Peru elections 2011

Peru held elections Sunday for president (first round), congress, and (if anyone cares) Andean Parliament.

The president is elected by two-round majority. The front-running candidate won just over a quarter of the vote: Ollanta Humala, with 27.4%. As is often the case with fragmented first-round fields, the race for the second slot in the runoff was closer than the race between the top two. Keiko Fujimori appears to have made it in with 20.8%, but as just over 82% of returns have been processed, her margin over Pedro Pablo Kuczynski, at 18.1%, is not safe yet.

Former president Alejandro Toledo ran fourth, currently on 13.6%.

Fujimori is, of course, the daughter of the former president, Alberto, who now resides in a jail cell. Kuczynski is a former prime minister who served during part of Toledo’s presidency. (Yes, Peru has a semi-presidential system, of the president-parliamentary subtype, and not a pure presidential system.)

As is typical in Peru, the party system barely deserves the name. The party of the incumbent, Alan Garcia, did not even have a candidate in this process. This party, APRA, has been a major party in Peru since the 1930s, although it has held the presidency only twice, both times with Garcia (elected 1985 and 2006).*

The names of the top five candidates’ parties tell us little about what they stand for: Peru Wins, Force 2011, Alliance for the Great Change, Possible Peru, and the National Solidarity Alliance (my translations). The candidate of Justice, Technology, and Ecology managed only 0.06% and National Awakening slumbered to 0.12%, while Forward remained stuck below 0.1%.

Far less of the congressional vote has been processed at this point. Peru’s unicameral congress is elected by open-list PR, with most districts having magnitudes in the 2-9 range, except for Lima (M=35).

UPDATE: Rici has some corrections on the congressional districting and other useful information in a comment, and boz also addresses the congressional result.

* Its “populist” founder, Victor Haya de la Torre, won a plurality in 1962, with 33% at a time when the rule was that one third of the votes was sufficient for the front-runner to be elected. Otherwise the legislature selected from the top three. A military coup annulled the results of the 1962 election.

20 thoughts on “Peru elections 2011

  1. Actually, the congressional districts are all in the range 2-7, except Lima which is now M=36, and Madre de Dios which is still M=1.

    Since 2006, 10 seats have been added, and the non-metropolitan part of Lima was hacked off into a 26th district, so Lima is now somewhat less under-represented than it was before.

    There’s probably a lot more substantive things one could say about this election, but I’m not sure that I have the energy. I did do a simulation of the current state of the congressional vote, resulting in the following: (I put the presidential candidate’s name in parentheses because, except in the case of APRA which has none, it’s about the only thing that might make the name meaningful to anyone not following closely.)

    “Party” (Presidential candidate) | total vote | % | total seats | %

    Gana Perú (Humala) 823321 26,82% 41 31,54%
    Fuerza 2011 (Fujimori) 673853 21,95% 35 26,92%
    Perú Posible (Toledo) 463622 15,10% 22 16,92%
    Alianza por el Gran Cambio (Kuczynski) 395126 12,87% 16 12,31%
    Alianza Solidaridad Nacional (Castañeda) 325211 10,59% 11 8,46%
    Partido Aprista Peruano (APRA) 204466 6,66% 5 3,85%

    (Cambio Radical would have won two seats had it not been for the 5% bar; it currently shows 2.8% and I doubt that it can change enough to reach the bar. The unrepresented parties, including Cambio Radical, sum to 6% of the popular vote.)

    Of course, this simulation has no official weight, and is based on early results, particularly in Lima (which is only at 7%), so it will probably change a bit. But, assuming it doesn’t change too much, it seems that the only ruling coalitions possible are:

    Humala + Fujimori


    either Humala or Fujimori, plus at least two of the other parties (roughly).

    We’ll see how that plays out…

    • Rici, welcome back to F&V! Thanks for the updates on the magnitudes and the data on congressional results.

      It is striking how much worse Kuczynski’s lists are doing than he himself did, with Toledo’s running third. I assume that’s because Toledo has much deeper support in the countryside than Kuczynski.

  2. I don’t know how deep Toledo’s support is, but it seems that he has more of it outside of Lima than PPK.

    At some point it will be interesting to compare presidential and congressional vote by district, but I suspect it will not yield to any analysis which doesn’t take into account the existence of significant regional actors who have entered into alliances with some presidential candidate, for whatever reasons, whether strategic or simply affective.

