Electoral reform advances in Chile

Both houses of the Chilean legislature have passed legislation to increase the size of both chambers and scrap the current electoral system.

The assembly size changes are modest. Chile currently has one of the smallest first chambers, relative to population, of any democratic country (see graph). The planned change from 120 to 134 will redress this only slightly. The Senate will go from its current 38 members to 44.

The current electoral system for both houses tends to be referred to by the rather uninformative name, “binomial” system. It is open-list PR in two-seat districts, using the D’Hondt formula. D’Hondt is probably the most common PR formula, but of course it isn’t very proportional when your district magnitude is two. It does, however, tend to over-represent the second largest list, usually the conservatives, and this was its intention. (Aside: I met someone in the US government in about 1988 who had served as an advisor to the Chilean government, and he said the goal of the system then being designed was, in his words, “to screw the left”. Interesting that it is a conservative government that is finally moving ahead with changes.)

The article does not make clear what the new system will be. Presumably PR with higher and variable magnitude. Or maybe the same M=2 districts but with the additional seats being for a national compensation tier. If anyone has details on either the planned new system or the politics behind it, please comment.

The new system will not be in place until the 2017 elections.

(H/t to Chris, in an earlier thread on the topic; yes, they’ve been talking about this for a while.)

37 thoughts on “Electoral reform advances in Chile

  1. Apparently, the parties are still discussing the details of a new system.

    I’ve always wondered why a non-federal country like Chile needs so much malapportionment in both houses of parliament. The respective Gini-coefficients (census 2002) are 18,6%, rather high for a lower house, close to Venzuela and Argentina, and 43,5%, higher than the Senate of Canada but lower than the Ständerat of Switzerland.

    A small change in the Senate could make it beter apportioned (to 30,4%, close to South Africa):

    * give the Metropolitan Region double of what it has now (2 binominal districts -> 2 four-seaters or 4 binominals)

    * take half the deputation of the two southernmost regions (2 binominal districts -> one seat each or one binominal) The two northernmost regions are already coupled in one senatorial binominal district.

  2. Not so fast: the recently approved legislation was the result of an agreement between National Renewal (RN) – the more moderate partner in Chile’s ruling center-right coalition – and the opposition parties. However, RN’s coalition partner, the rightist Independent Democratic Union (UDI) remains staunchly opposed to the changes and has successfully resorted to delaying tactics, as reported by CNN Chile.

    Meanwhile, emol.com has further details on the electoral reform passed by the Senate last September. According to the article, the regions of Valparaíso, Maule, Biobío and La Araucanía would have their respective Senate constituencies merged – with each region having a four-seat constituency – while the Santiago Metropolitan Region would get four extra seats and become an eight-seat constituency. Meanwhile, the recently established region of Arica and Parinacota would become a two-seat constituency on its own right, and the binomial system would continue to be used there as well as in the other nine smaller regions, which would continue to elect two senators each (including the Atacama region, which the article omitted).

    In the Chamber of Deputies, some constituencies would be merged as well, while fourteen new seats would go to Valparaíso, Santiago and Concepción. In all, there would be nine constituencies with six seats each, eighteen with four seats, and four with two seats.

    Thus, the average Senate constituency would have just under three seats, while the average Chamber constituency would have just over four seats; the binomial system wouldn’t entirely go away, but its scope would be drastically reduced, particularly in the Chamber of Deputies.

    The article also notes that coalitions would be able to nominate twice as many candidates as there would be seats to be filled i.e. four candidates in two-seat constituencies, and so on.

    Finally, latercera.com notes in an article from last August on the electoral reform that the Congress building in Valparaíso has office space for at most 136 deputies.

  3. 136? That’s odd. There were 200 seats in Congress (50 Senators + 150 Deputies) before the 1973 coup. Did Pinochet bomb the Congressional offices along with the Presidential palace?

  4. Historically, the congress sat in Santiago. The seat of congress suffered some damage in 1973. Pinochet constructed a new legislative building in Valparaíso. The limits of the new building may well have been deliberate.

  5. I know that this may be somewhat off topic. How powerful is Chile’s Senate? Does a unitary country like Chile need a Senate? Is Senate abolition off the table?

    How power are Senates in the other non-Federal Latin American countries? Are they as powerful as the U.S, Australian, Japanese, Italian, or Romanian Senates?

