Center for Election Science Comments on Electoral Reform Act of 2012 – Integrated Into Drafts 3.8 & 3.9


Below are comments on each of  the elements of the proposed Act:

01 Process I see the idea of receipts mentioned. There is a security procedure for paper ballots that uses receipts called the Rivest-Smith System. There’s a simplified version of this called the Twin system. See here:

02 Ballot Access This is a good reference for thinking about ballot access: Duopoly: How the Republicrats Control the Electoral Process.  Many of the books I’ve read suggest signature requirements just equate to the money needed to pay people to collect signatures. But the recommendation to just make the requirements the same for everyone probably indirectly addresses many ballot access issues in itself.

03 Voting IRRV is certainly a more descriptive name than Condorcet, which is what I assume you’re referring to. Be aware that there is not always a Condorcet winner. Think rock-paper-scissors. Also, the Condorcet algorithm is a real pain to do with paper ballots. And it gets super complicated when you need to find a winner and there is no clear Condorcet winner. Schulze methods are a prime example of this.

But, believe it or not, there’s a very simple method that is quite good at also selecting the Condorcet winner when one exists. This system is Approval Voting. You vote for as many as you wish (no ranking). Most votes wins. Approval Voting also behaves superbly with third-party candidates by giving them an accurate measure of support. This accurate support measure is important because even if they lose, they’re not marginalized. And you need that legitimacy for these candidates to practicably exchange political dialogue. Approval Voting is one of the very few systems to be immune to vote splitting and always allow candidates to vote their honest favorite.  Gaming the Vote: Why Elections Aren’t Fair (and What We Can Do About It) , by William Poundstone is a good read for single-winner systems; it does not cover proportional representation in larger bodies.

04 Debates The standard that debate hosts have for debates is to use objective criteria for allowing candidates to participate. For US President, the Commission for Presidential Debate’s standard is an average of 15% over five national polls. Plurality polling destroys support for alternative candidates. And, pollsters ask the wrong question. The question should be: which candidates would you like to see in the debate? This lets voters choose multiple candidates and asks a more pointed question. Polling using Approval Voting would get a more accurate measure as well. Some reasonable threshold can be used from there. The best reference for debate issues is, No Debate: How the Republican and Democratic Parties Secretly Control the Presidential Debates by George Farah.

07 Representation This seems a roundabout way of asking for proportional representation. Voting systems can achieve this for you when you use at-large districts (at least five members each) that use a proportional voting method. States can use this for their legislatures and cities can use this for their councils (as they have before). Federal law requires US House seats to use single-member districts so a federal law is required to change this. The US Constitution places a barrier for PR systems with US Senate seats.

Here are some examples of systems:

-District based (I recommend five or more members per district)
– Limited Voting (semi-proportional)
– Cumulative Voting (semi-proportional)
– Single Transferable Vote
– Proportional Approval Voting
– Reweighted Range Voting Asset Voting
– Whole Legislature Methods
-Closed Party List Open Party List

Party-list systems tend to be the most proportional, though any PR system will put you light years winner-take-all systems. The Sainte Laguë algorithm is the fairest for party-list systems of those most commonly used. For reference, Real Choices/New Voices, by Douglas Amy is really good. He also has a good site on PR systems here.

08 Districts Using single-member districts brings inherent problems even when you use computers or independent commissions. It’s just an incredibly poor sampling procedure to determine voter will. This whole issue can be sidestepped by using proportional methods with districts using at least five members. This is immune to gerrymandering.

Also, if you simply must use districts or you are wanting to do lines for five-member districts, I’d recommend the split-line algorithm which can be found here.

Single-member districts are also prone to false majorities. That’s when a party with minority support gets the majority of seats. This happens about half the time with systems that use single-member districts and just over 10% of the time with PR systems. For example, Canada’s last election gave their conservative party the majority of the seats even though they had less than 40% of the vote. Canada has also had independent redistricting commissions for over the last 50 years.

I hope you find this information useful. As a disclaimer, our group has no position on any Occupy groups since we are nonpartisan. But we are happy to share information on voting theory and election-related issues with any group.

Aaron Hamlin
Parliamentarian, Director
The Center for Election Science