A Louisiana Senate panel has sent on its way to needed passage a bill emulating law in several states that will help to ensure more coherent governance and more faithful translation of aggregate voter preferences.
SB 101 by Republican state Sen. Blake Miguez would prohibit ranked choice voting (RCV) for any election in Louisiana (save votes cast by military personnel overseas as by federal law in order to provide an “instant runoff”). There are many varieties to the concept, but basically it involves an electoral structure where all candidates by or regardless of party affiliation (if any) run together in an initial or only election, and then lower-ranking candidates are eliminated in either vote computation or by ballot redistribution.
Louisiana with its blanket primary actually uses a very diluted form of this. The initial general election eliminates all but two candidates by vote computation where the two highest then compete in a runoff election, unless one candidate receives a majority.
But the form most often associated with is concept in the U.S. is for single-position contests, whether for office or a partisan nomination, where only a single election occurs and voters rank order candidates. At the counting phase, unless someone has obtained an outright majority, the candidate with the fewest first-choice votes is eliminated and those voters’ second choices then are redistributed to the remaining candidates. The iterative process continues until a candidate receives a majority of all votes, whether redistributed.
Theoretically, the system is sold as doing a better job of echoing voter preferences because it allows for gradations of support rather than an all-or-nothing choice. Entire reliance on first choices could mask more general support for other candidates. Yet therein lies the fatal flaw for it as well.
The system assumes within a voter’s mind that gradations between candidates and intensities of support are monotonic. That is, let’s say there are five candidates for one position, and that in the voter’s mind he grades out, with 0 being entirely unacceptable, 25 as annoying, 50 as indifferent, 75 as acceptable, and 100 as enthusiastically supporting, candidates of a scale of 0-100.
In this hypothetical election with the hypothetical voter, his preferences are congruent with the assumptions behind RCV if he grades candidate A at 100, B at 75, C at 50, D at 25, and E at 0. But what if he grades A at 100, but B at 90, C at 50, D at 30, and E at 0? His ballot would reflect less support for B and D that he truly feels because the ballot scaling involved (called Likert scaling) imposes an intensity in preference ordering – same distance between all candidates – that may not match the voter’s. This could wash out in the aggregate – say another voter ranks things as A=0, B=10, C=50, D=70, E=100 – but there’s no guarantee of that and in fact is highly unlikely in a system where there are vast differentials in resources to inform voters about specific candidates, as well as vast differences in specific voter capacities to evaluate candidates.
This can create incongruities, if not perversities, in outcomes. For example, let’s use the above example where A and E are quality candidates identifying with opposing major parties, B and D are somewhat lesser candidates identifying with the same opposing major parties, and C is a no party candidate. Resource-rich A and E campaign heavily, B and D aren’t as fortunate but they do get the word out, and C does little. When the votes are in, no candidate has a majority, A and E lead the way as they attract most partisan support, B and D drag the rear with some partisan support, lagging C who picked up mainly nonpartisan support and attracted the bulk of his votes precisely because most voters knew little more about him than he wasn’t a Republican or Democrat. In this scenario, either B or D is dropped and his votes reallocated, leaving the other at the rear and next to be culled.
Prior to dropping the last of B or D, the vote was A at 40 percent, E at 30 percent, C at 20 and the other at 10 percent. But let’s say A and E were bitterly opposed by B or D supporters and they overwhelmingly had C as their second choice, chosen simply because he wasn’t A or E. So, in this redistribution, C narrowly jumps ahead of E, and A remains around 40 percent. And say C was the third choice for A and E supporters – B and D respectively being their second but those now are unavailable as sorting targets. So, what next happens is E is eliminated but as those voters almost exclusively had C as the next viable choice, C ends up winning 60-40 – not because many voters truly preferred him or even knew who he was, but because he was the least disliked by partisans on both sides as well as the least known and acted as an empty vessel.
And this scenario furthered can be skewed by removing the assumption that all voters use, in this example, all five of their votes. Some may give only their first choice, perhaps out of ineptitude in understanding the system or that they strongly dislike all other candidates. Others may not fill out all five choices for similar reasons. If so, a significant amount of roll-off may occur – the final vote count much less than the total votes cast – which leads to many voters not even having a say when it comes to the final result when a candidate achieves a simple majority. In fact, this has been observed in typical RCV-structured elections in America, with the least educated and informed voters disproportionately likely to roll off.
To some degree, these shortcomings can be made less relevant either by having a strong party system that can educate and encourage voters about candidates and the navigation of the complexity of RCV-based elections or by filling (for plenary bodies) offices through proportional representation – or both, as the two often go hand in hand. But we don’t have PR in America nor strong parties, and especially in Louisiana with the country’s weakest.
Indeed, introducing RCV would be a disaster for policy-making in Louisiana, which already suffers from a high degree of personalism in politics focusing too much on candidate attributes and too little on ideology. This is precisely what makes the system so vulnerable to special interests and more resistant to accountability and responsiveness. RCV only would exacerbate this by degrading party – a glue that holds candidates together to focus them on following an agenda on which they campaigned in common rather than as individuals trying to service parochial interests – and increasing the importance of personalistic campaigning and governing.
Thus, it’s best to forestall RCV use. In an election system like Louisiana has – which would benefit from closed primaries for all contests for precisely the same reasons RCV would be harmful, given the state’s political culture, as it strengthens party – aggregate voter preferences become best translated into coherent policy through its first-past-the-post structure than would RCV. This justifies SB 101 becoming law.
No comments:
Post a Comment