Yesterday, the website FiveThirtyEight released a fascinating interactive on gerrymandering in America, and it demonstrates how unfair and open to manipulation the single-winner district system really is. By creating and contrasting thousands of gerrymandered districts and assembling congressional maps maximized for various criteria, they found that:
- If Republicans could redistrict the entire country ahead of the 2018 elections, they could draw 275 safe Republican districts, 139 safe Democratic districts, and only 21 swing districts.
- If Democrats could do the same, they could draw 263 safe Democratic districts, 145 safe Republican districts, and only 27 swing districts.
- A single winner map that maximized partisan proportionality would still be expected to deliver Republican control in an even year, but only by four seats.
- Maximally compact single winner maps produce more competitive districts, but preserve roughly the level of pro-Republican bias present in the current map.
- A map created to maximize minority representation would be almost identical to a proportionally partisan map, but with a slightly larger bias towards Republicans (five seats).
These conclusions align well with what FairVote has found in the past, namely that:
- The current system creates tensions between compactness, proportionality, and descriptive representation that are fundamentally irresolvable (you can only maximize one at a time).
- Single winner districts create an ability for representatives to choose their constituents in ways that can subvert the will of the people to a staggering degree. Control of the redistricting process gives you the power to manufacture landslides for either party out of an even year.
- Independent redistricting is an important issue, but cannot independently fix the crisis of polarization and gridlock that is afflicting the US Congress (Harry Enten of FiveThirtyEight has a new piece making this point today).
As part of their research, FiveThirtyEight ranked these redistricting proposals against each other across 5 criteria. In order to show how multi-winner districts with ranked choice voting bypass many of the stalemates of single winner representation, we decided to throw our own sample map for the Fair Representation Act into the ring, and see how our map compares. Spoiler alert: it outperforms every single-winner district map in all but one category. The results are below:
Some of these cannot be measured exactly the same in multi-winner districts, mainly because multi-winner districts mean that “seats” are different than “districts.” A few notes on the assumptions we need to make to compare these maps:
- Competitive districts refers to districts containing a competitive seat, rather than competitive seats. These metrics are the same for single winner maps, but since a multi-winner district can only ever have one swing seat in a two party race, are very different for multi-winner races. Under the Fair Representation act sample map, 242 seats are in districts containing a competitive seat, which is far more than any single winner map. Technically, only 67 seats are competitive between the two major parties, but multi-winner districts also open up competition in ways not captured by this metric, like by incorporating competition between members of the same party and by being more inviting to candidates from outside the two major parties entirely.
- Minority majority districts refers to minority power to elect. Again, these are the same for single winner district maps, but different for multi-winner maps. Technically, fewer of the districts are majority-minority, but with larger district magnitude, a group does not need to be a numerical majority of a district to have the power to elect a seat with their votes alone. A minority group only needs to be 16%-25% of the population to achieve power to elect under a multi-winner map.
- Compactness refers to border length rather than district size. Multi-winner districts are inevitably far larger than single winner districts, but because of this can have far more compact borders relative to their size. This is a choice FiveThirtyEight explicitly makes in ranking their maps on compactness, but compactness can be defined many different ways.
We found that our sample map performed better than all the single winner district proposals across all but one criteria. Our map was created using a method similar to what FiveThirtyEight calls Compact(borders), that is, we used Auto Redistrict to create a map that was optimally compact while respecting county borders and maintaining an equal population per seat. The main difference between our map and FiveThirtyEight's is that we use multi-winner districts with ranked choice voting, which allowed us to drastically reduce county splits and create more compact, larger districts, while preserving competitive districts and proportional seat allocation. Notably, even though our redistricting program specifically ignored minority power to elect considerations, our map is the second best performer in terms of creating districts with minority power to elect. Our map also ties with the proportional seat allocation single winner map in terms of creating the most equal distribution of seats to Democrats and Republicans in a 50-50 year, as shown below.
Partisan Balance of Seats:
Not only does the Fair Representation Act sample map performance across categories make it a clear front runner in terms of maximizing these criteria overall, it’s not at all clear which map would be in second place. The majority minority seat maximization map comes in 6th on the efficiency gap and number of competitive districts, and 7th on both county splits and compactness, meaning that while it performs best on what it is intended to do, it is below average on all other criteria.
FiveThirtyEight’s map that maximizes compactness while preserving county boundaries ranks second across two categories, but it is only an average performer across the remaining three. This makes it the probable runner up, though its weaknesses on minority power to elect and middling efficiency gap score demonstrate that compactness is not really a proxy for communities of interest, and that compactness-only maps for single winner districts still suffer from many drawbacks.
We believe our map provides the best set of tradeoffs for Congress given its history, the preferences of the American people, and the needs of a diverse and highly polarized country. For more information on this proposal, read the Fair Representation Act report, or explore our sample map at http://www.fairvote.org/fair_representation_act_report#interactive_map.