John Lewis, the late Representative from Georgia, spoke often of good trouble. That may sound like an oxymoron, a contradiction in terms, but stirring up trouble to achieve a good end seemed to
Atlanta Womens March John Lewis.
(Image by Wikipedia (commons.wikimedia.org), Author: Daniellem4848) Details Source DMCA
Reactions to good strategic voting could be similar. Strategic voting seems anything but good; it is a practice for manipulating elections. But usually, it is done with good intentions and sometimes with good consequences. Strategic voting can be good because compensating for a dysfunctional voting system is beneficial , at least if it works as planned.
With plurality voting, voters have only the option of indicating a single candidate to support. The assumption that the election can select the best winner based on such limited information seems absurd. The winner with the most votes may also be the most widely disliked. A voting system cannot possibly avoid falling into this trap if, as with plurality voting, it fails even to ask voters about their dislikes. Nor can plurality voting possibly take into consideration the possibility that voters may be willing to compromise among similar candidates and so have difficulty in deciding which of several candidates to vote for. Compensating for such limitations is why voters often resort to strategic voting. Rather than voting, as they have been advised, for the candidate they most like, they may vote instead for a different who seems more likely to win. Still, after the election, pundits often pretend that voters did all vote for their favorite. This is what the voters choose, they say to make a point. But surely even these pundits understand that their underlying pretense, that votes honestly express the preferences of voters. to be widely off the mark.
Strategic voting is generally done with a good intent. Voters understand the need to take into consideration the limitations of a voting system and perhaps even weigh electability more than personal preferences. But estimating electability is inexact, making strategic voting in this manner seem much like a game of chance. Voters do the best they can with the options available to them, but what is needed is to adopt a better voting system, one that avoids coercing voters into playing such strategic games of chance.
Ranked-choice voting (IRV) has many enthusiastic supporters but, like plurality voting, it has significant flaws; in particular, IRV also fails to ask voters about who they oppose. But unlike plurality voting, IRV even fails to ask voters about which candidates they support. IRV is a complex voting system, and the complexities of IRV might seem to make strategic voting so difficult as to be impractical. But our previous article hint at how it might be done. As with plurality voting, the strategy would take electability into account. This strategy with IRV will appeal to the voter who, most of all, wants to avoid electing candidates from the opposed list and with that in mind, be willing to make some compromises in choosing among other candidates. Perhaps even more important, the strategy attempts to correct the overly punitive way that IRV deals with abstentions.
Notice that, in an IRV election, when a voter ranks the eventual winner first that voter loses influence on early rounds of counting. Conversely, a ballot that lists first candidates who are unlikely to win, is more likely to influence earlier vote counts, perhaps even casting a vote that allows the eventual winner to avoid elimination in an early round of the tally. For this reason, the strategy is to order preferences with increasing estimates of electability. Determining this ordering involves bit of guesswork of course, but much like the guesswork voters now in practice with plurality elections.
Polls often predict the electability of candidates and adoption of this strategy would depend on that continuing. A widespread adoption of the strategy would probably help polls and pundits retain their present influence on election outcomes. As noted earlier, strategic voting is likely to have a downside, despite it being an honest effort to solve real deficiencies in a voting system.
Following the proposed strategy, the voter would first decide, much as a BAV voter would, which candidates to oppose and which to support. The voter might choose to write out three lists, one consisting of the candidates to approve, the second of candidates to oppose and the third being a grab-bag of the remaining candidates (BAV-abstentions).
The strategy is, for both the list of supported candidates and the list of BAV-abstentions, to first adjust these two lists to be in increasing order of electability. The ranking on the voter's ballot would begin with support list and that would be followed by the BAV-abstentions. It is worth noticing that if this strategy were widely adopted, there might be a rash of elections in which the candidate judged as most electable would get few votes at first and be eliminated in the first round of voting. The pundits and pollsters would be forced to adjust their approaches for predicting electability.
The combined pair of lists could be too short, making it quite possible for the list of candidates to be exhausted before the re-counting of ballots ends. For example, suppose there are ten candidates with three support votes and four oppositions; that leaves three BAV-abstentions. The list prepared for the ballot is only six candidates long in an election that could have as many as nine countings of ballots.
In this eventuality, it is certain that one of the candidates that the voter opposes will be elected. But that disappointment aside, perhaps the voter does make distinctions among the four candidates that the voter opposes. In that case, the voter would be wise to extend the ballot list to include (in decreasing order of preference) three of the four remaining candidates. The winner would, from this voter's view, be one of the worst, but at least the voter will have tried to avoid electing the worst of the worst.
Preparing to follow this strategy might seem more trouble than voters would take on.
But this is the computer age; surely some enterprising programmer would construct a cellphone app to do most of the work. The app might even provide a selection of likely electability orderings. The voter would need only to indicate which candidates are opposed and which are supported. This would be relatively easy, exactly as easy as filling out a BAV ballot. Such an app might, even today, find a market in Australia and other places, like Maine and Alaska, where IRV is now used.
As suggested earlier, this strategy should appeal to voters who are open to making compromises as opposed to having a firm conviction about a detailed order of preference among the candidates. With the strategy, the voter exchanges the opportunity to specify a single most favored candidate in exchange for an improved prospect of avoiding a particularly bad alternative.
That same tradeoff is achieved by BAV, but more directly and efficiently. With BAV the voter is asked only to express these simple preferences, avoiding the unpleasant need for strategic.
