[MSS here: Thanks to RAC for the post below, and to all those out there who treated this and the previous post as open thread on these elections. There is a wealth of great information on Peru and Italy in the comments below.]
In Peru, as of this writing, ally of Hugo Chavez, Ollanta Humala (UPP) has secured first place, while former president Alán GarcÃa (Apra) has narrowly overtaken Lourdes Flores (UN) for second as counting continues. (updated coverage here)
Meanwhile, Keiko Sofia Fujimori has won a seat in the assembly.
Italy Update: After some projections showing Berlusconi’s center-right coalition with a narrow lower-house plurality, Prodi and the center-left appear to have captured control of the lower house by an extremely narrow plurality. Under the new electoral system (on which, see Federico Ferrara’s recent comment), this will produce a 340 seats — 55% — for the left coalition.
Final results in the two-day election that ended Monday showed Romano Prodi’s center-left coalition winning control in the Chamber of Deputies, the lower house of parliament, with 49.8 percent of the vote compared to 49.7 percent for Berlusconi’s conservatives. The winning coalition is automatically awarded 55 percent of the seats, or 340 seats, according to a new electoral law.
For a time Monday projections had predicted center-right control of both houses:
[The polling company] Nexus said its projection based on 44 percent of the vote gave Berlusconi’s centre-right alliance a wafer-thin 49.9 percent to 49.6 percent edge over challenger Romano Prodi in the lower house.
Initial reports, however, showed a center-left lead in exit polls and narrowly in the count.
Prodi’s alliance was set to win between 50 and 54 percent of the vote in both the lower and upper houses of parliament, giving it a working majority in the two chambers, a poll by the Nexus research institute said.
In Hungary, the incumbent center-left coalition leads after the first round of voting.
Analysts say the advantage of the two ruling parties in the first election since Hungary joined the European Union in 2004 was slim and could disappear in the decisive second round.
See Lewis Baston’s coverage of Hungary here
I take it you meant ‘340 seats, a majority of about 60′ (not sure of the exact number)
[this text is now clarified; thank you - ed]
Seed planted by Paul Davies — 11 April 2006 @ 04:22
For those who collect ballot papers (Wilf, for example); I found a PDF of the Peruvian ballot (for voters in Lima).
In real life, it’s quite large, about A3 I think.
The centre column varies from district to district, since not all parties run lists in every district. The first and last columns are identical throughout. The order of the parties is set by lottery prior to the election, and is the same on all ballots. It doesn’t appear that Javier or his party benefitted from their first place lottery draw. (It’s a shame; he’s a good guy, and has been a highly effective congressperson for a long time.)
The rules for valid ballots are intentionally very permissive: you can mark either the party logo or the candidate; for the congresspeople and andean parliamentarians, you can mark the logo, write one or two candidate numbers, or both. As long as the intention is clear, the vote counts.
I don’t believe many countries use the Double Optional Proportional Vote (DOPV), but it seems like a reasonably good system to me. Some peruvians don’t like it because they don’t see why they have to cast both their ballots for the same party. Probably more education needs to be done about what a proportional vote really means. Many politicians would prefer a closed-list system, arguing that DOPV is “too complicated”; in my view, this is completely self-serving and should be rejected by the electorate.
When the congressional results are available, perhaps next week, we will see the results. My prediction is that there will be 30-35 (of 120) women in Congress. That’s quite a bit under the list quote (40%) but it is considerably better than most countries manage with closed lists.
A couple of other things which will be notable from the congressional election:
1) UPP is predicted to win 43 seats over APRA’s 35, although the total popular vote of both parties is pretty similar, around 20% (which should translate into only 24 seats each).
2) UN is predicted to win 19 seats, with a popular vote of around 15%. AF, with a popular vote of 13.5%, is predicted to win 15 seats. These are pretty close to proportional to the vote.
3) Two smaller parties are expected to win 5 and 3 seats with votes of 7.6% and 4.1%. These are significantly underrepresented.
4) Peru Posible, the governing party, is not expected to reach the 4% threshold (but will not miss it by much). If it makes it, the above numbers will change.
5) Nearly 20% of voters will have voted for a party which did not receive any seats.
