Elections in the US are skewed: twice in the last twenty years, the winner of the popular vote in a presidential election has lost the Electoral College vote. Steven Mulroy argues that only by adopting Proportional Representation can the US overcome current electoral challenges like gerrymandering and the “Big Sort”. He writes that countries such as Australia have successfully instituted systems such as the Single Transferable Vote and now see far more proportional elections, and that the US could do the same very easily, starting at the state level.
The presidential candidacy of former Starbucks CEO Howard Schultz has once again stoked fears of a candidate splitting the anti-incumbent vote splitting the anti-incumbent vote and handing the presidency to a candidate not supported by a majority. This would be the second US presidential election in a row in which the nationwide majority will was thwarted, since Hillary Clinton won three million more votes nationwide than Donald Trump, yet still lost the Electoral College. Meanwhile, the US Supreme Court has taken two recent cases which only underscore the extent to which partisan gerrymandering routinely warps US electoral outcomes.
The common thread in all these narratives is a fundamental flaw in the US electoral system, a flaw shared by many systems around the world. When you carve up the electorate into arbitrary subunits—states for presidential and Senate elections, districts in House elections–and hold ‘winner-take-all’ elections within each, you create the potential and likelihood for the vote to be skewed—for electoral outcomes that defy majority preference, or otherwise skew from an accurate reflection of popular will. The common solution is to get rid of (or at least minimize) single-member districts and the winner-take-all paradigm, and move to Proportional Representation through adoption of Ranked Choice Voting.
US elections are skewed
This voting ‘skew’ is most obvious with Senate and presidential elections. Each state gets two US Senators, regardless of population. Thus, a Wyoming resident’s vote counts 60 times more than a Californian’s, since California has 60 times the population of Wyoming yet still has the same representation. This bolsters the influence of sparsely populated rural states, . skewing the Senate rightward.
A lesser version of this dynamic occurs with presidential elections, where each state gets a number of Electoral College votes equal to its number of House members (based on population) and Senate members (not based on population, as seen above). This again over represents lower-population states. Since almost all States award their Electors on a winner-take-all basis to the candidate that wins a state, a candidate with more nationwide votes can nonetheless ‘waste’ lots of votes by running up lopsided victories in a few states, netting fewer Electoral votes than a candidate who racks up narrow victories in more States. This allows the nationwide ‘loser’ to still win the White House, as has happened five times in US history.
But this ‘skew’ extends beyond state-based vote allocations to all single-member districts (SMDs). Carve a state, county, or city up into SMDs, and you are bound to have a disconnect between the percentage of the vote netted by a political party (or racial minority, or any politically cohesive group) and the percentage of legislative seats that party wins. Often this is due to intentional gerrymandering by the party temporarily in power at redistricting time, after the Census which runs every ten years.
Election reformers frequently call for stricter court policing of such gerrymanders, or the adoption of nonpartisan redistricting commissions to draw district lines. These are both good ideas; indeed, the US is the only industrialized democracy that still allows incumbent legislators to draw their own district lines come redistricting time.
Photo by Matt Artz on Unsplash
But these sensible reforms won’t completely solve the problem, because of the phenomenon of “demographic clustering,” also called “the Big Sort.” Voters do not settle down neatly in perfectly mixed ratios of Democrats to Republicans. Democrats tend to over-concentrate in urban areas, while Republicans disproportionately live in suburban and rural areas. Draw contiguous, relatively compact districts over such populations, and you will naturally over-represent one party or another. While this “Big Sort” has been most pronounced in the US in the last few decades, a similar phenomenon can be seen in other nations as well. Indeed, such a geographic/demographic “skew” is inherent in any winner-take-all system using SMDs.
Proportional Representation can unskew US elections
But it does not have to be this way. We can use Proportional Representation (PR), a system in place at the national level of almost every industrialized democracy in the world. Under winner-take-all, 51 percent of the vote controls 100 percent of the power, and a consistent minority of 40 percent gets nothing. Under PR, 51 percent of the vote would control (roughly) 51 percent of the power, and that 40 percent minority would get (roughly) 40 percent of the legislative seats.
