the frustrated voter
- 1 the frustrated voter
- 1.1 under mixed-member proportional
- 1.2 under approval voting
- 1.3 under ranked
- 1.4 allocation ballots
- 2 counting and complexity
the frustrated voter
Imagine the case of a fictional voter whose opinion of the current political options in his or her district are as follows:
- an independent candidate with no real organization, Zachary Zombiecrat, agrees with the voter on most things but has little chance because they lack money and visibility
- the growing All Night Party has the closest platform to the voter's desire, and s/he'd generally like to encourage them to keep competing and grow in the region
- the Pirate Party has been the established opposition for some time and though it has many clearly wrong policies, it's a tolerable way to get rid of the status quo
- the Greed Party is repulsive to the voter and the Preservative Party has engaged in acts that the voter considers criminal, such as bribery and defying campaign finance laws
Let's examine how these attitudes are expressed, or not, in the balloting systems offered by various proposed electoral systems. The companion article entitled the transparent scrutineer explains the issues in counting and verifying that the public will has been expressed. The third article in the series, the responsive party explains what parties should do with this increased information.
Thank you so much for this atircle, it saved me time!
under mixed-member proportional
Now, consider the mixed member proportional proposed (and rejected) in Ontario and Prince Edward Island.
The voter must still make exactly the same unpleasant choice to throw away their preferences of candidate in favour of a barely-tolerable one. However they can send that clear message of encouragement to the Kitchen Party. They are forced to make the unpleasant choice of ignoring the most-preferred platform, that of the Independent Zachary Zombiecrat, because the MMP system will only elect a few "party proportional" legislators and the threshold of electing them is too high for the electors behind Zombiecrat to break through.
The voter can now express separately their tolerances of candidate and platform, but still aren't expressing preferences nor more than one each. Why they do what they do is still a mystery as there is no direct expression of intolerance or disapproval. And they have two separate difficult "least worst" choices to make. Also, while the voter is expressing support for the platform, it is candidates from the Kitchen Party that will actually go to the legislature, and those may be party insiders or officials the voter disapproves of. This appears to have been the major reason why ON and PEI voters rejected MMP so soundly as an option.
Lack of any party marking could be interpreted as a spoiled ballot, or a selection by default of the party associated with the candidate chosen. The latter is preferable as it means that ballots marked exactly as they would be under first past the post are accepted and counted. This reduces resistance to introducing the new ballot system.
under approval voting
One of the basic problems or frustrations the voter has is that they actually consider more than one candidate or party to be acceptable but are forced to choose only one rather than expressing their approval of several and letting the count determine the winner. Approval voting lets the candidate indicate all of the acceptable candidates.
simple: candidate and party must be approved together
For those that would like to be able to approve of an individual candidate but are inhibited by the fact that their vote will also be interpreted as approving of a party, it's possible to support disapproval of the parties on the same ballot, with a separate optional mark such as crossing out the party associated with the candidate.
option to disapprove the party while approving the candidate
This ballot is much closer to expressing the voter's real intent: the most acceptable candidates have been indicated, a tolerable one is chosen as a backup despite hesitations about their party. Despite having three votes none went to the Preservative, Greed or Kitchen Party, and those parties might interpret that as antipathy to some attribute or policy they have in common. The Pirate Party might actually elect Elvis E. Edsel but they can't reasonably interpret this as a vote of confidence in their party or platform if there are many such party-disapproving votes. If public funds are being distributed based on the votes, the Pirate Party will get none as a result of this ballot.
There is no need to cross out the parties if one did not approve of their candidates, so this ballot may give the wrong impression that the voter has no problem with the Greed or Preservative parties.
Another, more common, approach to indicating uneven levels of support for parties or candidates is to rank them. A simple ranked ballot is used in both single-winner (instant runoff) and multiple-member (single transferable vote) systems. Some such systems, including the one used in Australian federal elections, require every candidate to be marked. Assuming that is not the case here, the ballot reads:
Now it's quite clear that the voter prefers the Independent over all other options, likes either the candidate or platform of the All Night Party next best, and will accept the Pirate Party only as a barely tolerable choice. If they had the option of ranking others 4, 5, 6 and didn't, that again tells the Preservative, Green and Kitchen parties that they do not appeal to this voter at all. Signals to parties are even clearer than the disapproval ballot but without explicit disapproval it's hard to tell whether the third choice is really as reluctant as it is for this particular voter. If funds are allocated or refunded based on the popular vote, the Independent and Kitchen Party will both be encouraged by this result, more than Pirate which was clearly ranked third and would receive very little or none as a result.
Other ballot options could exist, including first past the post ballots with party disapproval or ranking parties and candidates separately. These may add more complexity than they're worth to the average voter. If more detailed assessment of voter preferences are desired an allocation ballot might work. The simplest version lets the voter assign points to a particular party/candidate combination up to some fixed total - in this example, five points.
five positive points
Their strong preference for the independent is just as obvious as on a ranked ballot, but they've decided to encourage the All Night Party more and let the Pirate Party find its own way. If they had six points they would probably have assigned one to the Pirate Party.
