPDATE: I used the wrong presidential vote numbers. See comment below.]

As for the free-list electoral system, the voter may cast as many votes as the magnitude of the electoral district (which averages six, but is highly variable according to population). In much of my work, I study the process of “personal votes“; I see now that certain of my assumptions could be called into question: that candidates not only “seek” such votes but that they have consented to being candidates. There actually is an item in the Salvadoran press about a candidate who has asked to be de-listed because she did not consent to being a candidate!

Of course, another question is just how much do candidates campaign for personal votes? Another item from the Salvadoran press, from late December, showed a campaign billboard for a specific candidate.

In this case, the candidate had jumped the gun on the campaign period and was admonished by the electoral authority. So, yes, candidates (at least some) apparently do campaign for personal votes, albeit sometimes too eagerly.

Nonconcurrent elections, a post-civil-war party system, free-list PR. Plus, pupusa contests! El Salvador has a lot to hold my interest!


* The linked article is incorrect on the electoral system. It says that the Salvadoran electoral system has 20 nationwide seats. That was last true for the 2003 election. (The source linked to is from 2013, only a decade after such a system was last used!) There are 84 seats, 14 districts.

9 thoughts on “Salvadoran assembly election, 2018

  1. Pingback: Italy, 2018 | Fruits and Votes

  2. Candidates for the Legislative Assembly have had personalized campaigns for as long as I can remember.


      • Ugh – now you’ve got me doubting my own memory. I’m going to ask a few people, but perhaps I am wrong. I might be confusing races at different levels (mayors obviously ran personalized campaigns), and it also may be uneven. In some places individuals might have a great deal of electoral strength, and thus their name recognition was important. But let me check with folks in El Salvador, because I might be confused on this. (i was present during elections since the 1980s, although not all).

        This photo and caption would support your argument that personalized campaigns started with the ability to vote for individuals in 2012 race, although even then it was just the name, not the photo. However, the photographer and/or caption writer might also be too young to remember previous races!


      • Yes, please check if you can. El Salvador is an interesting critical case for testing the personal-vote thesis (as well as the electoral-cycle thesis).

        Under the Carey-Shugart framework, we would expect some personal campaigning even under the former closed-list system, in the many small-magnitude districts. But its prevalence should increase with the move to open list and even more with the adoption of the free list. And it should increase especially in the high-magnitude districts!


  3. I have to adjust my “prediction” a little bit. I said in the main entry that the FMLN could be expected to get around 24% of the vote. But I based that on a misreading of what their president had won in the first round in 2014. That was actually 48.9% (the 39% I cited was the ARENA candidate).

    Recalculating, based on the correct presidential share, would lead to a prediction of 30.3% for the FMLN. It appears actually to have won 22.2%. That is a good deal worse than expected!

    It is a ratio (of assembly to presidential votes) of 0.45, where 0.62 is expected.


    • One could perhaps use the combined FMLN-GANA vote, given that GANA supports the president and did not run its own presidential candidate in 2014. These parties combined for 32.8% of the vote. That is pretty close to 30.3%!

      In 2015, their combined total was 46.4%. That election should have had a ratio of 1.06, predicting 51.8%. That is also much better than we would get using only the FMLN (which had 37.2% in the 2015 assembly election).

      In Votes from Seats, we do say presidential “party” not alliance. But when the alliance partner does not compete against the larger party in the presidential contest (even when it is in two rounds), it is probably legitimate to use both parties’ combined totals in the assembly elections. But I will leave that to others to judge. It is not clear, on theoretical grounds, which should be preferred.


  4. Pingback: The Salvadoran result 2018: The electoral cycle counts! | Fruits and Votes

  5. Pingback: El Salvador presidency 2019–a president from a different party? | Fruits and Votes

Leave a Reply

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

You are commenting using your 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.