Suppose the government announced that we could choose between only two breakfast cereals: Sugar Jolts and Bran Bombs. Every two years, we would choose by election between a two-year supply of each type. If you and 499,998 others in your district preferred Sugar Jolts, but the other 500,001 preferred Bran Bombs, you would all be stuck with Bran Bombs every morning until the next election.It's kind of important how we choose our leaders.
Yet, somehow, most people give little or no thought to the election process. Most people don't seem concerned that our current system is awful, and they certainly don't lobby for change. It's certainly not that no one has thought of a better system; many people have thought of many better systems. I for one enjoy reading about them, but amazingly, normal people don't seem to find the topic quite as fascinating.
There are numerous problems with our "First Past The Post" (FPTP) electoral system. Here are two of the biggest:
- The geographic restriction: this is the rule that you're only allowed to vote for someone who is officially running for office "close to you", in the same district as your house. The district boundaries themselves are chosen undemocratically, and the person running in your district might actually live somewhere else (usually near the legislature). You might love the candidate running in the north side of the city, but if you happen to live in the south side of the city you're out of luck. The geographic restriction is the main limitation on voter choice; drop this restriction and you'll suddenly have hundreds of candidates to choose from.
- Your vote probably doesn't matter: Of course, there is a bigger problem. In FPTP, if there are four people running in a district, one of them can win with as little as 26% of the vote, and it common in Canada for winners to have less than 40% of the popular vote. That means more than 60% of the voters basically go unrepresented in the parliament or legislature. There is a different sense in which your vote doesn't matter: many, if not most, races are won by more than 10% of the vote. That means that the winner could have had 10% fewer votes and still won. So in another sense, even if you voted for the winner, your vote didn't matter because he or she would still have won if you had stayed home. True, if many other people think the same way and stay home, it could change the election result, but if just you stays home, the chance that it will make a difference is infinitesimal. Finally there's a third sense in which your vote doesn't matter: you might not actually like (or know) any of the people running in your district, and if that's true, any vote you cast is kind of meaningless.
Introducing Simple Direct Representation
Of the dozens of proposals I've read, my favorite electoral system is called Direct Representation. This is a fantastic idea which might solve pretty much all the problems with FPTP, but most importantly it will solve the two problems I just mentioned. However, Direct Representation is a radical overhaul that questions everything you assume about how elections should work, so I think it's worth considering if we could have most of the benefits of DR, in a simple system that doesn't feel quite as radical.So today I'll sketch out a proposal I call Simple Direct Representation (SDR). It consists of six simple rules:
- No Geographic Restriction: During an election, you can vote for anyone who is running for the legislature in your state/province. Obviously, some thought will have to be put into how the polling station can gather votes efficiently when there are so many choices available.
- Proportional Power: your vote always counts! The voting power of your representative will depend on the amount of "poll votes" they have (votes from the official election polls). A representative that got 10,006 votes will have twice as much power as the one that got 5,003 votes; whenever a legislator votes for or against a bill, that single vote is multiplied by the number of poll votes that he or she got during the election. A simple computer program would be used to tally up votes; in case of power loss or computer trouble, the legislature could agree to allow approximations (e.g. by rounding off each member's power to the nearest thousand poll votes instead of counting individual poll votes).
- Two Choices: On the ballot you can write one or two names: a first choice and an optional second choice. The second choice will be given your vote if your first choice doesn't get enough votes to win a seat in the legislature. Your first choice should be your favorite candidate, and your second choice should be someone you are sure will win a seat. (Note: perhaps voters often won't be sure who is likely to get a seat and who isn't. In that case, the system could allow three or four choices so that you can safely write your favorite candidate as your first choice, and still have enough "extra" picks to avoid the risk of wasting your vote, which would happen if none of your picks has enough votes to get a seat.)
- Fixed Number of Winners: The legislature physically has a fixed number of seats, and those seats are filled with the people who got the most "first choice" votes. For example if there are 100 seats total, the 100 most popular candidates (measured by first-choice votes) get those seats. Then, for every voter whose first choice did not win a seat, their second choice is given an additional vote (provided that the second choice won a seat). If the first and second choice both lost, the vote doesn't count, but a voter can easily avoid this problem by choosing someone they know is popular as their second choice. In case the election concentrates power in the hands of only a few legislators, there will be some legislators that win without getting a large amount of votes; I think this is a good thing, as it can give "the little guy" a voice in the legislature without giving him serious voting power.