    The fact that Peruvian voters can issue preferential votes within a congressional list makes it possible for someone well-known in a region to take advantage of a political aggrupation. It has been suggested that this creates a centripetal force acting against party cohesion. Although I’ve been pretty resistant to that theory over the years, I have to admit that that Peru seems to be very hostile ground for the formation of stable political parties, and it doesn’t seem to be getting any better.

    • I think open-list PR is a very small part of the explanation for why Peru’s parties are weak. However, given that they are weakly organized, the ability to pull votes to a list by nominating a regional notable only makes them weaker.

      Voters in Peru are allowed more than one candidate preference vote, correct?

  3. Voters can vote for up to two preferences but only if those candidates belong to the same congressional list. The only exception is the electoral district Madre de Dios, where M=1, so voters can only vote for one preference.
    If a voter votes for two preferences belonging to different congressional lists, then his congressional ballot turns null.

    It’s interesting rigth now what will hapen in the presidential election. This is not the first time that the leftist nationalist Ollanta Humala wins the first round: In 2006, Humala won with 30.7%, and Alan García was second with 20.4% (now: Humala 31.9%, and Keiko Fujimori 23.3% so far). In the second round, García got the support of most of other candidates and ended up with the 52.6%.
    If one follows the same pattern, the other three candidates in the 2011 election seem to be more reluctant to Humala than Keiko, mostly for one reason: all of them want to preserve the current economic model. Although Humala has promised he will not break down the model, but correct it (the “Lula way”), the other candidates don’t believe him. Thus, the rightest Keiko comes up as the “lesser evil.” As Mario Vargas Llosa said, Peru is really having two options: “The suicide or the miracle.”

    • Do parties in Madre de Dios (M=1) present lists with more than one candidate?

      Practice in other open-list countries where at least one district has only one seat varies. In the Aaland Islands district in Finland, parties routinely submit multi-candidate lists. In the two one-seat districts of Switzerland, however, each party has a single candidate, reducing the contest to FPTP.

      In these cases, the practice in the one-seat districts is a simple generalization of practice in the country as a whole: In Finland, parties may submit up to 14 candidates or M candidates, whichever is greater, whereas in Switzerland, M is the upper limit on the number of candidates per list.

  4. One thing that strikes me here is that the Congressional elections are now over but the presidential elections continue. Yet, the Congressional elections appear overshadowed by the presidential race and the “parties” are mostly just coalitions supporting one presidential candidate.

    We’re about to see a presidential election where one candidate could rise or collapse in the coming month, and with it, people’s preferred Congressional vote may change.

    Could a case be made that the Congressional elections should have been held concurrent to the 2nd round instead of the first, so voters had more information? Hypothetically, how would that have changed the congressional makeup? Are there any examples of this anywhere in the world? An obvious problem is that there was no certainty there would be a 2nd round.

    • The only case I know of in which a congressional election was scheduled to be concurrent with a possible second round of a presidential election was Ecuador in 1979 (and there was a runoff). After that, Ecuador switched to concurrence with the first round, and I have looked but never located another case of second-round concurrence.

      I think a reasonable case can be made for second-round concurrence, which might boost the prospects that the incoming president has a good block of legislators around which to form a legislative coalition. Coattails would probably enhance the winning presidential candidate’s coalition–even if that coalition was forged only after the first-round results were known–and, within that coalition, it would also boost the president’s own party. At least in theory.

      I suspect that this might even be precisely the reason it is so rare: the more fragmented and fluid the party system, the less willing the parties are to risk their legislative performance on the coattails of a two-candidate race. And it is precisely the more fluid and fragmented party systems where two-round presidential systems are likely to be adopted in the first place. Another objection parties in places where two rounds are adopted for the presidency is that, were the president to be elected in one round, the following “honeymoon” congressional election could be expected to give a very big boost to the by-then-known new president. That might just be too big a risk if you are in any party not confident of dominating.

      I think Rici’s follow-up post to boz’s suggests some reasons why, on the contrary, concurrence with the first round might be normatively preferred.

      I have always rather liked the idea of congressional elections first, followed by presidential. This has been used in a few cases, including Colombia (most cycles since 1958) and Cape Verde. Also it was done in El Salvador‘s most recent cycle, where the congressional election result probably is what prompted candidates other than the top two to drop out of the (first round of the) presidential election. There is also the case of Nigeria, which is in the midst of such a cycle just now.

  5. The case certainly has been made that congressional elections should be delayed until the second round, but there is a good counter-case, too: it is probably too risky. (I’m not talking about the risk of there not being a second round; you could simply hold the congressional elections on that day regardless, since you would already have had to budget for a second election.) It’s not difficult to imagine the arguments which would be made by both sides in such a hypothetical competition, particularly one as close as the previous (or current) Peruvian election.

