Topic Notes
Intro
- Back to the year 2000
- Bush and Gore are competing to become president
- There is a recount in Florida
- Ralph Nader was on the ballet
- Bush beat Gore by 537 votes
- Nader had received 97,421 votes
- A plurality would have preferred Gore - accounting for maybe 40,000 votes
Arrow’s Impossibility Theorem
- 1951 - Kenneth Arrow published Social Choice and Individual Values
- The theorem shows that ranked voting will run into problems with more than two candidates
- What is ranked voting? FPTP: First Past The Post & Ranked Choice
- There are three problems that can’t be eliminated
- Spoiler Votes (All like X over Y, does the group?)
- Moderate-cum-dictators (In a polarized voting block, the middle makes the choice)
- Third party scandals (Everyone feels the same about the top two candidates but number three just got busted)
Common Voting Systems
- FPTP voting like we have in most places in the United States. We’ve seen problems with it in the past.
- Ranked Choice Instant Runoff
- Ranked Choice Voting / Alternative Vote / Instant Runoff
- Australia, Fiji, New Zealand
- Both Ireland and Northern Ireland
- US Locals
Other Voting Systems
- Borda - Ranked Choice joins forces with Addition. Add ranks, lowest wins.
- Condorcet Method - Simulated election pairs. Shulze Method (widest path)
- Bucklin Voting - Rank choices (1st, 2nd, 3rd, etc). If you have a majority, stop. If not, add the next lowest level of preference.
- Approval - Check the boxes of whoever you like. Really, as many as you want
- Score Voting - Like Yelp or Amazon Reviews
Making a good voting systems
- Avoid or make the problems Arrow described rare or impossible
- Can be counted manually
- Is easy to understand (improve legitimacy)
How do they Stack up?
- FPTP
- Spoiler problem is really common
- Manual counting is easy
- Very easy to explain
- Ranked Choice
- Less chance of a spoiler
- Counting is super hard
- Complicated to understand
- Can lead to some weird results which leave someone who is less popular to be the winner
- Borda
- Has a reverse spoiler effect - a bad candidate who is close in position to another bad candidate will be helped
- Counting is harder than FPTP but not hard
- Fairly easy to understand
- Condorcet
- Can lead to a loop or cycle. Might be fine if the vote can happen again
- Counting manually is very difficult
- It is hard to explain
- The Shulze method gets to crazy levels of difficulty to explain/count.
- Bucklin
- If the majority likes a group of candidates, one member of that group will winner
- Counting is medium difficulty
- Has been used in many elections in the US for things like primaries. Some versions were found unconstitutional
- Explaining is medium difficulty
- Approval Voting
- No spoiler effect or minority dictators
- Very easy to count
- Easy to explain
- Score voting
- No spoiler effect or minority dictators
- Counting is harder than FPTP but not hard
- Harder than FPTP and Approval to explain but not hard
- People may feel like they need to distribute their score to express themselves fully turning into a kind of Borda vote for them.
Where do I fall?
I am in favor of electoral reform. I think that many of these systems would be an improvement. My preference would be Approval Voting.