The democratic ideal is to let every citizen eligible to vote have the representation of his or her choice. But we know at a mere glance that a large part of the electorate must be voting for electoral losers, meaning they don’t get truly represented.
Define rate of representation as the percentage of voters that get what they want in their own geographical representation. Districts are formed to ensure that local issues are brought to the collective government. But if your local issues are brought by someone with the wrong ideology, you’d probably rather have that legislative seat empty. And if people with the right ideology serve other districts, you still have no one to deal with your issues the right way. For true democracy, we should have the goal of letting (nearly) everyone have combined geographical and ideological representation in one legislator. But we should also still have the goal of making sure that the legislature as a whole looks like the overall ideological split of the country.
I’ve collected information on the last four senate elections (2008-2014), and it turns out that the winners, without considering party, consistently get about 57% of the votes (it ranged from 56.58% to 57.79%). This is true across a span of 8 years and with 3 different sets of states, varying widely in population. That means, just as consistently, that 43% of voters don’t get representation in the Senate.
I started with that example because I have hard numbers, and because states can’t be gerrymandered. But now let’s look at what can happen with districting thrown into the mix. To take ideology out and treat things fairly, I’ll just talk about parties A and B. Say a state gets 10 representatives, and has a party split of 55% A’s to 45% B’s.
If the A’s control the districting, they may be able to gerrymander every district to look like the state, 55-45, meaning they would win all 10 seats, and the 45% of the population that likes party B gets no ideological-geographic representation. That gives a representation rate of 55%, and the overall representation of the state is highly skewed from the state-wide party split.
If laws require making as many competitive districts as possible, they might be able to get 9 districts that are 50-50, and the remainders go into a 100% A district. The competitive districts will each have a 50% representation rate, and the tenth district will have a 100% rate, for a state-wide rate that’s still 55%. The overall representation is much better. Since the 50-50 districts essentially pick randomly, we can figure they’ll average 4.5 representatives from each party, and we end up with a perfect 5.5 A’s and 4.5 B’s.
Another option is to make equal numbers of perfectly safe seats, and make the rest average out among themselves. Each party would then get 4 districts and the other 2 districts would be 75-25 in favor of the A’s, making them very safe seats too. So only the 25% of those last 2 districts don’t get what they want, which adds up to only 5% of the population—a representation rate of 95%. And our overall representation is 60-40, reasonably close to the affiliation rates of 55-45. However, the people that don’t get what they want are always of the same party, and this promotes civil unrest. At least when you have competitive districts, the voters in the two parties alternate in getting screwed out of their legislative influence.
If party B somehow controlled the districting, they might give themselves as many favorable districts (say, 55% B’s to 45% A’s) as possible. They could get 8 such districts, with the other two being 95-5 in favor of the A’s. The representation rate is only 63%, which is strangely better than when the majority party gerrymanders, but still not very good. And the statewide outcome is a horrible 20-80 instead of 55-45. This is the great danger of letting a minority party engage in gerrymandering.
The problem in all of this is the single-member district system. I’ve laid out how to let most of the voters have true representation in another blog post, but I came up with an idea for the U.S. Senate, because it’s guaranteed to be undemocratic by the Constitution, and we would have to amend the Constitution twice to get that changed.
A More Democratic Senate. Since every state has 2 senators anyway, we could elect them not only at the same time, but together, choosing the first- and second-place candidates. Because second-place candidates very consistently get another 39% of the votes (ranging from 39.10% to 39.42%), totaling about ninety-six and a half percent. So only about 3.5% of voters would not get their first choice to represent them.
Once this system got going, there would be room for a third party, which would naturally lessen the proportion of the population that gets its first choice. That’s when a movement should arise pushing for instant-runoff voting in the Senate elections. Minor candidates would be eliminated until only 2 remain, meaning that there would be no way for a candidate to spoil the election, and the percentage of the population that gets its choice of representation would technically become a full 100%.
Since the ratio of votes for first and second place is about 3 to 2, there should be some aspect of the system that reflects the difference. Working within the existing setting of American politics, I would propose that the first-placers get 8-year terms, and the second-placers get 4-year terms. But for that to work, there would have to be a third senate seat with an 8-year term, so that there would be elections staggered by 4 years.