Filling vacancies in OLPR

A proposal in Hong Kong would change the method of filling inter-election vacancies, reports David at Ahwa Talk. The electoral system for the Legislative Council is open-list PR, and currently by-elections are used to fill vacancies.

Most (all?) jurisdictions with OLPR systems simply take the next highest-ranking unelected candidate from the list of the departing member.

However, the Hong Kong proposal would take “The first candidate who has not yet been elected in the list with the largest number of remainder votes in the preceding general election.”

Average district magnitude is around five or six. David comments:

Maybe [proponents of this reform] assume that the individual candidate who captured that seat was a big reason for the party’s success… According to Carey and Shugart, low district magnitude in open list PR decreases the incentive for a candidate to cultivate a personal vote. In contrast, it is in high magnitude, open list PR where candidate preference matters more. This is because in a larger list, candidates have a stronger incentive to distinguish themselves from their fellow list members. Ultimately, we don’t really know why any individual is voting the way they are, but I think the Hong Kong government’s assumption requires more explanation.

I agree with David’s take on what the underlying assumption must be: that candidate reputation trumps party in Hong Kong. Indeed, the proposed method decreases the prominence of parties in that it departs from the normal working of OLPR by which is the list-PR component that first divides up the available seats, and only afterward that the “open” part comes into play in determining who gets those seats.

A potential benefit of the proposal, however, is that it should reduce the incentive of parties to rotate some of their legislators between elections. Doing so is common in OLPR systems–elsewhere (I do not know about Hong Kong)–and undermines the connection of elected legislators to the electorate. Under the Hong Kong proposal, a party would often forfeit the seat if it sought to swap out a member.

As for the Carey-Shugart (1995, Electoral Studies) we only claim that low-M OLPR places less premium on cultivating a personal reputation than does higher-M OLPR. The story is seen from the competing candidates’ point of view. From the voters’ point of view, however, smaller magnitudes and shorter lists undoubtedly increase the visibility of those who are elected, who win with greater shares of their party’s votes. I actually think this method for filling vacancies makes more sense for smaller district magnitudes than it would for larger. Whether it makes more sense than the usual party-centric way is an open question, and one that might not have a clear answer.

9 thoughts on “Filling vacancies in OLPR

  1. So this would be the equivalent of, say, using D’Hondt or St-Lague divisors to assign every candidate a “comparison number” (which would initially ensure proportionality across party lists)… but then filling a vacancy, not by the next-highest candidate from the same list, but from the highest candidate from the original list?

    SNTV seems to have taken root the most deeply in Asian democracies (Japan, Taiwan, the former Republic of China, and IIRC South Korea and South Vietnam at times). This proposal seems to be treating open-list PR as a form of SNTV (although, ironically, I believe most SNTV systems fill vacancies through by-elections).

  2. “Indeed, the proposed method decreases the prominence of parties”

    Could this in fact be the intention of the not exactly completely democratic Hong Kong authorities? It could be easier for a more or less autocratic government to co-opt independent MPs one-by-one, than a well-organized opposition party. So an electoral system with more focus on individual candidates is perhaps less scary to the Chinese government than multiparty democracy is.

    And by following the links on the Ahwa Talk piece I found this, which indicates that the motivation for abolishing by-elections is related to struggles over the future political system in Hong Kong:

  3. The comparison to SNTV is certainly apt, since the Hare/largest remainders method in Hong Kong offers no incentive for parties to combine other than into medium-strength lists. Why waste votes filling a whole quota when with any luck these may elect two members on a remainder basis? That is not to say that both the pro-democracy and pro-Beijing blocs are not genuinely divided internally in terms of ideologies, interests or strategies, but if there were any electoral benefits to joining forces many parties certainly would.