- Power Sharing Required: Often, people will simply vote for the leader of a party (or the person they think should be the leader). Excessively unbalanced power in the legislature is potentially a bad thing, since a politician may not behave the way voters expected; therefore, there is a fixed upper limit on the amount of voting power that one legislator can wield, for example, 3% (or N%) of all votes cast. This limit is calculated on election day and fixed until the next election. For example, if 5,000,000 votes were cast in the election, 3% is 150,000 votes. If 500,000 people voted for candidate Smith, then Smith still officially "owns" 500,000 votes but is only allowed to use 150,000 of them when he votes on a bill. Normally, Smith will use the next rule to transfer his excess votes to his friends or allies in the legislature (otherwise he would be wasting his votes).
- Vote Transfer: A legislator who is above the N% power limit can transfer excess votes (i.e. voting power) to another member of the legislature. The legislator should only transfer votes to someone he or she trusts, because the giver cannot change his mind and take the votes back. That's because revocation power could be used to get around the N% limit by giving one legislator leverage over other legislators; the legislator with 20% of all votes could say "vote the way I tell you, or I'll transfer your votes to someone who will!". All transfers will be a matter of public record, so the voters can judge in the next election whether the transfer itself, and the use of transferred votes, was appropriate.
One important thing that is often overlooked about elections is that the kind of people that run in elections depends on the rules of the election itself. For example, if election campaigns are 100% funded with private donations, different kinds of people may choose to run, compared with a system where campaigns are 100% funded with public money, or tax-deductible small contributions or "democracy vouchers" (Lawrence Lessig's idea to let each citizen redirect the first $50 of income tax they pay toward one or more political candidates).
Likewise, the electoral system itself will change the flavor of politicians who choose to run for office. FPTP is a cutthroat system—it's all or nothing, you win big or you lose big, and if there are four people running per district, 75% of them are guaranteed to lose. What kind of people would enjoy running for office in a system like that? I wouldn't. On the other hand, in an SDR system, if there are 100 seats, it's likely that (to avoid wasting time and money), less than 200 people will decide to make serious runs (with door-knocking, ad-buys and the works), so less than half of the serious candidates will lose. Minor parties who don't expect to win a lot of votes don't have to run a lot of candidates either, so they won't be scouring college campuses for kids to run "symbolically" in "unwinnable" districts. Among the "heavyweight" contenders there is a lot less drama, since there is no particular "enemy candidate" that you have to "defeat". "Going negative" won't work well, because if you convince people not to vote for an opponent, that doesn't mean they will vote for you instead. Also, an individual candidate no longer has to convince half the voters in a tiny region to vote for him, but instead can pander to a small percentage of voters in a huge region.
All these factors will change which kinds of people choose to run for office. I, for one, expect that the legislature will enjoy increased diversity of opinion and expertise. For one thing, single-issue candidates may be common--they'll run on a platform of "let's fix this one thing that's wrong with our government". If candidate McFoo wants to, let's say, change the way hospital fees work, or start a municipal broadband network, or improve privacy rights, there's no way to convince half the voters in any specific "district" to care enough about one particular issue to vote for McFoo. However, McFoo could very likely convince 0.5% of all the voters in the entire state/province to vote for him on that issue. Once in the legislature, McFoo can't directly pass any laws with his 0.5%, but he could dedicate his time to lobbying on his issue.
Another difference with SDR is that candidates can probably have a thinner skin and lower income than typical winners in FPTP. FPTP candidates often have to be prepared to spend a lot of money campaigning and accept the risk of being attacked by the "opponent", as well as the risk of losing. Winning a seat under SDR should be a more predictable affair; the amount you spend and your chances of winning are no longer tied to whoever happens to run "against" you, and a person can have a chance of winning whenever they have a fan base of some kind—any large group of people, no matter how geographically dispersed, who like that person. It's important that the fan base need not have anything to do with politics; a popular author, or actor, or university professor, could more easily run for office and win under SDR because they don't need a lot of "campaign skills", they don't need to struggle to figure out how to get a large percentage of votes in a small geographic area and they don't have to run "against" anyone, they can simply ask their (geographically dispersed) fans to vote for them. This may mean that we'll end up with fewer "career politicians" in office.
Why don't politicians care enough to change the electoral system? In part, I've already answered this question. Legislatures are filled with career politicians, who might be threatened by a new system that allows new kinds of people to win. A bigger factor is certainly the big-party advantage of FPTP. In Canada it's common for a party that got only 40% of the popular vote (or less!) to get a majority government; big parties win more seats per vote than small parties. It's obviously unfair, and most politicians probably prefer it that way. In America it's worse; Duverger's Law has taken full effect, so that no party outside the big two ever wins a federal seat. When the system favors you or your party, of course you don't want to change the system.