What is “proportional” about robbing a majority winner of his majority?

Sent to The Los Angeles Times today, in response to an article that inaccurately refers to Republican proposals for several state’s allocation of electoral votes as “proportional systems”:

The LA Times refers today (Jan. 27) to Republican proposals in several states to replace statewide winner-take-all allocation of presidential electors with “a proportional system”.

These proposals are NOT proportional; they are still winner-take-all, but in each congressional district. As noted elsewhere in the article, had a district plan been in effect in 2012 Mitt Romney might have won 9 of Virginia’s 13 electors.

This means Barack Obama, who won 51.2% of the statewide vote, would have had barely 30% of the electors! This does not meet any standard of proportionality.

Even the House of Representatives, which is obviously allocated based on congressional districts, is not proportional: Democrats won the most House votes in 2012, but Republicans won a majority of seats.

Proportional representation is used by most of the world’s democracies. It produces allocations of political power that mirror how people actually vote. By contrast, the Republicans are proposing a house of mirrors to distort the vote for partisan advantage.

24 thoughts on “What is “proportional” about robbing a majority winner of his majority?

  1. These are fantastic points about something that should not be misunderstood. Proportional is proportional.

    However, we would be remiss if it was not mentioned that part of this is attributable to the Republican rebranding of the proportional concept as codified in the national party’s 2012 delegate selection rules. State-level delegate selection plans were deemed proportional if they contained some element of what the RNC defined as proportional. Similar congressional district delegate allocations were defined as proportional. In many, if not most, ways, anything that was not winner-take-all was considered proportional.

    The press (as are Republican state legislators pushing these plans) — and I talked to Paul West about this stuff in the midst of the primaries last year — is now extending this reasoning to these electoral college proposals.

    I spent way too much time at FHQ last year debunking the myth of proportionality. That message fell on a great many deaf ears.

  2. Even generally intelligent people such as Wonk Blog on nytimes.com make the same mistake. It’s infuriating.

  3. The Republicans have introduced the especially greedy twist of giving the two “senatorial” votes to the winner of the most CDs, rather than to the winner of the popular vote (as in NE & ME), raising the hypothetical 2012 split from 7:6 to 9:4, vs the actual 0:13.

  4. Josh, thanks for commenting. I do remember your posts on this theme, and should have made the connection. I guess this is the death tax of electoral systems.

    Infuriating indeed!

    I did indeed notice that some plans allocate the two statewide electors to the ticket that won the most districts. Speaking of infuriating. I think I have seen something like that only in the old South Korean electoral system that the dictatorship set up. It was a mixed-member system, but the list (don’t called it “proportional”!) tier was allocated based on districts won, not votes.

    I guess the Republicans have become the Chun Doo-hwan of American politics. Chun was, after all, a contemporary of Ronald Reagan.

  5. For years, I have seen Maine and Nebraska’s allocation rules described in the media as “proportional”, so this is not new.

    I recall a letter (which might have been published, but evidently not influential) that I wrote to the Times after the 1992 election, in which I pointed out that a 4-0 allocation of Maine’s electors was somehow not very proportional when there were three candidates in that state with over 30% of the vote.

    Moreover, the greatest split that Maine’s ever could allow is 3-1-0, which also pretty clearly fails the proportionality test.

  6. Of course the scheme in question is not proportional – nowhere in the article was it described as “complex” or “complicated”…

    Now, seriously, this is history repeating itself, the other way around: back in the early 1990s, the Democrats – which at the time had lost three presidential elections in a row – proposed a nearly identical plan, specifically in Florida if memory serves me well. Predictably, the Republicans were outraged, not least because the selective implementation of such plans.

    (Incidentally, this is also an issue in the current debate that the linked article doesn’t explain very well: while a nationwide switch to CD allocation of electoral votes would in all likelihood favor the GOP, the implementation of such plans only in selected states run by Republicans but won by Obama would be far worse.)

    At any rate, the Democratic proposals of two decades ago came to nothing, and Bill Clinton’s subsequent victory in 1992 under the existing system was in all likelihood the last nail in the coffin. I don’t know if the current GOP proposals will get any traction this time around, but I must say I find it quite fascinating that the Republicans – traditionally the staunchest defenders of the existing electoral arrangements – would be willing to toss them out in favor of a “if you can’t win a majority, manufacture one” approach: sounds like some of them are getting a bit desperate.

    That said, some of the saner minds in the GOP appear to realize that the proposals in question could backfire on them: if, say, Virginia were to adopt a variation of the Maine-Nebraska system and then the state switched back to the Republican column at the presidential level, they could end up giving away to the Democrats electoral votes the latter party would have never won in the first place…and which could well decide the outcome of a presidential election.

  7. Actually, Maine could theoretically allow a 2-1-1 split. In a theoretical 100 popular voter contest, ME-1 could vote Bush 22, Clinton 20, and Perot 8, giving Bush one electoral vote, ME-2 could go Perot 21, Clinton 20, and Bush 9, giving Perot the second electoral vote, and statewide would thus be Clinton 40, Bush 31, and Perot 29, giving Clinton the two remaining electoral votes.