    The proposal to abolish by-elections is surely meant to prevent what happened last year, when five left-leaning liberals and social democrats – one for each constituency – resigned simultaneously to force a quasi-referendum on democratisation. Now, this did not go very well for them since the moderate pro-democracy parties disagreed with this strategy and no government-backed candidates would stand against them. As a result turnout was low and the whole thing apparently seemed wasteful to many people. All the resigned members were returned with very large majorities though.

    It should be noted that direct elections now only apply to 30 of 60 legislative councillors. The other half is elected on a restricted franchise from “functional constituencies” representing various professional sectors (all are single-member seats except for a three-member Labour constituency). The pro-Beijing parties get around 40% of the general election vote, but achieve a slight majority through these functional seats (I believe there are various super-majority requirements within the Council, however).

    The Chief Executive as well as delegates to the Chinese People’s Congress are elected by a more select group of people, appointed or elected through various means.

    Wikipedia says that after the “referendum” the moderate Democratic Party cut a deal with Beijing. The Council will be increased to 70 at the 2012 election, adding five more members to the regular elected tier, and five more functional constituency members to represent the 18 district councils. Crucially, while these latter five will be nominated by and from the local councillors, they will be elected at the general election by all voters who do not otherwise possess a vote for functional members, using jurisdiction-wide PR. There were some other aspects to the reform package as well – which passed the Council by about a three-fourths vote but left the pro-democracy camp badly divided. It does not seem to put the pro-Beijing majority in jeopardy.

  4. This sounds like the David Davis manoeuvre – that’s David Davis the rightful Leader of the British Conservative Party, I mean, not David Davis the Victorian State Health Minister: ie, resigning so as to trigger a by-election which the resigning official can use as a de facto mid-term referendum on some controversial issue of policy: in Davis’ case, on 42-day (up from 28-day) detention without charge of suspected terrorists.

    This tactic has the advantage that an official who wants an immediate public vote on a particular issue has to put their money where their mouth is and risk their own office to obtain a poll: whereas, eg, Rep Darryl Issa didn’t need to forfeit his seat to push for a recall against Gray Davis.

    It is also good for political standards to see officials resigning over matters of principle – eg, Phil Gramm voluntarily submitting to a by-election after he switched parties in 1982-83. There were calls for Cheryl Kernot to do this in 1997 when she switched from the Australian Democrats (ADs) to the Labor Party, but since she was a Senator, there would be no by-election – instead the Parliament of Queensland would appoint a replacement Senator, with her former party at the time of election (the ADs) enjoying a Constitutional veto over anyone so appointed. I believe that in New Zealand, a by-election within the last 6 months of Parliament’s term can be waived by a 75 per cent vote of the House.

    Three main disadvantages of using voluntary by-elections as surrogate referenda:

    First, each by-election covers only a single electorate, so not all voters get a say on the issue. However, if a PR system is used where by-elections are required (Ireland) or allowed (Tasmania), then a small percentage of all MPs could force a “mini-general election” across a large proportion of the populace. In Tasmania, the Greens could bring 100% of all voters to the polls if their MHA in each of the five electorates were to resign. In Ireland, a dozen resignations in the right DÁIL constituencies (the largest four- and five-seaters) could bring nearly one-third of the total electorate to the polls again.

    Second, as with any election using live candidates (that’s any “election” in the non-US sense, ie, excluding referenda), the verdict can be clouded by personalities. (In 2005, Davis was the plurality choice for party leader by MPs who nonetheless voted for 42-day detention). Having said that, it would be naïve to assume that referendum verdicts aren’t clouded by personalities either. [2] And one could argue that it is better to vote for concrete officials with a non-legally-precise mandate than for a legally-precise-mandate which the actually existing officials may have no incentive to carry out.

    The third problem is cost. Wikipedia notes that “Culture Secretary Andrew Burnham called on Davis to fund the cost of the by-election to the taxpayer, estimated at £80,000, from his own pocket.” Whether one interprets the action as heroic self-sacrifice or a narcissistic drain on the public fisc probably depends on one’s opinion of the politician concerned. I note that Australian voters generally resent an MP resigning in mid-term and triggering a by-election.


