We have internalized belief in the two-party duopoly to the point that it is even ingrained even in habits of speech. We talk casually about "the other party" as if there could not possibly be more than one other
party. Experience with the duopoly confines our minds to a box and it takes effort merely to think outside of that box.
It seems ironic then that among people who do focus on voting methods, just the opposite problem also persists. Experts forget or perhaps just deny that there even is a two-party duopoly. This finds expression in the hope that we will automatically escape from the duopoly if we just adopt another voting system such as instant runoff voting (IRV) or approval voting.
IRV does avoid at least the most common forms of the spoiler effect, a problem that is clearly a barrier for small political parties. But while avoiding that effect surely is a step in the right direction, it is hard to ignore that after more than a century since IRV was adopted in Australia, that country still retains basically a two-party system. Avoiding the spoiler effect does improve elections by helping to avoid electing the least popular major-party candidate, but there is little evidence to suggest that adopting IRV will end the two-party duopoly.
Approval voting is yet another alternative system that appears promising. It clearly accommodates elections with many candidates and like IRV it avoids the most common forms of the spoiler effect even including ones that afflict IRV. But compared to IRV, approval voting has suffered from less promotion and it has fewer adherents; it has no record of being widely tested in Australia or elsewhere; BAV likewise suffers even to a greater degree from these deficiencies. While it is not impossible that with years of serious approval voting will be successful than IRV in encouraging additional parties; but that hope alone is no reason to feel much confidence. But until there being such an experiment it seems that we are limited to thinking carefully about what we should expect with approval voting.
Both approval voting and IRV do have a great appeal, in large part because each system welcomes voters to express how they feel about any or all of the candidates (even elections more than just two). But this freedom is of little value in elections that in fact include only two candidates. And it is clear that any transition away from duopoly must surely start with an electoral win by a minor-party candidate. At some time we simply must get off the dime and take that first step. And unfortunately, a single win by a third party will probably not be enough to reshape the attitudes of voters, particularly if that comes only many years after a new voting system is adopted. If a win seems merely a freak accident, voter expectations will not likely change.
As we have noted (see for example the previous article), unless voters can effectively express opposition as well as support for candidates, the voting system gives a significant bias favoring the very well known candidates. And whenever there is a two-party duopoly the most famous candidates will most likely be from the two dominant parties. The imbalance in many voting systems (those that are not balanced) presents a serious hurdle to to clear for other candidates but one might argue that even without balance it may still may be possible for a voting system to clear that hurdle.
Why there is resistance to allowing votes against candidates is a mystery, at least to me; perhaps it is just a resistance to change. The wishful thinking is that these other systems are good enough no doubt does play a role as well. But a priori, it does not seem impossible that there is some yet unnoticed magic in IRV or in approval voting that will some day allow minor parties to grow into viable competitors. But if so, there surely there must be a first such election. In that election, contrary to historical experience and everyone's expectations, a third party candidate wins. As noted before, even this might not be enough to make a dent in voter expectations, but surely it is necessary for there to be a first win by a minor party. So let us think about that particular and extremely surprising election.
Let's first consider such an election using IRV. If experience is any guide, voters will mostly ignore the campaigns of minor-party candidates. Such candidates may appear on the ballot but that seems to voters as a mere formality, nothing more than a pretense at elections being open to everyone. In any event, information about these candidates is not widely reported and it may even take an extraordinary effort for a voter to learn much about them. Despite this, there will probably be a few voters who do submit ballots showing a top preference for one of these minor-party candidates. And there will probably even be a few voters who, as a matter of principle, show a preference for any of the minor party candidates to either of the two (or three if there is a spoiler) major-party candidates. But even these few voters know themselves to be actually tilting at windmills. They are well aware that their most-favored candidates have no realistic prospects of winning the election, so in most cases, even these voters will express on their ballots a first preference for one of the major candidates.