  6. RN wants larger, even numbered constituencies because that will benefit them more. They are the smaller of the two right wing parties in most parts of the country, so going from 2 seats to 4 seats will likely see them take 1 of the 4 seats in those districts, whereas now they lose a lot of those seats to the UDI.

    They want even numbered constituencies because otherwise the left is likely to gain a significant advantage in many seats, particularly if the Nueva Mayoría coalition, uniting the center-left Concertación with the Communists and some smaller left-wing parties, holds together. Most 4 seat constituencies are likely to split 2 left and 2 right, whereas most 3-seaters would split 2-1 to the left.

  7. Sounds like Tullymandering … because: what if one of both ends up with a 3-1 spilt against them? (if >60% against <40% votes)

  8. Indeed, Bancki, it is a form of manipulation, of the same kind which led to the establishment of this ‘binomial’ system in the first place. With 2-member districts under D’Hondt, a party needs twice as many votes as the next party to get both seats. 4-member seats are certainly better (more proportional) than that, as a party only needs one-and-a-half as many votes to get a majority (3 seats, which also isn’t all seats).

  9. Thanks for the detail, Manuel!

    On whether the “binomial” system would fully go away or not, of course many districted list-PR systems have some 2-seat districts, meaning they have a “binomial” system in parts of the country, if one wants to call it that. The label has stuck in Chile only because it is unusual among districted PR systems in having all districts of the same magnitude (2).

    As for the Senate of Chile, as far as I know it has co-equal powers with the first chamber. Such symmetry of powers seems to be typical of bicameral systems in Latin America.

  10. MSS @10: ‘many districted list-PR systems have some 2-seat districts, meaning they have a “binomial” system in parts of the country’

    From memory, Spain, Portugal and Switzerland have List-PR districts with widely ranging magnitudes that include 2-seaters.

    The only cases I know of beside the Australian Territory Senators and formerly the Commons University seats with STV 2-seaters were (1986-1996) some of the constituencies for the House of Keys in the Isle of Man’s Tynwald Court: http://www.tinyurl.com/lmmuj4e.

    • There are eleven 2-seat districts in Uruguay, at least as of 2004. There are thirteen in Colombia (including the special district for Colombians abroad). Peru had 8 as of 2006. Paraguay used to have eight (of only 18 total) districts with M=2, but now has six.

      Clearly it is not a common occurrence, and of course Chile’s use of only M=2 is unique, as far as I know. I think it is also fair to say that 4-seat districts are not very common, yet it seems as if Chile will have a lot of these low, even-number, magnitudes if this reform passes.

      • I should have noted that Uruguay’s electoral system makes the 2-seat (or any M) districts work a little differently. Parties may present multiple lists, so while each individual list is “binomial”, and obviously at most two candidates can win, there can be many more options than under Chilean (or Peruvian, etc.) rules.

  11. In Spain almost each province starts with 2, and then the rest of the seats are apportioned in proportion to population. This leaves only one province with two seats (Soria).

  12. How discipline are the parties in Chile’s Congress? Are they as discipline or more so than U.S political parties? How discipline are they relative to party discipline in other European countries?

  13. M=2?

    France is currently redistrciting the ‘cantons’ used as electoral districts for their departmental councils. It was about time, but the novelty is that every canton will elect (by majority as usual in France) a ‘binôme’, a closed list of a woman and a man. (the number of cantons will always be uneven)

  14. I thought the same thing. But then I realized that if both representatives from each canton are elected as a pair, they would be likely to be expected to vote the same. If the number of cantons was even, and votes aren’t splitting, then ties would be likely.

  15. I think the binôme is actually an interesting idea for countries that insist on retaining single-member districts, although obviously PR with zippered lists solves many more electoral problems than any single-memebr system can.

    I promise to resist the temptation to start calling single-member districts mononomes, if I can.

    As I understand it France has a spectacular history of gerrymandering, especially for elections below the national level. Does anyone know why?

    • (My latest comment missed the boat and wasn’t transferred over here)

      Only the departmental councils are elected by SMD. Those SMD are named ‘canton’ and most of them date from 1801 (Napoleon I), only the fastest growing cantons got split (mostly after 1970) I wouldn’t call it gerrymandering (deliberately drawing new boundaries), but malapportionment in a passive sense : implicitly but deliberately sticking with old districts. This is a problem in almost every department.

      Regional and municipal councils are elected by PR with a huge bonus.