    Actually, I was just thinking exactly the reverse: I wonder if the composition of Congress might influence the second round vote, which I suspect will be a very difficult choice for many Peruvians.

    Now that ONPE has progressed in the count (and at an amazing speed compared to previous elections), I have to correct my previous numbers — I did warn that they were unreliable. The current simulation (which, of course, is still unofficial and subject to variations, and still is only based on about a quarter of the vote in Lima, but is probably closer to the final result) shows a different set of coalition possibilities from the earlier simulation. The unlikely coalition Humala+Fujimori is still possible, of course, but there is also the interesting possibility Humala+Toledo. Aside from H+F, the only coalition possibilities for the Fujimoristas involve Perú Posible and at least one other party, and APRA is almost irrelevant.

    In other words, Perú Posible is in a very good bargaining position, and this may well have a significant influence on the second round.

    That seems to me to be a somewhat healthier position than we might see if the congressional election had not taken place, since it is likely to lead to actual negotiations, and not just posturing. Of course, there will be a lot of posturing, too; I don’t mean to sound naïve.

    I also think that it’s worth having a clearer understanding of why political parties as institutions are doing so poorly (and not just in Perú). I don’t think they are weak so much as fantasmal, but I also don’t think that characterizing them as coalitions around presidential candidates quite captures it, either. Certainly, all three “outsider” candidates — Humala, Fujimori and Toledo — benefit from some sort of organization sufficient to attract votes in parts of Perú which are not particularly visible from Lima. But that organization may not bear much resemblance to what a European might think of as a political party. Although, even there I have some doubt: I am no longer sure I understand what the modern institution of political party is, either in Europe or in the country in which I grew up. How does one account for a Berlusconi or a Sarkozy? Or a Le Pen?

    But too much speculation for a little comment. Here’s today’s congressional numbers, anyway, same format as yesterday:

    Gana Perú (Humala) 2067881 26.98% 45 34.62%
    Fuerza 2011 (Fujimori) 1735344 22.64% 37 28.46%
    Perú Posible (Toledo) 1128467 14.72% 22 16.92%
    Alianza por el Gran Cambio (Kuczynski) 989198 12.91% 13 10.00%
    Alianza Solidaridad Nacional (Castañeda) 775194 10.11% 9 6.92%
    Partido Aprista Peruano (APRA) 514754 6.72% 4 3.08%
    Other 453341 5.92% —

  6. In Peru, parties must present lists of minimummaximum(3,M) candidates, and the lists must be gender-balanced (at least 30% of each list must be of each gender).

    I don’t believe parties can present smaller lists, but it is possible for the electoral organization to invalidate candidates (and then the party has the opportunity to replace them with other candidates), and this can result in shorter lists, I believe. However, the resulting list must still obey the gender-balance rule, which can — at least in theory — result in the entire list being invalidated.

    In the M=1 district, voters only have one “preferential” vote (although they’re not obliged to use it; they can simply vote for a party). In all other districts, voters have two preferential votes. In all cases, the voter must select a party and the preferential votes, if used, must be for candidates of that party.

  7. Indonesia is other case where congressional elections are first; moreover, only a party or coalition winning 25% of votes (or 20% of seats) of the House of Representatives (PDR) can nominate the presidential and vice-presidential candidates.
    In terms of potential government coalitions in Peru after the second round (in two more months), I am wondering how disciplined each “loser” party/alliance will be at that time. Are parties cohesive after elections? If Perú Posible (Toledo) gets an agreement with Gana Perú (Humala), it could be possible to see elected Perú Posible’s congressmen defecting, and joining Fuerza 2011 (Keiko) for instance? And on the other hand, are loser presidential candidates really capable to endorse their electors’ votes to either Humala or Keiko effectively?

    As for limited number of congressional candidates in Peru, parties/lists can nominate as many candidates as the magnitude of the district allows; so, max. number of candidates = M. Exceptions are made with districts of M=1 and M=2, where parties/lists can nominate up to 3 candidates. I don’t know if there is a minimum, since I saw a list (Alianza por el Gran Cambio) presenting only 2 candidates in Madre de Dios.

    • How did I forget Indonesia? After all Eduardo has taught me…

      (He is a student of mine, and he and some other students have been doing some projects on Indonesian elections and parties.)