When the IRV ballots are counted, all of the minor-party candidates will first be eliminated. Ultimately, the only question remaining is which of the two major-party candidates is preferred by the larger number of voters and so a major party candidates will win. What this shows is that with a polarized electorate, there will be no first election with IRV that marks the end of duopoly politics. And without a first there can be no second, etc. The duopoly will be maintained.
Turning now to using approval voting for this same election, we might recall that like IRV, approval voting suffers from a somewhat altered form of spoiler effect. A voter in an approval voting election is allowed to vote in favor of any number of candidates; however a voter is only allowed to vote against only a single candidate. This limitation to only one vote of opposition can force voters of like mind to split their opposition votes between two or more candidates they find equally unsatisfactory. This splitting of votes creates a problem very closely resembling the spoiler effect.
In exercising this option of voting against a candidate (which capitalizes on the subterfuge of voting approval for all of the other candidates), it should be noted that the voter sacrifices the opportunity of expressing any opinion at all about the other candidates; this makes the option seem an unlikely choice but that is only when viewed from the context of there are many candidates. For elections with only two candidates, voting for all the other candidates amounts to simply voting for the one other candidate, something many plurality voters do quite frequently.
An example with three candidates is particularly interesting and that is the electorate that is commonly used to illustrate the spoiler effect in a plurality election. In addition to the two major-party candidates there is a third, independent candidate, perhaps a disappointed Democratic politician who competed for the Democratic party's nomination but lost. Republicans of course vote for the Republican candidate. And most Democrats vote for both of the Democratic candidates (one running as a Democrat and the other who is masquerading as an independent). From examining the ballots, this miight be explained as the Democrats choosing to cast a vote against the Republican candidate. But it is likely that the Democratic voters honestly support both of the other candidates and were, in their own minds, simply voting honestly; it seems doubtful that many were even aware their ballot might be interpreted as a vote against the Republican, though they probably would be quite comfortable with that interpretation.
Might this example generalize in the sense of there being essentially two Republicans competing with two Democrats? Or perhaps to one Republican running against three or four Democrats? It does seem possible and in either instance the election could result in a tie. These situations would likely be rare, but if approval voting is adopted, some planning is need for handling these awkward situations. But however these odd events are handled, these situations provide an opportunity for a very technical win by an independent candidate. In a very real sense they only affect the question of which duopoly candidate will win.
In the example election, now held using approval voting (rather than IRV) he major-party candidates compete with one or more minor party candidates. For the majority of voters, the situation is like that with IRV; the minor party candidates are mere footnotes. The ballots, for the most part, will profess approval of just candidates from one of the two major parties. A small number of ballots will likely also show approval of one or more of the minor-party candidates but not in significant number compared to the many votes in favor of either of the major-party candidates. In the very few cases where voters have formed an opinion, a negative opinion is apt to be as common as a positive one, but in these cases of heartfelt opposition, the voter can do no more than to join the majority of other voters, indicating (falsely in this instance) to having no opinion about the minor party candidate.
So again, as with IRV, for this election which is assumed to provide the first win by a minor party, the only winner will be one of the two major-party candidates. We have to conclude that, as with IRV, approval voting offers little hope for ending duopoly politics. And without ending the duopoly, these two alternatives offer little hope of much change beyond a reduction in the frequency with which the spoiler effect corrupts our elections (by allowing the wrong duopoly candidate to win election).
The point here is that in order to end duopoly politics, a voting system has to offer a polarized electorate some way to get off the dime and start (actually, in reasonably frequent elections) to elect a candidate that is not from the two major parties. Voters need to be convinced that the minor party candidates are possible winners, even if not probable. Fortunately there is very good reason to believe that BAV can do that.
There are two important keys to this potential from adopting BAV. One key is balance, allowing voters to express opposition to a candidate as easily as support. The other is for the voting system to be evaluative, asking voters to merely to express their opinion about the candidates on their own, rather than in comparison to other candidates.
Whether BAV can produce third-party wins frequently enough or soon enough to change the way voters think about minor-party prospects is, for now, guesswork. That is in large part a question about voter psychology so it will take real-world experience using BAV in elections to find out.