Several design issues are at work here:
1) Lima, with 35% of the vote, only has 29% of the seats (35 seats).
2) All the other districts are small. One has one seat; eight have two seats; four have three seats; two have four seats; seven have five seats; one has six seats and one has seven seats.
3) The D’Hondt system is used, which favours leading parties.
By and large, the UPP strongholds are small districts, while APRA strongholds are larger districts. The UN stronghold is Lima.
The 40% quota (each gender must have 40% representation on all lists) is maintained by having a minimum list size of three, even in districts with fewer than three seats. I don’t know how many countries do this, either, but it also seems like a good idea.
Seed planted by Rici — 11 April 2006 @ 13:37
Update: I just did the seat allocation based on current ONPE figures (which represent about 45% of the congressional vote, and not a random sample), and my count differs slightly from Apoyo’s. I get (with Apoyo in parentheses)
UPP: 43 (43)
Apra: 39 (35)
UN: 18 (19)
AF: 12 (15)
FC: 5 (5)
RN: 3 (3)
That actually fits my thoughts on how the various biases in the system ought to work out in practice, but the results are probably not representative enough to justify the time it took to do it.
Seed planted by Rici — 11 April 2006 @ 15:02
Thanks to those who have kept the orchard watered while I was away.
I missed some interesting elections! But I am catching up.
I will be back to tending the orchard myself either Wednesday afternoon or some time Thursday.
Seed planted by Professor Matthew Søberg Shugart — 11 April 2006 @ 17:55
Ollanta vs. Garcia. Now, that is a BAD choice!
I am curious to know if Rici (or others) believe that Flores would have beaten either of these in a runoff had she made it. (Maybe I should say if she makes it, as I am unclear on whether the current ranking might yet change.)
And could Italy have its two chambers divided? Last radio update I heard had the left pulling ahead in the Senate, but again, that may not be final.
Italy is a rare case of parliamentarism with confidence required before both chambers.
Seed planted by Professor Matthew Søberg Shugart — 11 April 2006 @ 17:58
Flores might still make it, but it seems unlikely. The preliminary results are slightly biased by the fact that the last ballots to be counted will be those cast outside of Peru. The exterior vote makes up 2.8% of Peru’s voting population, and is heavily biased towards the right.
It should be noted that rather more than 2.8% of Peruvians live outside of Peru. However, many of those are “undocumented” and are reluctant to register with the Peruvian embassy. (Some do, though, and there is no evidence I know of that the information is leaked. However, it’s easy to understand the reluctance.) On the other hand, Peruvian law does not allow mail-in or proxy votes, so in order to vote you have to physically present yourself at a polling booth; this might be quite a distance. So the turnout is generally lower in the exterior than it is in Peru.
Voting is mandatory in Peru; you are fined if you don’t vote. So the turnout is higher than it would otherwise be.
Having said all that, my back-of-the-envelope computation indicates that GarcÃa will win by something like 30,000 votes (0.2%).
Another point of information: the 2001 election was between an outsider, Alejandro Toledo, running on a platform of opposition to traditional politics. In fact, running in opposition to traditional politics is pretty well a tradition in Peru. In any event, the other two protagonists were the same, and the results were somewhat similar: Toledo received 36.5% of the vote, GarcÃa 25.8% and Flores 24.3%.
At the time, Toledo was trying his best to look left wing, although I couldn’t help noticing that his CV was not exactly consistent with that: he’s an economist who has been a consultant for the World Bank, for example. He also played the “poor boy from Cabana” card to the hilt. Of course, it’s true: he did grow up in unimaginable poverty. Humala grew up in a middle-class family. Both of them live in very nice parts of Lima. But that’s all beside the point.
So, in the second round in 2001, Toledo edged GarcÃa out by 53% to 47%. However, he was starting with a rather larger base than Humala. My guess is that unless something particularly outrageous happens in the next few weeks, GarcÃa will edge Humala out by roughly the same 53-47 majority. If I were a professional political analyst I might be less shy about making a prediction about a Peruvian election.
Pre-election polls showed Humala beating GarcÃa in a hypothetical second round, and Flores beating Humala; in both cases, the margins were slim. But that actually seems unlikely to me. It is very difficult, perhaps impossible, for the right to win a fair election in Perú.