But except for the party disapproval option above, the voter still has never been able to express their actual motive for voting. Imagine an allocation system that allows not only positive approval but also negative disapproval points to be allocated to party/candidate combinations. Under this system it's much clearer what motivates the voter:
five positive or negative points
They have abandoned all compromises and clearly indicated that they disapprove of two parties strongly, are not really opposed to the others, and strongly prefer the Independent. This ballot might have less impact on the power struggles between the Pirate, All Night and Kitchen parties except that Kitchen was left out of the disapproval. Receiving a great many such ballots might cause the Kitchen Party to look at the Independent's platform because the opponents of the Preservative and Greed parties may be easier to unify under their banner if they embrace some of the Independent's ideas. Individual candidates are less likely to benefit from negative assessments than parties because they can change their personalities less easily than parties change platforms - accordingly, allowing only positive approval of candidates but negative approval or disapproval ("striking out") of parties might lead to less resistance to this kind of balloting by the public at large. It's also possible that the absolute approval system might be confusing to voters though it's actually difficult to spoil a ballot - you'd have to assign more points than you were allowed, and it's hard to assign any points if you don't understand how to use the ballot in the first place. Even a simple X can be interpreted as a +1 (or even +3 or +5) if it's desirable to let people vote as they always have.
A problem with this kind of voting is that it expresses the strong preferences and tolerances of the voter, or it lets them choose a compromise, but it doesn't let them do both very easily. Expanding the number of allocation votes to ten lets the voter more fully express both their strong beliefs and their preferred compromises:
ten positive or negative points
Given five extra points, the voter uses them to express stronger distate for the Preservative Party, a mild or perhaps strategic preference for the Pirate Party over the Kitchen Party, and again gets to encourage the All Night Party more strongly. While they express less relative support for the Independent, they are now actively choosing the compromises inevitable if the Independent is not popular enough to really win. A party official, candidate, journalist or political scientist can assume from this ballot that the voter reacts negatively to the Greed Party and very negatively to the Preservative Party or to their candidates, is utterly ambivalent to the Kitchen Party and would mildly prefer the Pirate Party to them, finds only the All Night Party has much appeal as a political formation, and would really prefer an Independent to win. This is more or less exactly the specification of the voter's original intent above.
counting and complexity
Since a ballot marked exactly the same as an FPTP ballot can be interpreted as a first-rank or maximum-points-allocated ballot under the more complex systems, ballot complexity is not an issue for either. By contrast, MMP ballots absolutely require two marks unless a vote for a candidate is interpreted by default as also a vote for his or her party.
marking, hand-counting/recount and interpretation issues
Not all of these methods are equally amenable to hand-counting and scrutineering. As the complexity and cost of counting the ballots by hand increases, some of these methods will become impractical. As electioneers begin to change to mechanical vote-reading methods, some of these remain easy to read with mark-sense, punch-card, or optical scan, and others become unlikely. Worst case, electronic counting methods might become relied upon exclusively, so that even if ballots were in principle possible to recount, in practice they would not be, leaving these methods open to all the fraud potential of e-voting.
Interpretation of hard-to-read or non-standard ballots gets harder under the more expressive ballot schemes also. In the first-past-the-post and other traditional methods, any ballot not marked exactly as indicated is 'spoiled'. In the methods shown later, many ballots marked in even slightly unusual ways will remain subject to debate. What if they were clearly marked, and a clear preference expressed, but in ways not prescribed? Did they misunderstand the instructions or just wish to express their individuality? In some close or very contentious votes, marginal votes like that could become a significant proportion, making "hanging chads" look like a simple issue by comparison.
Still, the value of political expressiveness and retaining the possibility of a practical hands-and-eyes paper recount by qualified scrutineers is so great that some solutions to these problems should probably be considered. A few such:
- Accept voter input electronically in the booth, with extensive help, but print a paper ballot they can verify and mark as acceptable. An electronic tally could be kept, or not, but should have no legal significance other than to help signal potential frauds to investigate.
- Set an objective standard for the user-friendliness of ballots to both the voter and the counter. For instance, 99% of citizens must be able to mark the ballot correctly and readably when motivated (say by a reward for participating in a test) and 85% of literate citizens must be able to count these ballots to within 95% accuracy given a training program of a single working day. This permits balloting and counting schemes to be tested in advance to these thresholds, and any chance to legally object to them would then be exhausted well before the pressure of an election. If the thresholds are too low, they can be adjusted before the next election.
- Educate people in primary and secondary school how to use such ballots.
- Accept a marking method of numerals as Roman-style marks (III for 3) and implement any negative or disapproval marks by putting them in a different column rather than using "minus" or "negative" marks. This may reduce the dependence on literacy as Arabic numerals need not be employed.
Note, also, that the complexity of some STV counting schemes is significant and the complexity of district border "gerrymandering" under MMP and FPTP schemes is actually much greater than that of anything above. It is probably a false hope to expect the overall system of districting, nominating and electing representatives to get so simple that all aspects of it could be fully understood by a majority of the public. A general education program on the full complexity of the system as it exists is probably the best antidote to objections of any new complexities.
As with e-voting, systems can be simple for the end user but more complex and less reliable as a result of seeking that simplicity. This principle should guide any advocate who encounters objections regarding "complexity".
democracy demands transparency increase over time
Democracy probably requires over time a generally higher assumption of public education and the removal of complex hidden processes that have been exploited by power elites and used to "simplify" democracy for the public. Where the public chooses to rely on some group of experts to say draw district borders, the process by which this is done should remain visible. So in such schemes as BSTV+C+L, while ecologists would be the primary authorities drawing objective district borders, there would be many opportunities to object to their recommendations or challenge experts, certainly not fewer opportunities than exist under gerrymandering.