  16. Thanks, and sorry about that Bancki. That was why I shut down the comments once I knew we were moving. But I was afraid I might have missed a few in there.

  17. The Chilean Chamber of Deputies’ Constitutional Commission approved yesterday the idea of reforming the country’s binominal electoral system; La Tercera reports that once more RN joined forces with the center-left parties to vote in favor of the reform, leaving UDI alone in opposition to it.

    As it currently stands, the reform calls for expanding the Senate from 38 to 50 seats and the Chamber of Deputies from 120 to 155. For upper house elections, each one of Chile’s fifteen regions would be a constituency electing two to seven senators, with seven regions remaining two-seat constituencies. Meanwhile, for the lower house there would be 28 constituencies, each one of which would be entitled to a minimum of three seats; the remaining 71 seats would be distributed among the constituencies in proportion to the number of registered voters, but no constituency would be assigned more than nine seats. Finally, constituency seats in both houses would be distributed by the D’Hondt rule.

    Chile’s Chamber of Deputies has further information on the bill here.

    • Isn’t that upper house system quite unfair to whichever party has more rural strength? If there were just two parties, we’d expect the urban party to take 4-5 seats in the 7 seat urban districts, and still be able to split the rural 2-seaters.

      • But presumably the existing alliances will split to some degree, and there will be more than two viable lists in most districts.

        Still open lists?

      • Not at all, if higher district magnitude coincides with higher population desity, the Monroe-Rose variance effect described elsewhere on this blog will favor rural parties; added to this there will still be some malapportionemnt favoring rural districts.

      • Would Monroe-Rose apply even in the case of two-seat districts and a strong two-alliance system? I’d be quite surprised. It only really works when we can expect rural parties to gain an advantage in rural areas, and that’s not at all the case when M = 2.

    • Following up on the subject, earlier this month the Chamber of Deputies finally approved the bill to reform Chile’s binominal electoral system, with UDI being once more the sole party voting against it. La Tercera has more information here about the recently approved bill, which now goes to the Senate.

      • And last week the Chilean Senate’s Constitutional Commission gave its conditional backing to the reform of the binominal system – conditional because the number of Senate seats to be assigned to each region – especially the Santiago Metropolitan Region – has become a point of contention; La Tercera has the details here.

  18. Actually, the Santiago Metropolitan Region – which would become the sole seven-seat Senate region if the reform bill were to be passed as is – has been a highly competitive area in the last two congressional elections, as shown by aggregate regional results available on my website’s
    Chile page.

    As detailed on the text of the bill, the intent is to uphold the principles of equality among all fifteen regions of Chile and of the Senate as a chamber of territorial representation, which is not exactly the easiest thing to do when the largest region in the country accounts for almost two-fifths of the electorate. The result is a compromise whereby each region is allocated a minimum of two seats, for a total of 30 assigned seats, and the remaining twenty are distributed according to the number of registered voters. That said, it’s not one-person, one-vote by any stretch of the imagination.

    I’ve run simulations of the 2009 and 2013 elections under the proposed Senate system (using lower house results out of necessity, as Senate elections in Chile are staggered, with approximately half the body elected every four years), and the notional distribution of all fifty seats among coalition tickets would have favored the center-left over the center-right in both instances: 26-to-23 in 2009, and 30-to-18 in 2013; the Santiago region would have split 4-3 in both instances (in 2009 for the center-right, and in 2013 for the center-left).

    • Matthew, to answer your first question lists will remain open.

      I wouldn’t rule out splits in the existing alliances if the proposed system is ultimately adopted, but that said the average Senate and Chamber constituencies would increase to just over three and five seats, respectively. That doesn’t seem to me like much of an incentive for alliance partners to go their separate ways, especially under the D’Hondt rule.

      • Fair point, Manuel. I guess what I should have said is that existing third alliances would have a shot now to win a few more seats, and so the assumption of two-party (or two-alliance) competition should not be taken at face value. I believe the two big alliances’ combined strength has been eroding anyway, and a higher average magnitude would open up some seats for other alliances. Not a lot, though, as you suggest.

  19. If I’ve calculated it correctly (on the basis of the number of registered voters mentioned the bill), the Gini-index of malapportionment of the Senate goes from 43% to 38%… Still rather high for a non-federal country?

  20. Pingback: Chilean electoral reform | Fruits and Votes

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.