In the 2016 U.S. House election, Jim Bridenstine (OK-1) won reelection in a race with just 62,655 votes cast (or 8.1% of the district’s 2010 census population), Bridenstine won 50,595 of those votes, meaning he could have turned out less than half of his support base and still won handily. Meanwhile, Ryan Zinke (MT-AL) won his 2016 re-election bid with more than 507,000 votes cast, and earned more than five times as many votes as Bridenstein on his way to a victory that was (relatively) close. How can this be? Enter the complicated world of apportionment, demographics, and competition.
How many votes does it take to be elected to Congress?
If you’re Montana’s at-large Congressman Ryan Zinke (who has now resigned to become secretary of the interior), you need 285,358 votes in 2016 to sweep to victory with 56% of the vote or a full 36% of eligible voters. However, if you’re incumbent Congressman David Valadao in the California 21st district, you can hit a comparable 57% of votes win with only 75,126 votes, or 15% of eligible voters. These two are the candidates who won the most and the fewest votes, respectively, among the candidates elected to the 115th U.S. House who faced strong general election competition. Below is the full ranking of the members of congress with the top five highest and top five lowest vote counts.
The difference between Zinke’s and Valado’s vote counts is stark. As shown above, Valadao got little more than a fourth of the votes Zinke did, but both won moderately competitive re-election bids. Indeed, this actually undersells Valadao’s achievement, as he won as a Republican in a district in which the Clinton campaign got 55% of the vote.
The constitution mandates that states draw up their districts so that they are equal in population. So, then, why is the disparity between the most and the least votes needed to win so large?
The first reason is that while congressional redistricting only happens every 10 years after the decennial census, people are constantly moving. Montana for example has been growing steadily in population since 2010 and has almost grown enough to regain the congressional seat it lost in 1990. So there are more people in the Montana at-large district than other districts. Additionally, because Montana is currently an at-large state, the effects of population growth on district population and vote count are much larger than in more populous states that are close to gaining a seat. In Texas, which currently has 36 seats and is poised to gain at least one seat after the next census, for example, the average congressional district would have to be 3% more populous than the national average in order to merit a new seat. (100/37 = 2.7, although the actual apportionment formula is somewhat more complicated). However, since Montana is currently an at-large state with only one seat, its population needs to be double that of the average congressional district in order to gain its next, and while it has grown significantly, it is currently not growing quite fast enough to hit that target by 2020.
The second reason vote counts can vary by so much is that congressional seats are apportioned by population, not voters or even those eligible to vote. This means that both the eligible electorate and the number of voters who turn out vary greatly district-by-district. Valadao’s district is primarily hispanic, a demographic which has historically both registered to vote and turned out to vote proportionally less than the national average. Valadao’s district also happens to be in the San Joaquin Valley, a largely agricultural region which contains a high number of non-citizen immigrants. Non-citizen residents are counted towards the district apportionment population for the census despite their ineligibility to vote, and recent immigrants and other ineligible classes such as children are responsible for a significant portion of the vote count disparity.
Demographics of a district, average age, wealth, racial breakdown all play a role in driving turnout up or down, but straightforward disenfranchisement also plays a role. While felons in all but three states regain their voting rights after the completion of their sentences, all but two states prevent prisoners from voting. Prison populations are counted in population counts for the purposes of redistricting. In districts with large prison populations, this can have a significant effect on the eligible electorate. This is particularly problematic in places like Florida’s 5th district which since being redistricted in 2015 now has 5 prisons with roughly 17,000 prisoners. After redistricting, this district went from having the 10th lowest number of votes for congress in Florida in 2012 (Florida currently has 27 House districts) to the 5th, despite a turbulent election in which incumbent Corrine Brown was defeated by a primary challenger, Al Lawson.
A third reason for low vote counts is lack of competition. We found that for every 1% increase in the two-party win margin in a congressional district, 2000 fewer votes were cast. This drop in turnout is particularly severe for Representatives who ran unopposed, as unopposed races tend to see a dramatic increase in voters voting for invalid write-in candidates or not voting at all.
So how do you win a congressional election with 50,595 votes? As Oklahoma's 1st District Congressman Jim Bridenstine did, win your Republican primary and have no one declare as a Democrat or Independent, at which point Oklahoma and at least one other state, Florida, simply do away with the formality of holding a general election. Even stranger is that 50,595 votes is far more than Bridenstine needed to win to lock up the seat. His two opponents in the primary combined for 12,060 votes, meaning he could have won with 12,061 and still avoided a primary runoff. This phenomenon is not isolated to Republicans either. Florida’s 24th District Congresswoman Fredrica Wilson won her primary with 50,822 votes, and given that she had no general election challenger, would still have won her district with only 14,024 votes. By comparison, the average district size when these districts were reapportioned in 2010 was supposed to be 710,767 people. Bridenstine and Wilson were only held accountable by about 7% of their districts' 2010 census population, and could have won with about 2%.
So how can we make our congressional representatives answerable to more voters? Multi-winner ranked choice voting, as envisioned by the Fair Representation Act , would replace the winner-take-all single-winner districts with fewer multi-winner districts. In each multi-winner district, three, four, or five winners would be selected by ranked choice voting. In each district, the majority would elect most of the seats, but voters outside the majority could elect their fair share too, meaning that nearly every voter would have a representative they supported and helped elect. The total number elected in each state would stay the same, but district lines would not determine winners: voters would.
While this unfortunately does not solve the issue of single winner representation in states that are already at large, such as Montana and Delaware (State boundaries cannot be redistricted), it would make a massive difference in other states like Oklahoma, Florida, and California. Oklahoma for example, would elect its representatives at large, and would likely elect congresspeople from a broader range of political positions and demographics, since 33% of the state's voters vote for Democrats and around 30% of the state is non-white. Jim Bridenstein, Fredrica Wilson, and others like them could still win re-election, but they would need to take a lesson from Zinke, and learn how to turn out voters in a competitive, high stakes election.