I've been asked to explain and justify some of the claims I've made in my advocacy for Strategic Election Reform or the use of 3-seated State Assembly elections.
Here's a summary of a source that describes the results of comparative analysis of electoral rules across countries.
Choosing an Electoral System takes its title from a larger work that was edited by Princeton scholar Arend Lijphart.
See my reader's guide and thoughts below...
As I have written, Strategic Election Reform is a position in the electoral debate over what sorts of election rules we should use that affirms the need for both winner-take-all, or single-member, and winner-doesn't-take-all, or multi-member, elections. We need to use both types of elections in different sorts of elections. The obvious example is to use winner-doesn't-take-all 3-seated state assembly elections and winner-take-all state senate elections. Another example would be to use a two-staged election that has a winner-doesn't-take-all first stage and a winner-take-all second stage. The "top two primary" used in Washington State and recently adopted by California is a flawed version of this type of election rule. I also tacitly argue that the sorts of options given to voters are of second-order importance to working out the right balance between these two basic types of elections.
But how do I prove this? Why take my word?
Here are the main conclusions from the first source linked to above.
- The most important determinant of the proportionality of a system (the percent of seats won equals percent of votes received) is the magnitude of its electoral constituencies. Ie, elections with more contested seats tend to be more proportional on average.
DLW:Why does this matter? Proportionality is often held to be the gold standard for a fair vote. However, proportionality in winning seats does not entail proportionality in gaining power. And it's power that matters more to most politicians than the ability to win seats. Also, if we are using both single-seated and multi-seated elections then the outcomes of the single-seated elections will always tend to favor bigger parties. As such, why be preoccupied with making the multi-seated elections as proportional as possible?
- Turnout is usually higher at elections in countries with Proportional Representation(PR) than in countries without. In general, the more choice potential voters are offered the greater the likelihood that they will turn out and exercise it. It also seems that PR with fewer seats induces more turnout than PR with greater numbers of seats. And if voters have to vote for long ordered lists of party candidates, that is less effective than letting them vote for individual candidates in getting them to turnout to vote. Also, more complicated election rules do not seem to help to increase turnout as much.
DLW:An underlying criteria here could very well be whether the elections are competitive or uncertain in their outcomes. If we believe our votes might very likely make a difference then we're more likely to turnout to vote. 3-seated Hare Largest Remainder elections are about as small and simple of PR as you can get. It does not use lists of party candidates. If a party's candidate does very well then (s)he could win two of the three seats and get to pick a (somewhat lesser) partner (with a very similar politics) to hold the second seat. This would permit third parties to win seats. These third parties would have an incentive to persuade habitual nonvoters to vote. On average, a higher voter turnout will tend to help the more progressive/liberal party, like the Democratic party do better in single-seated elections.
- FPTP(The US's current primary election rule) and AV(Britain's term for IRV) are more likely than PR systems to produce single-party majority and long-lasting governments. This is because the largest party gets an even larger share of the seats. IRV reduces this tendency some compared with FPTP, but it still remains. But this tendency is not strong and lots of other factors matter.
DLW:Single-seated regional (state) elections tend to produce a dominant party in a region or state. It need not be the same party in each region/state. In the US, we had strong third parties in Wisconsin and Minnesota and New York earlier in the last century. Our presidential election and the freedoms given to $peech helps to keep us with a two-party system that often tilts to effective single-party rule at the state level.
- Multi-member (or seated) elections tend to increase the number of political parties. Other rules are often used to keep the number of parties down for the sake of political stability and to keep small parties from wielding an undue amount of influence.
DLW:If we still use single-member elections for most of our elections, as we are currently constrained to do so with federal elections, then we do not need to worry about the undue proliferation of parties. Also, a 3-seated election would never lead to an undue proliferation of parties. The two biggest parties would tend to be almost guaranteed one seat each and then there'd likely be two or three small-but-not-too-small third parties trying to win the third seat. The rule with 3-seated Hare Largest Remainder is to award the top three candidates one seat each, unless the top candidate beats the third place candidate by more than one-third of the total vote. In that case, the top candidate wins two seats for her/his party and gets to pick who would hold the second seat. So a third party could win with as little as 10% of the vote if the top candidate got less than 43.3% of the vote.
- When single-member elections are used, small shifts in support between the dominant parties among "swing voters" can lead to many seats changing hands between the dominant parties.
DLW: Can you say election of 2010? This is important to keep elections interesting... If we already knew which parties would be in power then it'd be easy to lose interest in the election. This is why many countries are using mixed systems where some of the seats are elected in a single-seated election and others are elected in a multi-seated election. In Mexico, they gave non-PRI parties more of a chance when the 1986 Electoral Reform Law enlarged the Chamber of Deputies from 400 to 500 seats and doubled the number of congressional seats filled by proportional representation to 200. If we used 3-seated and 1-seated elections for our state legislatures then larger swings between the two major parties in the state senate would be compensated by their relative parity in the state assembly. Of course, if our state-districts were redrawn by a state assembly elected with 3-seated elections then the parity between the two major parties and presence of third party reps would put the kibosh on the ability of one party to gerrymander the state districts. This would result in more competitive state senate districts and the party in power in the state senate would have a higher chance of falling out of power in an election.