Many nations get PR by using a ‘party-list’ system, a parliamentary system in which voters vote for parties, not candidates; parties are awarded a number of seats based on their proportional share of the nationwide vote; and party leaders fill the seats their party has won from an internally pre—approved list of potential party candidates. Americans, by contrast, are used to voting for people, not parties, and choosing party nominees through primary election.
A better fit for America is the Single Transferable Vote (STV), a method of achieving Proportional Representation through Ranked Choice Voting (RCV), where voters rank their 1st, 2nd, and 3rd choices rather than just voting for one candidate. STV has been used to elect the Australian Senate for over 70 years. In America, it has been used successfully for decades in Cambridge, Massachussetts, and for years in Minneapolis, Minnesota.
STV works when there are multiple legislative seats to be filled in a single election. It takes place either in an at-large election or one from a multimember district. It sets a minimum quota of votes needed to win a seat, based on the number of seats up for election. In a 5-seat election, for example, the quota would be 1/6, or about 17 percent of the vote. Any candidate reaching the quota wins a seat. Any ‘surplus’ votes for that candidate over and above the quota are then redistributed among remaining candidates based on the 2nd choices of the ballots for that winning candidate. If any other candidate now meets the quota, she is seated. Otherwise, the candidate with the fewest votes is eliminated, and the ballots for that eliminated candidate are redistributed among remaining candidates based on 2nd choice votes. This process of redistributing ‘surplus’ votes of winning candidates, and all the votes of lowest-rung eliminated candidates, continues, drawing on voters’ 2nd and 3rd-place choices as necessary, until all the seats are filled.
Advantages of Proportional Representation
STV has proven over time to provide roughly proportional results. A politically cohesive group that nets, say, 30 percent of the vote, will end up with something close to 30 percent of the seats. Australia is a natural experiment proving this case. Since 1949, it has used single member districts for its federal House and STV for its Senate. Senate elections are consistently more proportional in outcome, with better representation for third parties, and no ‘majority party loses’ anomalies (which have occurred twice in recent decades in the House). Such differences can be seen in recent Australian elections, too.
This has numerous advantages. It ensures that the majority preference is respected, while also ensuring that significant minorities are not completely shut out. Instead of ‘winner-take-all,’ it’s “majority take most, minority take its fair share.” It eliminates or minimizes the inevitable gerrymandering inherent in SMDs by eliminating or minimizing the drawing of districts. Thus, it would permanently and decisively end the anomalous election results as we have seen in recent US presidential elections.
But STV has yet further advantages over our current system. In almost every at-large or SMD general election, the outcome is a foregone conclusion, because the jurisdiction or SMD tilts heavily toward one party or another. Elections are not competitive, which depresses turnout. In partisan elections, the only real competition is in the primaries, which pushes candidates to the extremes of left and right and discourages across-the-aisle compromise. In contrast, STV makes elections more competitive, spurring participation. The ranking system tends to elect consensus candidates with broad-based support. Ranking also encourages cooperative campaigning: a candidate wants to be the 1st choice of her own base, but also the 2nd choice of her rivals’ bases. So she won’t risk alienating voters with mudslinging attacks. Instead, she will respectfully ask to be the 2nd choice of her rival’s base. This has been the experience where RCV has been used in the US
The good news for PR proponents is that it would not take a federal constitutional amendment to adopt it. Each state legislature has the authority to adopt it for state, local, and federal elections within its borders. Congress could adopt it nationwide. Indeed, pending in Congress is the Fair Representation Act, which would provide for STV elections for the US House.
Whether at the federal, state, or local level, Proportional Representation is an idea whose time has come. It would remedy a host of fundamental US electoral flaws in one fell swoop.
- This article is based on the new book ‘Rethinking US Election Law Unskewing the System’ by Steven Mulroy
Please read our comments policy before commenting
Note: This article gives the views of the author, and not the position of USAPP– American Politics and Policy, nor of the London School of Economics.
