Araucaria

Joined 12 September 2018
Revision as of 01:33, 16 September 2008 by imported>Araucaria (Simplify primary vote proposal: augmented Top Two primary)

The etymology of my username is explained on my Metafilter User Profile.

It is important that we not be confused by the false choice that there are only two viewpoints on any issue. This is why it is important that there to be enough candidates in the general election to challenge common wisdom, talking points and the implicit agreement in the two-party duopoly to avoid discussing issues that actually affect us.

Instead of our current (Single vote) system, I would prefer a voting method that does not force one to vote for a compromise based on assumptions of what other voters would do.

In the interest of simplicity and expediency, I favor an incremental transition as follows:

  • First Choice plus Approval: Single First Choice vote, plus Approval of any number of other candidates. The First Choice is also approved. If no candidate wins >50% of the First Choice votes, elect candidate with highest approval.
  • First Choice plus Range voting. Single First Choice as above, plus a score of 0 to 99 can be given to any number of other candidates. First Choice is also assumed to have highest score (i.e., 99). In the event that no candidate has a majority (>50%) of First Place votes, use the same Approval fallback as above, except cumulative ratings score is used instead of total approval votes.

Why do I favor Range voting over Approval? It allows the voter to show a lesser preference for some candidates, if only to encourage them to keep contributing to public discourse. And perhaps more importantly, it can be used to infer candidate rankings by sorting the candidate's ratings in order from highest to lowest. Candidates with equal ratings would be given equal rank.

Once a Ratings ballot is used, the scores could be tabulated and reported with 5 different methods for comparison: Top First Choice (Single vote); Range voting; plus 3 robust Condorcet completion methods: Schulze; Definite Majority Choice; and Cardinal pairwise using River.

As voters recognize the difference between FC+RV and more sophisticated methods, support would increase for the transition to a more robust voting system.

In the case of no candidate winning a majority of the First Choice votes, we could select from one of the other four robust methods at random to discourage strategic voting. Since all four methods have similar sincere voting strategies but different strategies in the case of cycles, it would give the voter a strong incentive to vote by sincere preference instead of trying to game the system, thus avoiding Arrow's paradox.

I would prefer to avoid primary elections. However, they may continue to be required during a transition period. If that is the case, I would recommend using

  • First Choice plus Approval. As above. Pitch the method as an augmented Top Two Primary: Narrow field to at least 2 candidates, comprising the two candidate with highest First Choice totals (single vote winner plus single-vote runner-up), plus any other candidates with equal or higher approval than the top two single-vote candidates. This would automatically include at least the top two Approved candidates. The main advantage of this is simplicity, plus it would be a marked improvement of the Top-Two Louisiana-style primary using Single Vote --- voters would be assured that a good selection of alternative candidates would face the Single-vote winner in the general election, instead of dividing the opposition.
  • First Choice plus Range voting. As above. Narrow field to at least 2 candidates, who would include the Single Vote winner and runner-up, plus any other candidate with a pairwise beatpath to the cumulative ratings winner, plus any candidate with higher cumulative rating than the Single Vote winner and single-vote runner-up. The improvement here over First Choice plus Approval would be that the entire Smith set (plus first- and second-place range voting winners) would be included in the slate of candidates going to the general election.


You can contact me at araucaria dot araucana at gmail dot com.

Election Reform Priorities

My highly opinionated views on what can be done to improve the US form of government:

Trust in the process

Before anything else, we need to trust the mechanics of voting:

Redistricting reform

Redistricting is too entwined with partisan politics. By federal law, states should be required to enact non-partisan restricting reform along the lines of Iowa's successful law. In a nutshell:

  • First, after the census, the state sets up a 5 person redistricting commission. The Democrats and Republicans choose two members each. These four then choose a fifth non-partisan member who serves as the chairperson. By law the fifth member cannot be a holder of any political party office and cannot be a relative or employee of any member of the legislature.
  • This five member commission then sets about redrawing the districts using several legal guidelines. Among these are:
    • Population - The districts must be as equal in population as possible.
    • Respect for political subdivisions - for example county lines.
    • Contiguousness - The district must be contiguous. Iowa law says that any area that meets only at the points of the corners is not contiguous.
    • Compactness - This is the relative "squareness" of the district. As much as the other factors will allow, the district must be as compact as possible. This gets rid of those oddly shaped districts seen in other states.
    • Political and racial neutrality - The Iowa Code states that districts shall not be drawn to favor any political party, an incumbent legislator or member of Congress, or any other person or group. The district also cannot be drawn for the purpose of augmenting or diluting the voting strength of a language or racial minority group.
  • To insure compliance with number 5 above, Iowa law prohibits use of the following data:
    • Addresses of incumbents
    • Political affiliation of registered voters
    • Previous election results
    • Any demographic data other than population headcounts.

Transparency and efficiency of legislative process

The current US legislative process moves at the same leisurely pace it did in the mid-19th century.

  • Riders are attached with no warning and bills are not always voted on by all members.
  • There is no reliable way to determine accountability. This harkens back to the days of "smoke-filled rooms".
  • Career politicians are relying on lobbyists to write legislation.
  • The only way to debate is on the floor of the House.

We need a method of legislation that would works quickly and transparently, whatever the size of the legislative body, be it 50, 500 or 50000. The rules need to be open and non-partisan. And the system of making those rules needs to be protected by checks and balances.

One way to streamline legislation would be to compress the highly inefficient technique of Robert's Rules of Order, which reduces every decision to a series of Yea/Nay votes. A strong Condorcet method could be used to resolve the preference between many options at the same time.

Better Representation

The House of Representatives is not representative

The US House of Representatives has been frozen at 435 members for nearly 100 years. That's one member for every 640,000 people. The Founders originally thought there would be one member for every 30 to 50 thousand people. The number of representatives should be increased to at least 600. This would also decrease the over-representation of small states and under-representation of large states in the Electoral College.

Proportional Representation

Local representation is best implemented using Proportional or Full Representation, with at least 5 members per district. The best way to do this is with Single_transferable_vote (STV).

A better single-winner Voting system

Proportional Representation (PR) proponents tend to think that if STV is good for multi-winner elections, it will be good for single-winner elections also, in the form known as Instant Runoff Voting (IRV). But it isn't. That's because STV is very good at ensuring representation of a wide variety of different viewpoints but isn't so good at aggregating many different viewpoints into a single compromise.

The best voting system for single-winner elections (e.g. senator, governor, president) is Condorcet, also known as Instant Round Robin or Pairwise Voting. It uses a ranked ballot, just like IRV, but is counted differently.

  • Unlike IRV, it doesn't require recounting ballots after each elimination.
  • Unlike IRV, all ranked preferences are accounted for, not just those for the eliminated candidates.

See also