Yglesias on STV

US political blogger Matthew Yglesias has suggested single transferable vote as ” one solution to polarization” in the US Congress.

I would note that his specific suggestion that New York City could form a single 13-seat district might not be the best way to sell STV. But perhaps one should not quibble with such details, important though they are, at this point.

I did not look at many of the comments (55 at last check), but I did notice that the first comment advocates expanding the size of the House (as an alternative, but why pick just one of these?), and another makes the all-too-common mistake of conflating the increased district magnitude of PR with “at large” plurality (with reference to such a provision in the Puerto Rican legislature).

And at least one of the comments mentions the looming referendum on STV in British Columbia.

9 thoughts on “Yglesias on STV

  1. Don’t be too hard on the commenters. Some professional journalists – not amateur comboxers – have argued, in print, that Iraq should use single-seaters because these are – wait for it – more favourable to dispersed minorities like Iraq’s liberals/ democrats than a single nation-wide constituency is.

    See, eg, the Wall Street Journal, Larry Kaplan, and Larry Diamond. OK, you expect the WSJ – home of John Fund – to have odd ideas about how PR works, but The New Republic?! Hendrik Hertzberg, call your old office!

    Arguing for districts (single- or multi-seat) for Juan Cole’s reason – because these guarantee the Sunnis per capita communal representation even if they sulk and sit out the elections – now, that would make sense.

  2. New York City is actually entitled to just under 12 congressmen, remember the size of a Congressional district is going to something like 720,000 and the city’s population is just over 8 million.

    There is an interesting point here. You could implement proportional representation for the House of Representatives without amending the constitution. But no matter which version was used, the multimember districts would have to be based on states, since congressmen are apportioned to the states first.

    There are 12 states that have only one or two congressmen apportioned to them that would have to stick with single member districts in this situation. There are another 26 states with between 3 and 10 congressmen, and these select something like a third of the House. The thresholds for these states would be at least 10%, probably too high for a party list system (you would get really distorted votes to seats results). These states might as well use STV.

    In the larger states there are no regional or metropolitan area wide governments, with the large southern California counties the possible exception. There are no real natural divisions for multimember districts. I’ll bet if STV was implemented some large state would actually try for a twenty-plus statewide district unless the legislation specified an upper limit to the districts. You would see a 12 member New York City district because people are just too used to thinking in those terms.

    If the US were to use proportional representation, STV would be a better technical fit for the US federal system, with its large discrepancies in state populations, than the party list system. And minimum and maximum sizes for the STV districts would have to be provided for. Given the American political culture, there would inevitably be attempts to tullymander.

  3. The legislative council of New South Wales has 42 members with 21 MLCs elected every 4 years. The state is a single STV electorate for that election. It has not caused any severe problems, apart from the tablecloth sized ballot paper.

  4. The U.S House should be increased in size along with a change in electoral system. 435 members doesn’t represent 300 million anymore like it use.

    If the Wyoming rule is used, then Wyoming is the only state that would have 1 congress member using a majority system, but all the other small states can STV electing 2 members, and the rule can be establish. States that have more than 3 seats, must use multiple districts of 3, 4, and 5. The U.S Congress would increase to 669 members. Is that true?

    New York City can be broken up into Three 4 seat districts if it were given 12 members. Probably get more representation if Congress is increase.

    NY would probably get 20 seats. Just use 4 5 seat district or whatever fits using communities of interest. Each of the boroughs can be a district within reason and 10% deviation.

    However, any system of PR in the U.S needs to be tried at the state level to see if it will work before it can tried at the Federal Level.

    The Senate can use the Majoritarian Alternative like Australia uses in it’s lower house. So we basically followed Australian method reversed.

    [Links within F&V added by MSS]

  5. Does NSW allow above-the-line voting, and if so do most voters use that option? Does NSW elect dozens of other offices at the same time, as US voting jurisdictions tend to do?

  6. > “Does NSW allow above-the-line voting[?]”

    Yes – in fact, I think NSW had ticket-voting a few years before the Feds adopted it.

    > “and if so do most voters use that option?”

    Yes, it’s about 95% for the Senate and all mainland State upper houses with ticket-voting. The exceptions are (a) Tasmanians electing their federal Senators, where I believe it’s only 80-85%, due to use of rotated ballots for State elections; and (b) on occasions, ACT Senate elections – Canberran Territory elections also have rotated ballots and with only 2 Senate seats vacant, the addition of “above the line” squares tends to make voting marginally more, rather than less, confusing.

    Tasmania is the only State where a Senate election since 1948 (STV-PR introduced) has come close to electing a party’s candidates out of the order they appear on the Senate ticket. There was some talk in 1980 this might happen in Qld, after Premier Bjelke-Petersen (who was opposed to working mothers in other contexts) shoehorned his wife Flo into #1 spot on the Nationals’ ticket.

    Note that whereas Senate elections require all candidates to be numbered for a valid vote, NSW only requires 15 preferences for 21 seats. (Originally 10 preferences for 15 seats). Of course, this is still a complex task, and the habits formed by Senate ticket-voting carry over to State elections even when…

    > “Does NSW elect dozens of other offices at the same time, as US voting jurisdictions tend to do?”