    Because the two states are small, and because we rarely have three candidates strong enough to win a cong district, the statewide plurality winner has won a majority of both states’s votes every time. The statewide plurality winner must win at least half of ME’s votes, and the worst for the statewide winner in NE would be 2-2-1 or 2-1-1-1.

    5 is the most electoral votes a state can have without the theoretical possibility that a state could give more electoral votes to the second place candidate statewide than to the first place candidate.

    I do think that if any state adopted this plan with more than 5 EVs, and even perhaps for 5 or fewer EVs, the plan would be unconstitutional as a violation of the 14th Amendment. Because different numbers of voters vote in each district, it would violate one man one vote for a state to distribute EVs in a statewide race in sub state districts

  8. Why would it be unconstitutional to distribute electoral votes in the exact same manner as congressional votes?

    Mind you, I would actually contend quite strongly that a more rational jurisprudence would disallow congressional gerrymandering.

  9. Honestly I think the current Congressional election system is unconstitutional.

    I think the current limit of 435 seats is not just unfair, but an Equal Protection violation as well. The seats must be apportioned to States by the entire free population according to the Constitution, but I think an arbitrary seat limit that means some states have 500k people per rep while others have over a million per rep is illegal. Congress may set the apportionment method and number of seats by law, and obviously there must be differences in reps per capita to an extent, but such a disparity as exists now should be found to be unconstitutional. I would love to see a state that gets screwed, like Montana, sue over apportionment, but I don’t know if any ever have.

    I would also say the current requirement of equal total population per seat within a state is illegal, though that is a result of a SCOTUS ruling, Westbury v. Sims. Apportionment constitutionally is by population by state, not by district. Because districts have different numbers of eligible voters, registered voters and people who actually vote, it means a vote in one district does not have equal weight to that of another, and should be seen as an Equal Protection violation as well. The GOP’s Texas congressional plan was denied preclearance in federal court last year for drawing a ‘Latino opportunity’ seat by putting the Latino precincts with the lowest historical turnout in the district. I would think the principle of actual equal voting power, and not just potential voters (and even non citizens and persons deprived of their civil rights as the current system includes), must apply to all seats and not just minority opportunity seats. Whether the potential remedy would mean distributing congressmen by statewide votes (either proportionally or, God forbid, winner take all), or whether a distracted plan (either MMD or SMD) could be used that guaranteed, to a reasonable extent, equal voting power, I don’t know.

  10. Frank Luntz probably suggested that Republicans use “proportional” in this discussion.

    After all, unless they use this “proportional” system, they may never get a chance to repeal the “death tax.”

  11. As for why I think that, even assuming that the current House system is Constitutional, the CD+2 electoral model is unconstitutional, I think there are two valid arguments.

    One is that each House election is an independent election, not a statewide election as the election of a state’s electors is. By giving voters in certain arbitrary subdivisions the state more weight in the statewide vote than others, that violates the equal protection clause.

    Even assuming that the CDs are voting for individual electors and not a slate or.portions of a slate associated with a statewide ticket, I think the setup, including that of ME and NE, is in violation of Westbury v. Sims. It ruled that states couldn’t have SMDs and statewide at large districts on top of them. That was because states weren’t redrawing districts and thus they had vastly different populations, but I think it prohibited the SMD+at-large set up overall (which, as an aside, may mean MMP is unconstitutional as well). Since EVs are constitutionally assigned to the state as well, I would think the CD+2 method for EVs also violates the Westbury ruling.

  12. I think the only way to ensure equal protection in a MMD, which all states are in the EC, is to do so proportionally statewide.

    It might be constitutional, if people are ruled to actually be voting for individual electors, for a state to draw one SMD per elector, so 13 in VA’s case. These would have to pass VRA muster, but could be partisanly gerrymandering under current law. This method was used in the very distant past. Let’s hope no GOP operative reads F&V and picks up on the idea.

    It also is constitutional for state legislatures to pick the electors without popular consultation, but
    even our apathetic public would flip at that.

  13. I like the idea of having the Congressional Districts allocating 1 elector each and have the 2 at-large electors be used as a “top-up” to try and give a more proportional approach.

    But my definite favorite is allocating all electoral votes proportionally.

  14. One minor quibble: if a truly proportional system is used (ie, the “complicated” or “European” type of PR, as in “if your slate polls 25% of the votes, it wins 5 of the 20 delegates,” not the “simple” or “GOP” type of PR, ie, “East Dakota-XII snakes around the Southwestern Highway for 210 miles, then assumes the shape of a Vermicious Knid in mating season…”), then that too can “rob a majority winner of their majority” in the sense that, say, 51% to 49% might produce an even division of 10 delegates to 10, even if D’Hondt is used and a fortiori most other PR quota or divisor systems.

    Perhaps more accurate in future to talk about “redistributing” something that the popular-vote winner had rightfully worked for – this might get under the skin of the typical Republican legislator. Just saying “denying [them] a majority is broader” and can sound from a distance like standard Madisonian checks ‘n’ balances.

