UPDATE: See the comments by Manuel for important detail and alternative scenarios!
____________
The results of the Salvadoran legislative (and municipal and the Central American Parliament) elections are now posted at the site of the Tribunal Supremo Electoral.
The seats in the assembly are as noted earlier:
-
35 FMLN
32 ARENA
11 PCN
5 PDC
1 CD
There is also a votes summary that can be downloaded (under “Otros Reportes” strangely enough, as if it’s an afterthought). As far as I can tell there is no posted national summary of votes. If one wanted to add up all the departmental votes one could confirm the final vote total for the FMLN (last reported at around 43%) and other parties. But this one does not want to.
The departmental summary is interesting, however, in that it gives the full allocation process, showing the quota, seats by quota, remainder, and seats by remainder. Once again, the PCN lives a charmed life, winning one seat in each of 11 departments, always on a remainder. The large parties use up most of their votes on quotas, leaving the PCN to win a seat even in a 3-seat district like Chalatenango where it had only 11.4% of the vote.
While the FMLN is probably not happy with its showing, it is worth remembering that if it indeed won 43% of the vote, that’s the highest total a party has obtained in any of the last five legislative elections. The only higher share since the FMLN began participating was in its first election, 1994, when ARENA scored 45% (in the election that was concurrent with the presidential election). It would also be the FMLN’s own highest total ever, by around 3 percentage points.
While I am not in the business of making predictions, the FMLN’s prospects heading into the presidential election in March look about as good to me as they did before the election.
Nationwide vote totals of the 2009 legislative election in El Salvador, available on the TSE’s website under “Consolidado Nacional Votación de Diputados,” are as follows:
FMLN - 943,936 (42.6%)
ARENA - 854,166 (38.6%)
PCN - 194,751 (8.8%)
PDC - 153,654 (6.9%)
CD - 46,971 (2.1%)
FDR - 22,111 (1.0%)
Seed planted by Manuel Alvarez-Rivera — 27 January 2009 @ 03:15
Thanks, Manuel.
OK, how many electoral systems would not only under-represent the largest party but also the second largest?
Advantage ratios (% seats/% votes):
0.978 FMLN
0.987 ARENA
1.49 PCN
0.863 PDC
0.567 CD
Very odd, but a direct result of the use of simple quota and largest remainders in (mostly) low magnitude districts.
Victor D’Hondt, where are you when we need you?
Seed planted by MSS — 28 January 2009 @ 20:47
You’re welcome, but you may not like what the D’Hondt rule would do to El Salvador’s Legislative Assembly:
42 FMLN
38 ARENA
2 PCN
2 PDC
0 CD
And modified Sainte-Laguë wouldn’t really be much of an improvement either:
41 FMLN
37 ARENA
3 PCN
3 PDC
0 CD
Meanwhile, pure Sainte-Laguë would deliver a rather peculiar seat allocation:
36 FMLN
35 ARENA
9 PCN
3 PDC
1 CD
However, there is another alternative…and we don’t have to stray far from El Salvador to find it.
Nearby Costa Rica also uses simple quota and largest remainders for the distribution of seats on its unicameral Legislative Assembly, but the Costa Rican system provides for a sub-quota (subcociente in Spanish), equal to half the simple quota. If a party has won no full quota seats and its remainder votes are below the sub-quota, it’s not allowed to take part in the distribution of unfilled seats. That provision prevents parties from capturing remainder seats with very low vote totals.
So, under the Costa Rican system, El Salvador’s Legislative Assembly would look like this:
37 FMLN
35 ARENA
8 PCN
3 PDC
1 CD
Compared to the actual election outcome, FMLN would have picked up two additional seats and ARENA three, at the expense of PCN (-3) and PDC (-2).
Seed planted by Manuel Alvarez-Rivera — 29 January 2009 @ 03:36
What’s not to like about the hypothetical D’Hondt allocation? Fifty percent of seats (not a majority if the other parties stand firm against a proposal of the plurality party) on 43% of the votes. Not too bad an outcome, but if a 7 percentage point boost for the largest party (and A=1.17) seems like too much, a good solution would be to increase average magnitude. Or keep the current districting but make the national tier compensatory rather than supplementary.
In any case, I concur that the “Costa Rican” variant allocation somehow seems about “right.” And I am grateful to you, Manuel, for clarifying the rule used there. I am writing something about electoral systems and reforms in the region, and I was having a hard time figuring out just what Costa Rica’s rule was, based on the description in the Nohlen volume. Your description is so much more clear!