  8. Obviously, I meant maximum(3,M). Also, I checked the elections law, and it seems that if a candidate is disqualified from a list, that does not put the rest of the list in danger, even if the result is that the list is not gender-balanced.

    In any event, in practice, most lists are made complete although disqualifications are not uncommon.

  9. Does this mean Aaland Island could see its sole seat won by the candidate with a plurality of votes on a 14-candidate list?!! – if, say, the Kokoomus list leads the field with 43%, and its highest candidate has 4% while the other 13 have 3% each?

    Is this possible in principle? If so, how unlikely is it in practice?

    • On Aland island, I just happen to have the data for four years (1995-2007). Nothing quite as interesting as what Tom asks about has happened during that time, although in principle it could. Maybe not as dramatic, as I am not sure now about the 14 candidates. That may be an upper limit only to districts with M>1 (and less than 15). All parties contesting the district in these elections have nominated 3 or 4 candidates.

      The winner had 75% to almost 90% of his or her list’s votes in 1995-2003, as well as a plurality of the overall district vote. Only in 2007 was there close competition for the seat, but it was intra-list, with the winner having 45.9% of the list’s vote and the runner-up 42.1%. Each of these candidates in the winning list in 2007 had more than three times the votes of the most popular candidate in the only other list.

      I would like to check results from Peru’s one-seat district for comparison. Aside from Aland, it is currently the only case I know of with OLPR, M=1, and more than one candidate per list allowed. Maybe there are others…

  10. Of course, OLPR was used to elect the Uruguayan president for many years, even after the collective presidency was abolished. So there was another M=1, c>1, case (where c is the number of candidates per party, in case that was not obvious).

    It did happen sometimes that the winner had fewer preference votes than a loser in another party. For example Bordaberry of the Colorado Party was elected in 1971 with about 60,000 fewer votes than Ferreira Aldunate of the National Party. Bordaberry won because the combined votes of the PC trumped those of the PN by around 17,000 votes (41.0% to 40.2%). (Data from the Nohlen handbook.)

  11. Sorry, I can’t find much historical data for Madre de Dios. ONPE doesn’t seem to have full congressional results from before 2006 online anymore, and I’m not sure if I archived the data I grabbed at the time.

    Let me emphasize here that the Peruvian system is significantly different from the Finnish system[1] in that Peruvians vote for a party, always, and optionally for a candidate (or two, other than in Madre de Díos). Just voting for a party is pretty common, at least in the case where the party is well-known (whether or not it is “traditional”, whatever that means).

    In Madre de Díos, at least, it’s easy to tell how many people chose to issue a candidate vote[2]. It’s harder to tell in other districts because you don’t know how many people cast both preferential ballots. Anyway, the current results[3] show that from 55% to 77% of party voters cast a candidate vote (55% in the case of Gana Perú, which will undoubtedly win the seat; 77% in the case of Solidaridad Nacional).

    In the case of Gana Perú, the top two candidates received 4459 and 3393 votes, respectively, showing some sort of competition. In all the other parties, one candidate received the bulk of the vote. In the particular case of Perú Posible, which came second (but a long way back), the first candidate received 5943 votes, far in excess of the Humalista who won (“probably won”, I should say.)

    As it happens, the Perú Posible candidate received a somewhat similar vote in 2006, also running for Perú Posible, and won the 2001 election for a different (now defunct) party.

    So Madre de Díos does not present a clear picture either way; it shows both inter- and intra-party competition, in different parties and different years. This will be the first time in three elections that Madre de Díos will elect a representative from a major party, but the major parties do generally do reasonably well. (“Major” in the context of the election in question, I mean.)

    I believe that the preferential ballot does function as a way for party voters to express their preferences within their party, at least in some cases. In that sense, it could be perceived as a reasonable alternative to “primaries”, and I don’t know if anyone has ever suggested that primaries are centripetal.

    [1] By the way, my reading of the Finnish documentation is that in the case of Aland, lists are allowed to include up to four candidates; 14 is the limit for other districts where M<=14.

    [2] With one proviso: if someone fills in the preferential vote and doesn't fill in the party, even though the voting intention is obvious, that is probably counted as a spoiled ballot.

    [3] The current results are based on all 74% of polling stations where there were no apparent errors in the reported results, so they're probably pretty indicative.

  12. I propose the “Square Root Limited Vote”, where voters are given a number of votes equal to the square root of the number of seats in a state or region.

    Voters must use all preference votes and can give a candidate more than 1 vote.

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 )

Google+ photo

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

Connecting to %s