Shortened URL for this post: http://bit.ly/2HHvxEt
About the author
Steven Mulroy – University of Memphis
Steven J. Mulroy is a Professor of Law at the University of Memphis. A former litigator for the Voting Section of the U.S. Justice Department, he has authored numerous scholarly articles on voting rights. As a law professor, Professor Mulroy has participated in the litigation of over a dozen cutting-edge cases in areas which inform his scholarship, including the challenge to the Palm Beach County, Florida “butterfly ballot” in the 2000 presidential election; the first-ever federal court injunction against a state senate’s ongoing internal election recount proceedings; and the first federal case imposing “cumulative voting” as a non-district remedy for minority vote dilution under the Voting Rights Act.
A lucid, helpful description of the benefits of STV and RCV.
I’m especially taken by the possibility of reducing rancor in elections, as the process would seem to incentivize candidates to underscore shared values, and I assume to frame any key differences in a respectful manner, to avoid making themselves unpalatable to 2nd-choice voters.
It’s also easy then to hope that today’s rancor is borne simply of the cold-hearted incentives of today’s winner-take-all mechanics. But I trust in my heart that even as this more equitable and representative election system becomes the norm our candidates and fellow voters will innovate new and exciting means to treat each other like shite.
There’s tradition, after all.
LOL. Tru dat, as the kids say.
I just had an idea hit me, inspired by your post.
The Senate is a conundrum when it comes to proportional representation, breaking the two party duopoly, and representing the people. One key reason is that the constitution requires that every state send exactly 2 representatives. But what I’ve realized is that *because states get to choose their electoral systems* …
A grouping of 5 states, say
Wisconsin, Michigan, Illinois, Iowa, and Minnesota
Arizona, New Mexico, Nevada, Utah and Colorado
… those 5 states could band together into a single multimember district which would elect 8 senators. Those states could choose to collaborate and synchronize their 3 senatorial tier elections (senators are elected for 6 year terms and every senator will either be up for reelection in 2022, 2024, or 2026).
So say by 2024, we got some states to agree to this. In this scenario AZ, CO, NM, UT and NV would be electing all of their senators together via ranked choice voting in a multimember district. If any person running for senate in that term between those different states reached a certain percentage threshold, they would get a seat.
It might work better with a larger grouping of states, but bear with me.
Nevada has 1 Senator up for reelection in 2024, New Mexico has 1, Colorado doesn’t, but Arizona also has 1 in that year. Utah has a Senator up for reelection in 2024 as well.
Every state would eventually get someone up for reelection, so Colorado would still be able to vote in this one.
So there are 4 seats to fill and anyone in those 5 states can rank their favorite candidates that are running. Since there are 4 seats, the quota threshold to get a seat would probably be something like 1/5th of the electorate voting for you, so 20%.
It’s only quasi proportional, but apparently 90% of people would see their first choice represented.
But those five states would probably always have at least one Navajo Senator, at least one Mormon Senator, at least one Libertarian senator, and at least one Democrat per senate tier electoral cycle.
And that’s just one single Senate tier. They would have 3 tiers and 10 total senators as a block.
As an aside, I’d personally prefer the US house of representatives have mixed-member proportional representation, similar to Japan, New Zealand, Scotland, etc. In that case a party list would supplement the district representatives and I’d be open to various systems for district seats because the political alignment would be proportional anyway.
Rank choice would be decent with the president, especially as an executive executing policy, a head of state, etc.
I’m glad there are states trying out rank choice voting, but it’s mainly in single-member districts. It is truly inspiring to see change, but multimember and other preproduction systems would be truly life-changing.
Pardon my typos, I was apparently drowsy. In an earlier draft of this post I had been using 4 states rather than 5, so pardon the part where I refer to 8 senators in the block, as I said later it would be 10.
I’m not married to the notion that it has to be a certain number of states, this is just the first time I’ve realized that states could band together in this manner. Though it’s probably a significant political task (prisoner’s dilemma kind of thing).
And then the other one I noticed is my typo when I tried to say ‘other proportional systems’ near the end, but it came out preproduction for some reason.
I think your proposal as it stands would not be acceptable to the smallest state (population) in any proposed multi-state district, as the voting power of residents in that state would be diluted strongly.
The senate system in the US was originally designed to protect the smaller states from being overruled by larger states.
While still adhering to Article 5, perhaps senate seats could be increased to 3 per state, and then align the elections so all 3 seats in any particular state are elected at one moment as a 3 seat proportional district.