Seed planted by MSS — 29 January 2009 @ 17:12
Actually, I don’t find the hypothetical over-representation of FMLN under D’Hondt to be particularly objectionable or for that matter exceptional. In Portugal, which uses the D’Hondt rule to elect its unicameral Assembly of the Republic, the Socialist Party won 120 of 226 seats in Portugal proper (53.1%) with 45% of the popular vote back in 2005, while in Spain - which also uses the D’Hondt method to choose members of the Congress of Deputies - the People’s Party (PP) secured 183 of 350 seats (52.3%) in 2000 with 44.5% of the vote.
However, the potential under-representation of PCN and PDC under D’Hondt (or even modified Sainte-Laguë in the case of PCN) comes across as rather extreme when compared to Spain or Portugal, even though smaller nationwide parties are usually under-represented in the legislatures of both countries (especially the former, where the average constituency size is smaller). For example, in Portugal’s 2002 general election, the People’s Party (CDS-PP) won 14 of 226 seats (6.2%) with 8.8% of the vote, while in Spain’s 1989 parliamentary election, the United Left (IU) obtained 17 of 350 seats (4.9%) with a 9.1% share of the popular vote.
By the way, El Salvador’s Legislative Assembly electoral system no longer provides for a national tier: it was scrapped a few years ago, and all Assembly seats are now allocated at the multi-member constituency level. At any rate, the current apportionment has its share of problems as well. For example, the departments of Cabañas and Cuscatlán both elect three deputies, but in the former only 18,003 votes were needed to secure a full quota seat, whereas in the latter the quota stood at 32,404 votes.
Finally, I’m very happy to hear you found my summary of Costa Rica’s electoral system useful.
Seed planted by Manuel Alvarez-Rivera — 30 January 2009 @ 01:35
Thanks for the clarification on the national tier. I noticed there was no voting result for such a district in the summary I had downloaded, but I had not yet added up the magnitudes of the regional districts to confirm that they covered all 84 seats.
So, magnitudes must have been adjusted upwards, as it appears the assembly size has been fixed for some time (back to 1991, according to the Nohlen volume).
By the way, the Nohlen volume refers to the national tier as of 2003. Assuming it is correct, that would suggest maybe 2006 as the first election without this tier.
On the under-/over-representation of small parties under D’Hondt, of course this is small magnitudes and regional concentration (or its absence) at work. Many of the small parties in Spain are quite concentrated (and thus aided by small magnitudes in the case of regional parties, or by the large magnitudes in Madrid and Barcelona in cases like the IU). The PCN is pretty uniform in its size across districts, and so it is precisely the sort of party that would be under-represented with low-M D’Hondt (and over-represented with low-M SQLR).
Seed planted by MSS — 30 January 2009 @ 17:23
In an April 2003 decision, El Salvador’s Supreme Court of Justice ruled the Legislative Assembly’s national constituency was unconstitutional, and struck down Article 13 of the electoral law (Código Electoral), which deals with the apportionment of Legislative Assembly seats; the full text of the court decision is available in Spanish here. Note that the citizens who brought forward the case were all FMLN leaders, headed by none other than the late Schafik Handal.
As a result of the court ruling, the electoral law was amended in 2005 to allocate all of the Assembly’s 84 seats among El Salvador’s fourteen departments; the full text of the amending decree can be found here, while the TSE website has the full (Spanish-language) text of the electoral law here.
Interestingly, there is not much difference between the average multi-member constituency size in El Salvador (six seats) and Spain (almost seven). However, while the smaller nationwide parties in Spain - and by this I mean the smaller parties than run throughout the entire Spanish territory and not just a region or two (in Spanish I would refer to them as “partidos menores de implantación nacional”) - usually concentrate their efforts in the larger constituencies, especially Madrid (Barcelona can be a tough sell because of the additional competition from the Catalan nationalist parties), in El Salvador both PCN and PDC fared poorly in the largest constituency, San Salvador, which returns 25 deputies. Of course, given the way the Salvadoran electoral system works, both have little incentive to pull all the stops in San Salvador, since they can easily pick up seats elsewhere on remainder votes.
Finally, you’re absolutely right that in Spain some smaller parties fare comparatively well in the distribution of Congress seats, but as you suggested these happen to be nationalist parties from the peripheral regions; I’ve pointed out this well-documented phenomenon in Elections to the Spanish Congress of Deputies and Spain Votes 2008: Zapatero’s election to lose?, where I wrote: “Because votes cast for nationalist parties are concentrated in the regions in which they operate, these can circumvent the limitations of the rectified proportional system in varying degrees.”
Seed planted by Manuel Alvarez-Rivera — 30 January 2009 @ 21:28