Of course, Toledo managed it, but his popularity plunged so fast he must have had nosebleeds; his party, such as it is, has been completely devastated in this election. And that despite the fact that he managed to do exactly what he said he would do — reading between the lines — that is, implement a successful neoliberal program which kept Perú’s macroeconomy steadily growing. The problem is that the macroeconomic success was actually counter-productive: if you’re poor and you know some people are getting richer, you are much more inclined to notice that you are not yourself getting less poor.
Flores is just not in a position to tap into that resentment. Toledo could, and so could Fujimori, because both of them are visibly not part of the establishment (whether or not they are actually part of the establishment). So Flores (somewhat like Vargas Llosa in 1990) finds that the more she presents herself in public, the less support she has; her vote slipped steadily in the polls during the election campaign.
Flores does have one thing going for her: the gender gap. Polling shows an enormous gender gap between Flores and Humala, and I think that is what was driving the hypothetical second-round victory for her. However, by voting day the gap had diminished somewhat, and it probably wouldn’t resist another six weeks of campaigning. For one thing, Humala now has a caucus. By my count, 14 of Humala’s predicted 45 seats were won by women. That’s 13 out of 39 seats outside of Lima. By contrast, UN seems to have elected four women, two from Lima and two outside.
An even more interesting hypothetical question is: how well would Fujimori have done had he managed to get onto the ballot? His daughter Keiko will receive far and away the highest preferential vote in Lima (she has more than 80% of her party’s vote). But I don’t feel like going down that road…
I haven’t reverified my congressional calculation, but I did put names to the seats, and it seems like there will be 34 women elected, 11 in Lima (of 35) and 23 in the rest of the country (of 85). We’ll see in a week or two how close I came.
Seed planted by Rici — 11 April 2006 @ 19:04
[...] Matthew Shugart at Fruits and Votes also has a tally, but be sure to check out the comments section, for a guy named Rici, who seems to know up close what the balloting was like. One piece of useful information (I wish I had known earlier) is that the Lima voters are underrepresented in Peru’s legislature. That would naturally work against Flores’ party, because her base of support is in Lima. I wonder if it would also put her at an analogous disadvantage in the presidential race itself. This system of disadvantaging the capital’s votes sometimes turns up in different elections (e.g, Indonesia’s), and when you get a result like this, it’s a real shame. The item can be read here. [...]
Scion grafted by Publius Pundit - Blogging the democratic revolution — 11 April 2006 @ 19:08
Publius: the presidential vote is simply one person, one vote; congressional districts do not play a role. In fact, Lima is slightly overrepresented in presidential elections; for one thing, the effective turnout is slightly higher (I saw news footage of people who’d ridden three hours on horseback to vote at a rural poll; obviously, not everyone is going to do that) and for another thing there an estimated one million people who do not have identity cards, almost all of whom will be in remote parts of Perú.
Also, there’s no evidence I’ve seen that the foreign vote is going to Humala, although it’s hard to know since it isn’t showing up in the official counts yet. Apoyo estimated 60% of it would go to Flores, though, which is what early returns are showing.
The problem with counting the foreign vote is getting the poll reports physically to Lima; that takes a couple of days. (According to El Comercio, the early returns in the count were faxed in, but it turned out that many of them were illegible.) It also takes a couple of days to get reports from some of the remote parts of Perú.
Seed planted by Rici — 11 April 2006 @ 19:26
Thanks, Rici, very informative.
Seed planted by A.M. Mora y Leon — 11 April 2006 @ 19:50
Last radio update I heard had the left pulling ahead in the Senate, but again, that may not be final.
The Union did win - it took four of the six expatriate seats, and has a 158-156 edge in the Senate.
And if you think that’s close, it could easily have been closer. The right-wing parties actually won most of the vote in North America, but because they were running separately rather than as a coalition, the Union got that seat. If the right had its act together in North America, then that seat would have gone CDL. The Senate would then have resulted in a 157-157 tie, with the balance of power held by an independent party representing the interests of the Italian diaspora in South America.
Seed planted by Jonathan Edelstein — 12 April 2006 @ 04:25
[...] See the rest Fruits and Voteshere. A.M. Mora y Leon @ 7:37 am | [...]