  15. MSS: the old South Korean system you mentioned reminds me of the way special women’s seats are distributed in Pakistan – parties get additional seats distributed according to the number of constituency seats they win.

  16. Yes, “proportional systems” as actually practiced in the USA do sometimes deny a majority of seats (i.e. convention delegates) to a candidate who won a majority of votes–at the district level. The issue is largely a matter of widespread use, in Democratic presidential primaries, of small and even-numbered district magnitudes, and also of apparent allergy to the Jefferson method (also know by a funny foreign name, D’Hondt).

    In fact, one of the other articles I read about the congressional-district proposals for some states’ electors quoted a Republican operative referencing the Obama-Clinton race in defense of the proposal. It claimed that Obama won because of a “proportional” system. Of course, Obama actually won because he beat Clinton in the accumulated popular vote. The proportional systems mirrored that popular-vote win in the aggregate, majority denials in some individual districts notwithstanding.

    As for Maine, Chris (#7) is correct that it is mathematically feasible for the state to split other than 3-1; however, politically, that is almost impossible.

    Yes, JD, the Pakistani system for women’s representation bears a similarity to the old South Korean system. So I guess we know the models of democracy preferred by some Republicans come from such bastions of democracy as Pakistan and 1980s South Korea.

  17. I think a Maine 2-1-1 split is unlikely, but far from impossible. I used Bush, Clinton, and Perot in the example because Perot came second in ME in 92 (though my fictional example had him third). The state is also refreshingly unorthodox in its two-party politics. Until 2012 they had two moderate Republican senators, who won with huge majorities, while at the same time electing Democratic Reps. In 2010, they had a three way gubernatorial race, with a Tea Party Republican winning due to a split in the center and left vote. In 2012, Romney won the GOP caucuses in ME-1, while Ron Paul dominated in ME-2, and in the general centrist independent Angus King won the Senate race. Redistrictinv has made the.1st CD quite a bit more left wing than the 2nd.

    All of this put together means that the right 3rd party candidate has a very good chance of at least 1 EV in Maine. Either a libertarian like Ron Paul or Jesse Ventura or a moderate like Jon Huntsman Jr, or especially a centrist Mainer (any of King, Collins or Snowe) would Maine very interesting, especially of the GOP nomineewas very right wing.

  18. I hate to again comment on the Chilean system, but really we should call it the Pinochet system.

    This is a system specifically designed by one of South America’s harshest dictatorships to entrench its power. The system remains in force because, despite controlling a majority of elected seats in the congress, the Concertación has been consistently denied, by this very system, a sufficient majority to change it. It perhaps has a place on a political science blog as how not to design an electoral system, no more.

  19. All else being equal, if a State is majority-GOP enough for the Republicans to control both Houses and the Governorship, you would assume it would also consistently vote GOP in Presidential elections, in which case the “unit rule” would favour them more than districting.

    This sort of ploy only works in States where the majority in a presidential election is likely to be different from the majority in the State legislature. Reynolds v Sims notwithstanding, this can happen for various reasons. Two that come to mind are where either or both:

    (a) district boundaries favour one party (by accident or design) to win a majority of State legislative seats: whereas the unit rule operates at-large.

    (b) a presidential candidate is personally popular and attracts voters who might otherwise vote for the opposing party (Reagan with blue-collar Democrats) or not vote at all (Obama with African-Americans).

  20. It was a mix of the two: B allowed the GOP to take majorities in 2010, which gave them control of redistricting; then A allowed them to hold onto the state legislatures in 2012, as well as the congressional delegations.

    They’re trying to act now because in most of those states, a continued GOP “trifecta” of both houses of the legislature and the governorship is unlikely. The Democrats need to only take one chamber or the governorship to block the GOP agenda at the state level. Indeed, given the 20-20 split of the Senate in Virginia (due, in part, to Democratic gerrymandering), the GOP’s “trifecta” was so ephemeral they had a single day on which to act, inauguration day.

    It remains to be seen how well the GOPmanders will stand up over time; given the fact that the GOP districts are not as safe as Democratic districts, I think that in many of those states, the Democrats will make steady gains in the legislature; whether the GOP majorities will hold up to 2020, when they would have a chance to shore up their majority again, is a major question.

  21. A third factor might be that some US States hold State and local elections in odd-numbered years (NYC and NJ, from memory). Given that (a) political opinion can shift sharply within two years (1964-66, 2004-06, 2008-10) and that (b) the subset of the electorate who vote in “off-year” elections may be different from the subset who are drawn to the polls by the quadrennial presidential race (which dovetails in with my second point above), this might account for the principal/ agent slippage in these cases.

    (I often catch myself and other Australians forgetting how much compulsory voting changes these equations. 96% of the total adult citizenry always vote, and those who don’t vote, don’t vary according to the presidential/ off-year difference as noted above… although the turnout for local government, which is voluntary in some States, does consistently tend to be lower.)

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 )

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