    … Federal and State elections are always (by Federal Act) held on different days. Under State law, State and local elections are also always on different days. So a NSW voter gets a large, Senate-like “bedsheet ballot” with 100 candidates for 21 Legislative Council seats, and a short ballot with half a dozen candidates for one Legislative Assembly seat (1 preference suffices). Occasionally – 1978 and 1991, from memory – a State referendum as well, on one or two propositions, all initiated by the State lower house (ie, the Cabinet).

    Since I lived in Queensland and the ACT before moving to Canberra, it wasn’t until my late 30s that I voted for the first time in a State Upper House election.

  7. NSW does have above the line voting but it is different from the federal version. You get to express preferences among the party tickets. The federal ballot-paper lets you put in an X in the box for one party. The NSW ballot allows you to number the boxes for different parties as you please.

    I do not have any hard numbers on how many NSW electors vote for more than one preference among the party lists, but the raw figure of how many NSW electors vote above the line does not mean the same thing it means in a federal election.

    While I am not aware of a change in the ticket order there are several cases like former Premier of Tasmania Doug Lowe whose claim on the party leadership was greatly strengthened by his ability to draw 2 quotas in his own right.

    The US does have significant long-ballot problems. Those could be minimised by above the line voting, better ballot design and better public information. In Scotland they found there were significant difficulties asking electors to vote by MPP for the Scottish Parliament and STV for local councils.

    I’ve always thought that is an excellent argument for STV because neither MMP nor list systems can be used for electing single chief executives.

  8. A comment on terminology, in particular on clarifying the term “district magnitude”.

    A far as I know this term was coined by Douglas W Rae in his influential 1971 book The Political Consequences of Electoral Laws http://www.tinyurl.com/p9bsfo8, as a useful way to avoid the confusion engendered by other terms such as “district size” (which could be taken to mean population numbers or geographical area).

    I don’t have a copy of Rae before me, so apologies to his memory if he does address this and I’ve forgotten it… I’m not sure whether, in his usage, the (say) US Senate has a district magnitude of 1, or of 2, or even (arguably) of 0.66667 (since 2 seats are filled over a three-election rotation cycle). Can anyone remember, clarify or assist?

    On the off-chance that Rae did not nail down this particular point, I would like to propose the following stipulative usage:

    1. A district’s “delegation” is its total number of seats/ representatives elected over one complete electoral cycle (whether a single-election cycle as for pretty much every non-Argentinean lower house, or a multiple-election cycle as for most elective upper houses outside Italy, Spain, Belgium, Victoria, WA, and, IIRC, a couple of US States).

    2. A district’s “magnitude” is the number of seats being filled at a particular election.

    3. A district’s “quota” is [100% of votes x number of separate votes a voter may cast], divided by [district magnitude + number of separate votes a voter may cast]. Thus, if a winner-take-all system is used, the quota is a 50% majority – whether [100% x 1 vote] divided by [1 seat + 1 vote] in an SMD, or [100% x 10 votes] divided by [10 seats + 10 votes] for the typical US State in the Electoral College. With PR or SNTV, of course, it’s a Droop quota. For Limited Vote (more than one vote, but fewer votes than there are seats), say 3 votes for 5 seats, it’d be [300% divided by 8, or 37.5%], as writers like Enid Lakeman and Andrew McLaren Carstairs have pointed out.

    Each quota polled by a party usually assures it (assuming either preferences, lists or prudent vote-management) as many seats as the number of separate votes a voter may cast (or all remaining seats, whichever is less). Thus, with 3 votes for 5 seats, the larger party gets 3 seats, and the second-largest (if over 37.5%) gets the 2 remaining.

    Following this usage, then, if a polity’s constitution were to specify that “Each State elects 10 Senators, 5 at each election in rotation, and no voter may vote for more than 3 candidates on one ballot”, then each State would have a delegation of 10, a district magnitude of 5, and a quota of 37.5%.

    [I have commented elsewhere on this blog that it is misleading to say, eg, “PR and SNTV give you only one vote, but MNTV gives you a generous five votes, for five seats” since “a single vote” in an STV or List PR system really means unrestricted, self-executing cumulation and dispersion, rather than pre-set percentages. But “five votes for five seats” is common usage so I’m going with it for the sake of argument here.]

    4. And finally, a district’s “threshold” means its quota, OR such higher percentage (if any) as the electoral laws may set. Thus, the German Bundestag has a quota of about 0.16% but a threshold of 5%.

  9. … The downside of this taxonomy (proud though I am of it) is that it leaves “district magnitude”, Rae’s most influential contribution to the literature, attached to the “least interesting” of these four measures.

    That is, political scientists would be interested in knowing how large a district’s total delegation is for the purposes of tracking pork, and psephologists would be interested in knowing what a district’s quota or threshold is so they can assess a minor party’s chances of winning any seats. But on its own, knowing that (eg) Illinois and Nicosia are both 20-seat districts (for the US Electoral College and the Cypriot House of Representatives respectively) isn’t very useful if you don’t know that one uses closed-list MNTV and the other votes by open-list PR – two systems having close to diametrically-opposed political consequences.

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.