    [1] The main lesson to be gleaned from the UK AV referendum of 5 May 2011 is that 68% of British voters really, really dislike Nick Clegg – 34% of them because minor party leaders like Clegg get to boss the whole country around when there’s a hung Parliament, and the other 34% because Clegg has rolled over for his boss Cameron and supinely allowed the Tories to impose their vicious spending cuts without a peep of protest.

  5. This by-election-as-referendum concept is interesting. But in Canada at least, the PM gets to schedule by-elections, and has as much as six months to decide on a date. If an opposition MP resigns to force a by-election, the “mini-referendum” might have to wait until the issue isn’t terribly salient anymore.

    Are other countries with by-elections much more strict about the timing?

  6. Tom said: “In Ireland, a dozen resignations in the right DÁIL constituencies (the largest four- and five-seaters) could bring nearly one-third of the total electorate to the polls again.”

    This actually happened in 1925 in Ireland. Nine deputies of the ruling Cumann na nGaedhal party broke away to form a “National Group”, dissatisfied with the Governments lack of nationalist fervour, as well as their dismissal of insubordinate army officer after the misnamed “Army Mutiny”. Unfortunately all of them lost their seats, and I don’t think such a manouvre has been attempted in the South again.

    In the North Unionist MPs resigned their seats in Westminister en masse in 1986, attempting to provoke a province wide referedum on the Hillsborough agreement between the British and Irish governments. All but one of the resigning MPs were returned, but the agreement stood.

  7. First of all, I agree with the overall claim that the new system is more consistent with the principles of a PR system. The old system was indeed a majoritarian way of dealing with things which is quite inconsistent with the overall PR system. Now, regarding the argumentation undergirding the reform, I’m a bit less negative than the author of the original Ahwa talk article.

    To begin with, (and in line with the distinction between the candidate’s and the voter’s point of view) I’m not sure that the distinction between small and large districts is that important. After all, in general, candidates in open-list systems have incentives to cultivate a personal vote (more so than in flexible or closed-list systems). Admittedly candidates in small districts may have less incentives than candidates in large districts, but they nevertheless have substantial incentives to cultivate a personal vote as it is an open-list system. The question then is: how big is the difference between large and small districts? …And does it matter to the voter?

    Ultimately I guess this is an empirical question: candidate campaigning is just one factor determining who and why voters vote the way they do (media coverage is another – there’s plenty of candidates in large districts that campaign but are ignored by the media while this is less the case in smaller districts). The author of the Ahwa article feels that we don’t really know why people vote the way they do. I’m not sure if this is the case. At least the Dutch election surveys (flexible list) ask questions about whether they vote for a candidate or a party (and why so). If such data are available it is probably better to use them rather than to rely on fairly ‘distant’ candidate incentives.

    Additionally it may well be that voters have embedded preferences: they may have a preference for a given candidate of a certain party. In such a case voters feel it matters which candidate gets a party seat. When voters have embedded preferences, picking the next candidate based on preference votes may well be in line with the reasoning of the Hong Kong government spokesman.

    Now if we really want to design a system that accommodates for all this we should probably use a variant of the Belgian system which has a separate list for candidates who replace vacant seats. One could for instance think of a system with such a list whereby voters have to cast two votes: one for the effective candidates and one for the reserve candidates. But this is obviously purely a theoretical exercise. And perhaps an attempt to show that the Belgian system isn’t that bad after all…

    (PS this post is originally from a forum on electoral reform. it is slightly changed to fit this discussion)

  8. Also not strictly on-topic, now that democrats in Hong Kong seem to be losing the battle over an open election of the executive, why don’t they shift their attention to the legislature (functional constituencies : remove them or move them to a less powerful 2nd chamber) and to executive – legislative relations (parliamentarism?)

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 )

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.