Scion grafted by Publius Pundit - Blogging the democratic revolution — 12 April 2006 @ 06:41
The ABC has a good interview with the expatriate senator and deputy from the Asia-Pacific. They’ll be commuting from Melbourne to Rome. Their district extends from the west coast of Africa to the Chatham Islands and from Antarctica to Mongolia. The electoral division of Kalgoorlie (2 295 354 sq km ) has been well and truly deposed as the largest district on the planet.
Seed planted by Alan — 12 April 2006 @ 10:13
Another update on congressional seats.
Based on ONPE’s current figures, the current seat allocation should work out as follows:
UPP: 43 (Lima 6)
Apra: 38 (Lima 7)
UN: 18 (Lima 9)
AF: 13 (Lima
FC: 5 (Lima 3)
RN: 3 (Lima 2)
If RN doesn’t make the threshold, their 3 seats would be allocated to AF and Apra in Lima, and to UPP in Madre de Dios.
If Perú Posible makes the threshold, they would win 3 seats in Lima (none outside) and UPP, UN and RN would lose one seat each.
If Perú Posible makes the threshold and RN doesn’t, PP would win 3 seats in Lima, replacing the two RN seats and one UN seat; in addition, UPP would win the seat in Madre de Dios.
All of these results should be taken with a great deal of caution. Between a third and a half of the polls will be revised; the foreign vote is largely uncounted (which might give UN another seat in Lima, where all the foreign congressional votes count); and I may have made any number of errors.
If I’ve done this right, there are only three parties which are affected by the 4% threshold: RN and PP (as shown above) and Natale Amprimo’s Alianza para el Progresso. The last of these has no hope of reaching the threshold, but would have won one seat (in La Libertad) had there not been any threshold.
Seed planted by Rici — 13 April 2006 @ 08:10
Rici, thanks for the continuing updates on the congressional vote. How is the threshold applied? This is now a districted PR system, so a national threshold does not make sense to me. Is it applied in the districts? If so, it would be applicable only in Lima (because the next largest district has a magnitude of only 8, in which the “natural” threshold would be greater than 4%). Or does a party that gets less than 4% nationwide forfeit any seats it had the votes for in a specific district (which would shut not only small parties nationwide, but also any regionally concentrated parties)?
As Rici notes above, the average magnitude outside Lima is very small (3.5), and the overall representation is somewhat malapportioned. Lima has 35 seats, but if districts had magnitudes proportional to their population, it would have 42 (of 120). Of course, if Lima had 42 seats, then the remaining districts would be even smaller.
Unless Peru ignored its provinces in districting, its PR would be necessarily somewhat disproportional, with some combination of malapportionment (under-representing parties that are strong in the capital) and small magnitudes.
Before 2001, the unicameral assembly was elected in a single national district. Before 1993, the senate was a single national district, but the lower house was districted.
A better solution than either the current districting or the nationwide district would be to combine adjacent smaller districts and break Lima into three or more districts. Otherwise, Peru, like all countries with districted chambers and widely divergent population across administrative units that serve as districts, must tolerate severe magnitude variance effects.
(One could use some form of two-tier districting, though that would be difficult without also combining smaller districts, given that a national compensation tier would reduce the number of seats available for all the basic districts; it would also, however, make malapportionment or variance of the basic districts less relevant at the national level, because party biases would be corrected for by the compensation tier.)
Seed planted by Professor Matthew Søberg Shugart — 13 April 2006 @ 09:20
Matthew, here is my understanding. I’ve seen various interpretations in the press, but I believe some of them are wrong.
First, the seats are distributed according to the D’Hondt rule, with no threshold. If a party wins five seats in total, not all in the same district, or the party wins more than 4% of the total national vote, then they are included. On that basis, the distribution is done again with only the included parties.
Given the size of districts, for a minor party to win five non-Lima seats without achieving 4% of the vote would be tricky, but I suppose not impossible.
With the current vote, the quotient in Lima is about 2.4%. So a Lima-based party could win five seats in Lima with 12% of the Lima vote, but that would represent 4.2% of the national vote, so the party would reach the 4% threshold anyway.
