In this third article of a series we introduce Balanced Approval Voting and argue that it is not only a better system of voting but one that we should seriously consider using.

::::::::

The second article of this series, What's Wrong with First-Choice Voting?, explored deficiencies of First-Choice Voting (Plurality Voting), the voting protocol that we generally use today. That article was intended to make the point that we should use a better voting system.

The first article, Are More Political Parties Possible Here? argued that a better system would be balanced, meaning that a voter could choose to cast a vote against a candidate as easily as for a candidate. Described in that article seems to be as simple a balanced voting system as possible - allowing each voter to cast a single vote but to require it specified whether that vote is for or against a specific candidate. When counted, a vote for a given candidate adds one and a vote against subtracts one from the net vote tally. This simple and natural example of balanced voting we will call, BV-1.

Other possibilities come to mind, however. A voter could be allowed two (BV-2) or three (BV-3) votes in place of just one. Is there a voting system that is best? It is not clear how to make that judgment, but as important as voting is for democracy, it is surely a question that is worth asking.

We can't dismiss the possibility that one of these systems, BV-5 for example, might be judged to be the best voting system, but that choice does seem arbitrary. But a system that does not seem so arbitrary is to allow a voter to cast as many votes as there are candidates. Accustomed as we are to casting only a single vote this may seem overly generous, but on further consideration, it appears to be a particularly good alternative.

A voter is asked to cast either a for or against vote for each candidate in the race; but abstention is offered as yet a third option. In effect, the voter partitions the candidates into three disjoint sets:

1. The candidates that the voter likes (or at least finds acceptable) will get a vote For (+1),

2. those the voter dislikes (or finds unacceptable) will get a vote Against (-1) and

3. about all the remaining candidates, the voter is undecided or ambivalent (0).

Thanks go to Steve Unger for pointing out the system of Approval Voting (AV is discussed at http://www.electology.org/) that is similar, though not balanced. With that system, the voter partitions the candidates into only two sets,

1. The candidates that the voter votes For (+1),

2. all the remaining candidates (0).

Recognizing the great similarity, this balanced system (with a three-way partition) will be called Balanced Approval Voting (BAV). Is this the best system of voting?

A curious feature of BAV is that typically, a voter does not indicate a first-choice candidate; the only way to do this would be to forgo the opportunity to vote For more than a single candidate. In this respect, BAV is nearly at the opposite pole from FCV which asks the voter only to specify a first choice from among the candidates. But much the same can be said of AV.

After the polls close, for BAV the votes For and Against each candidate are tallied and a net vote for each candidate is computed as the difference. These net vote counts tells us the net satisfaction with electing that candidate. A positive net vote means that number more voters would feel pleased than would feel displeased with the election of that candidate. Electing the candidate with the largest net vote would mean that the largest possible number of voters would be pleased rather than displeased with the outcome. Surely that is a result that we should be ready to accept and defend in a democracy.

A consideration in judging a voting scheme should be how stressful it will make voting. To a large extent this comes down to how accurately the ballot allows the voter express what he or she wants. Voters who feels unable to accurately express preferences for the candidates may feel cheated at the polls and may grow less inclined to vote in future elections. Much the same can be said for voters who feel pressured to make arbitrary (flip of a coin) decisions to satisfy what may seem like arbitrary demands of the ballot.

The first article of this series, Are More Political Parties Possible Here?, notes that balanced voting encourages elections to competitions between mort than just two viable candidates and that having more than two viable candidates naturally discourages polarized politics with its negative advertizing.

With BV-1, a voter in a polarized race with two major candidates may choose to vote for the candidate that the voter most prefers or possibly against the other major candidate. Choosing that second option improves the chances for some third-party candidate to win; each leading candidate's supporters and opponents may cancel each-other and this opens the possibility for a third-party win. With BV-1, whether to cast a positive or a negative vote is a choice that is forced on the voter.

The other balanced voting systems under discussion allow a voter to cast two or more votes and in a polarized political environment it is predictable how they will cast those first two votes. Most voters will likely vote for the major party candidate they prefer and they will use the second to vote against the other major party candidate; in a polarized environment that seems like getting to cast two votes for your favorite candidate, but with roughly equal numbers of voters for both poles of the political spectrum, the two major party candidates will net close to zero votes. What seems to individuals to be voting twice for their favorite candidate turns out, when that choice is made by everyone, to open the door for a win by someone other than in the two major parties.

In an already polarized political environment, any of these multiple-vote balanced systems will disadvantage candidates of the major parties much more than BV-1 does. Given how accustomed we are to seriously polarized politics, we might well wonder whether this disadvantaging of the major party candidates to this significant a degree is a wise thing to do. Imposing BAV (or even BV-2) suddenly on an already polarized political environment would probably deliver quite a shock.

If instead, BV-1 were adopted for the first few elections, voters would have a chance to get accustomed to the notion that someone who is neither a Democrat or a Republican might be elected. Likewise the media would have a chance to come to the same realization and to begin informing the voters about other candidates. A the same time that polarization decreases, voters would become educated about about balanced voting. Initially, while there are a relatively few candidates, BV-1 should work well; eventually, as voters become accustomed to balanced voting and there are more candidates and the limitations of BV-1 become more apparent, BAV could be adopted in its place.

With BAV, just as with any other voting system, there is a distinct possibility of a tie or a near tie in any given election. Electing a candidate on the basis of one or even a few hundred votes is apt to result in hard feelings if not calls for recounts, protests and perhaps even law suits. If there are two or more candidates in a statistical dead-heat after a BAV election an option might be to have a runoff election but this time to take into account the first-choice preferences of the voters. A consideration in changing the voting method is that a runoff using the same method as before is apt to yield much the same inconclusive result. A second consideration is that the level of enthusiasm for a candidate really is important, though not as important as it is to respond to the wishes of the greatest number of voters.

But elections are expensive to conduct and they disrupt people's lives so a decision to call voters to the polls a second time is not generally a realistic option. Fortunately, there is another possibility and that is to anticipate the possibility of such a situation and prepare for it in the first (an only) ballot. This need not require a complicated ballot. All that is needed is a third column where the voter can check off which one candidate (For or Against) is most important to the voter.

In the event that there is a tie or a near tie in the first-round BAV count, there can be a runoff re-count using only those votes marked for use in the runoff. Of course there will be some votes marked for the runoff will not count because they are For or Against a candidate that was eliminated in the first-round BAV vote count, but these are from voters who have indicated a relatively low interest in the second round, given that their most favored candidate lost in the first round.

In both rounds of counting (should both be needed), the counting is balanced - opposition to a candidate weighs equally against support. With this two-stage approach, in the first (and possibly only) round of counting it does not matter how strongly a voter feels about a candidate; it is only the voter's judgment of approval or disapproval that counts at all. But in the second round of counting the factor of importance does enter importantly into the decision of which of the runoff-candidates is to win. But whichever candidate wins the runoff, democracy is well served because these runoff candidates have already been judged in the first round as (within a small margin of error) as being equally acceptable (on balance) to as large number of voters as possible.

Authors Bio:

Attended college thanks to the generous state support of education in 1960's America. Earned a Ph.D. in mathematics at the University of Illinois followed by post doctoral research positions at the Institute for Advanced Study in Princeton. Taught for several years at Lehigh University prior to a short stint at Bell Laboratories but followed by a much longer career at NEC punctuated by ten U.S. and international patents in the general area to semiconductor applications.

Now living in a comfortable Maine retirement community and focused on the prospect of upgrading democracy by means of an improved voting system.