Predicting with partisanship: VA and NJ 2017 elections

Posted by Madeline Brown on November 15, 2017

Virginia and New Jersey were among the states to hold elections for their state legislatures this past Election Day. The results from Virginia, in particular, have been called surprising, but of all contested state legislature seats between the two states, partisanship predicted over 92 percent of the winners.

Partisanship is a measure of how much a given area will tend to vote for someone from one of the two-major parties over the other, all else being equal. It is measured entirely by how those voters voted in the presidential election compared to how the two major party presidential candidates did in the national popular vote.

For example, if Hillary Clinton outperformed her national percent of the vote by two percentage points in a district, we would say that district favors Democrats by two percentage points. It would be a 52 percent Democratic district, and all else being equal, a Democratic candidate can expect to get about 52 percent of the two-party vote there. Candidate quality, campaign spending, and other factors can make some difference, but partisanship remains the best predictor of who will win most congressional and state legislative races.

Virginia’s House of Delegates has 100 seats, all of which were up for re-election in the November 2017 election, though 33 candidates ran unopposed. In the 2015-2017 legislative session, Republicans controlled the state house 66-34, meaning that Democrats needed to pick up 17 seats to take control of the legislature. As of 11/4 at 5pm, Democrats had picked up 14 of the 17 needed seats, with three races still too close to call. There were no State Senate races in VA in 2017.

Though some are calling this “democratic wave” an upset, in reality, these results were largely predictable, based on partisanship. Of the 60 contested and called elections as of 11/10, only three Virginia House districts voted with an opposite lean from their 2016 partisanship. Incumbent Cheryl Turpin (D) defeated N.D. Holcomb (R) 51-49 in District 85, which is only a 48.7 percent Democratic district. Incumbent Robert Bloxom (R) fended off his challenger Willie Randall (D) 52-48 in a District 100, which is a 50.3 percent Democratic district. And finally incumbent Tim Hugo (R-40) could be the third to win in a district that has an opposite partisanship (54.4 percent Democratic), but the race is one of those that is currently too close to call.

infogram_0_a88b6671-ddd1-4300-8e97-1962b3bf0217Virginia State Legislature Raceshttps://e.infogram.com/js/dist/embed.js?yKftext/javascript

In New Jersey, 2016 partisanship was predictive of 86 percent of State General Assembly and Senate races. Only four general assembly races (Districts 1, 3, 8 & 21) and 6 state senate races (Districts 1, 2, 3, 8, 21 &16) were won by a candidate from the opposite party of the district’s partisan lean. It is worth pointing out that New Jersey uses the same districts for both its state general assembly and senate, electing 1 senator and 2 assembly members per district.

Voters in New Jersey use block voting in multi-winner districts to elect members of the state’s general  assembly. Block voting is the most common voting method in multi-winner elections in the United States. Under block voting, voters may vote for as many candidates as there are seats available. So in each of New Jersey’s assembly districts, voters vote for two candidates each, and the two candidates with the most votes win. This is a winner-take-all system - a single majority group always has the power to elect both seats.

New Jersey’s ballot design probably reinforces this trend, encouraging partisan voting by listing the party nominees in separate columns as shown to the right. It should therefore be no surprise that every single district seems to have elected two assembly members of the same political party (though District 8 remains too close to call). Indeed, 38 out of 40 districts also elected a senator of the same party as well (only districts 2 and 16 elected opposite party senators).

This multi-member district system in New Jersey has been beneficial for women’s representation in the legislature - women currently hold 39 of 120 seats (33%) in the state legislature, following a net increase of three women after Tuesday’s election. These figures do not include District 8, which is currently too close to call, but the two front runners are both men. The use of winner-take-all block voting, however, has lead to most of them being single party rule. A transition to ranked choice voting would alleviate some of the party monopolization of districts.  

infogram_0_21a1f97a-8afb-4b46-a863-7aff382c4969New Jersey State Legislature Raceshttps://e.infogram.com/js/dist/embed.js?KPZtext/javascript

The results from the New Jersey and Virginia state legislature elections demonstrate the predictive power of partisanship in our current elections. Even in states with so many close races, partisanship still predicted over 90% of winners. FairVote projects U.S. House races two years in advance using partisanship in our biennial report, Monopoly Politics 2018. Reforms such as ranked choice voting and redistricting plans like those proposed in the Fair Representation Act would help increase competitiveness in elections, and could be implemented for state legislatures, too.

 

Show Comments
comments powered by Disqus

Join Us Today to Help Create a More Perfect Union