Consequently, I think the five-seat rule is unlikely to have much effect.
For the next election, in 2011, the threshold is to be raised to 5%, as I understand it. That would have excluded both minor parties from the congress, I believe.
So with a 5% threshold, there would be five parties, and with no threshold at all, there would be eight (the last three of which would have a total of 6 seats). With the 4% threshold there, will probably be six parties, but it could be anywhere between five and seven.
Note that the foreign vote is included in Lima for congressional purposes. This is 2.78% of registered electors, which if it were a separate district would merit four seats. Lima actually has 33.98% of the registered electors; with the foreign vote, the total is 36.76%.
Also, parties which do not make the threshold will be deregistered (but not until after the municipal elections scheduled for November). To register a party, you need to collect signatures corresponding to 1% of the valid votes in the previous election; that would be about 120,000 signatures, I think. No-one is allowed to sign more than one petition, and there are currently 35 registered parties. (Draw your own conclusions
). So the real effect of the threshold may be visible in the next election, with far fewer parties participating.
Seed planted by Rici — 13 April 2006 @ 09:54
Matthew, I checked the law itself and I am now certain that my interpretation is correct. However, there seems to be a contradiction about party registration. (I’m not a lawyer, as they say, so I could be wrong.) In one place, it says that parties remain registered if they won a seat in the previous election. In another place, it says that they remain registered if they reached the threshold in the previous election.
In theory it’s possible to reach the 4% threshold without winning any seats.
The law that was passed last year actually says 5%, not 4%; there is a transition clause for the current elections. So the threshold for the next election will be 5% of the national vote, or 5% of the seats (6 seats) divided between at least two districts.
The law was challenged as unconsitutional and a judgement was delivered by the Constitutional Court. See the text of the judgement. I haven’t read it yet.
Seed planted by Rici — 13 April 2006 @ 10:25
There probably are magnitude variance effects, but I don’t know how severe they are. It is probably the case that UPP won several more seats than Apra (43 to 38 by my count), despite having fairly similar congressional votes (apparently: according to Apoyo UPP got 20.5% and Apra got 20.1%, but Apoyo seems to be consistently underestimating Apra’s vote by a little bit).
It’s interesting to look at the congressional vote in Lima, which I did get out of ONPE’s results:
UN: 20.3%
AF: 19.6%
Apra: 17.3%
UPP: 14.2%
FC: 8.5%
PP: 7.2%
RN: 4.8%
Partido Socialista: 1.4%
FIM: 1.3%
No other party reached 1%
That gives Apra one more Lima seat than UPP, but I don’t think that’s the magnitude variation in play. Outside of Lima’s, Apra’s strength is mostly in larger ridings. Apra came first in the two largest non-Lima districts (La Libertad where it should win five of seven seats, and Piura where I believe it will win three of six seats).
On the other hand, in almost all of the two-seat districts, UPP got the first seat and Apra got the second seat (in one of these ridings, UPP got both of them; in one, Apra narrowly came first). So that should mean that Apra’s votes were more effective.
Vote-splitting is very much in evidence in the above figures. Compare them with the presidential vote in Lima, also according to ONPE. (I don’t know how to format tables in a comment: the first figure is the presidential percentage in Lima; the second one in parentheses is the congressional figure, copied from above.)
UN: Lourdes: 34.3% ( 20.3% )
UPP: Humala: 23.7% ( 14.2% )
Apra: GarcÃa: 21.8% ( 17.3% )
AF: M. Chávez: 6.8% ( 19.6% )
FC: Paniagua: 5.7% ( 8.5% )
RN: Lay Sun: 4.5% ( 4.8% )
PS: Diez Canseco: 0.6% ( 1.4% )
PP ( 7.2% ) and FIM ( 1.3% ) did not run presidential candidates.
Seed planted by Rici — 13 April 2006 @ 12:56
[...] Rici: Another update on congressional seats. Based on ONPE’s current figures, the current seat allocation should work out as follows: UPP: 43 (Lima 6) Apra: 38 (Lima 7) UN: 18 (Lima 9)… [...]
Scion grafted by Fruits and Votes » Blog Archive » Peru: Possible alliances — 14 April 2006 @ 12:14