Point 6/7 are not as important.
- It has been argued that PR makes it easier for underrepresented minority groups to gain election.
DLW:Why should it matter whether a small group can get their own representation! What matters more is whether they can gain influence on their key issues as a group. 3-seated Hare LR would make it so a third party could win a seat with as little as ten percent of the vote. A smaller group, whether it be ethnic/race or economic(the poor) or ideological(green/libertarians) would have a better chance of being the decisive voters for this third seat than if only single-seated elections were used. The lower voter turnout among many historically disadvantaged groups in the US can be attributed to their inability to afford an exit threat from the Democratic party and thereby difficulty in getting attention to their key issues.
The vitriol with which African-American philosopher and radical political activist Cornel West has recently decried president Obama shows that the election of an African-American doesn't necessarily lead to policy changes that help the majority of African-Americans. Obama's incentive is to cater to the voters and donors who are more likely to help keep him and his party in power. Most of those voters/donors are not African-Americans.
- Under FPTP and IRV many seats are noncompetitive. It would require a very large proportion of the electorate to change their partisan preference between elections to change the election outcome. Parties have a low incentive to campaign in non-competitive elections.
DLW:This is why we need to use more multi-seated elections in the US! There can be good things about non-competitive elections too, but the distribution of competitive and non-competitive elections will be more even across regions if we used 3-seated state-assembly elections. If third parties specialized in contesting winnable 3-seated elections then they could vote strategically together in single-seated elections and make more of those elections become competitive.
- Political targeting is too effective in winner-take-all election campaigns. The swing-voters get phone-banked too much and get an undue amount of attention given to their issues, contrary to the democratic ethos.
DLW: If we used a variety of election rules then who are likely to be the swing-voters would be a wider set of people. If the two major parties tend to change their policies so that they are relatively equal in number of supporters then the swing voters in single-member elections would often be members of smaller third parties that vote strategically together. These groups would likely force the major party candidates to debate the issues, more so than they do today with political campaigns heavy on low-content slogans designed to appeal to low-information independent voters.
The article then considers two tacit trade-offs among the election rules. The first trade-off is between the representativeness of parliament and government accountability. More PR tends to produce a highly representative parliament. Single-seated elections help to deliver accountable single-party government. The rub is to achieve both of these objectives. Solutions include: 1.The use of rules that discriminate against smaller parties. 2. The use of smaller multi-seated elections (as in like 3-seated elections). 3. The largest party gets a "winner’s bonus" after a PR election, which guarantees them a parliamentary majority or a less unwieldy alliance. Unfortunately, sometimes attempts to combine both rules into one rule results in less representative governments that are also not accountable.
DLW:Mixed elections that combine winner-take-all and winner-doesn't-take-all elements in one election tend to be watered down winner-doesn't-take-all elections. My Montesquieu-inspired hypothesis is that the ideal system must use both winner-take-all and winner-doesn't-take-all election rules. The US's bicameral legislative system could help in this regard. If the newly elected state assembly-persons elected by plurality vote which party is in power then the leader(s) of the party in power could get extra procedural controls to help them to get things done transparently even though their party only controls 1/3+ of the seats. The third party representatives would essentially get to pick which of the two major parties is in power after the parties have chosen their leaders. The smallness of the third parties due to their ability to win a seat with only ten percent of the vote would make it harder for the leader(s) of the party in power to abuse their extra powers to try to stay in power indefinitely. So Strategic Election Reform essentially combines options 2 and 3.
The second trade-off is between the accountability of individual politicians and the cohesiveness of political parties. We want both but they tend inevitably to be in tension. Many forms of PR with large lists tend to ensure that political parties are cohesive. But the legislators are not given any autonomy from their party.
DLW:In a 3-seated Hare LR election if the top candidate wins two seats then the partner they would choose to hold the second seat would probably not be able to vote their own conscience. This does not mean that the top candidate would treat their partner as a "slave". But there would be a clear-cut and justified hierarchy, since the top candidate was the one on the ballot who got a very large portion (likely a majority) of the total votes. If the party leaders are given more procedural controls over their parties then that would serve to increase their cohesiveness some. The leader(s) of the second place party in the biannual plurality vote to determine which party is in power in the state assembly might be given some additional procedural controls over their party members. These controls would mitigate the ability of individual legislators to game the system to their own advantage. The leadership of a party would be transparent and the outcome of a vote by elected assembly-persons within a party. Which party is in power would be determined publicly by a vote of elected assembly-persons in a relatively proportional manner, weighted somewhat in favor of minority parties that are unable to be in power.
dlw