Has Donald Trump consolidated the party base less than previous Republican nominees?

The extent to which Donald Trump has consolidated the Republican base is of large importance for the nominee’s general election chances this fall. Most paths to the Presidency for the Republican nominee require him to receive votes from an overwhelming majority of the Republican base, as well as to win over traditional blue-collar Democrats and swing voters in rust belt states. Recent polls suggest he has been closing the gap with his rival, Hillary Clinton.

To examine his level of support, I use observed voter data from the primaries to compare previous non-incumbent Republican presumptive nominees to Donald Trump. I look at the percent of total votes each candidate received after becoming the presumptive nominee.

While polling data is most commonly used to predict how candidates will perform in an election, observed voter data has the benefit of noting how the electorate has actually chosen to vote.  I focus on Republican nominees because there has been a lot of discussion over Donald Trump’s ability to consolidate the Republican Party, both elected officials (http://fxn.ws/2d84SDa) and the base (http://wapo.st/2aonpWT).

A candidate becomes the presumptive nominee when the candidate no longer has any viable rivals for the nomination (https://goo.gl/HHkuUC). For Donald Trump this was on May 4th, when John Kasich suspended his campaign, and a day after the Indiana primary (when Ted Cruz suspended his campaign).

As a point of reference, Hillary Clinton became the presumptive nominee on June 6th, using the Associated Press’ count of superdelegates, or the evening of June 7th, after New Jersey concluded its Democratic primary. Democratic presumptive nominees also tend to clinch quite late in the primary season (http://goo.gl/iAumES). A candidate becomes the actual nominee when they are formally elected at their party’s convention.

In previous presidential elections, there have been many states that had yet to vote in the primaries when a Republican candidate became the presumptive nominee (what I call the ‘presumptive period’). The average proportion of the vote a presumptive nominee receives across states in the presumptive period can be interpreted as a measure of the candidate’s ability to consolidate their base.

I compare Donald Trump to all previous non-incumbent Republican presumptive nominees from 1980 onwards – Ronald Reagan in 1980, George H.W. Bush in 1988, Bob Dole in 1996, George W. Bush in 2000, John McCain in 2008, and Mitt Romney in 2012. I use the same presumptive nominee dates as Harry Enten in his June 1, 2016 FiveThirtyEight article (http://53eig.ht/1ZcZg7M).  Data for 1980 and 1988 come from Cook Rhodes’ United States Presidential Primary Elections 1968-1996: A Handbook of Election Statistics (Washington, DC: CQ Press; 2000).  Data for 1996 (http://bit.ly/2d4HHqk) and 2000 (http://bit.ly/1VQOmpP) come from the Federal Election Commission.  Data for 2008, 2012, and 2016 come from the Green Papers (http://bit.ly/2d0Rf3G), a site hosted by Richard Berge-Andersson and Tony Roza.

The presumptive period states are largely representative of the overall country. Each of the six non-incumbent Republican presumptive nominees prior to Mr. Trump had at least 20% of states vote after they became the presumptive nominee. Some had a significantly higher proportion – 37% for Mitt Romney, 39% for Bob Dole, and 49% for George W. Bush (see table 1). As well, these states were located across all four regions of the country for each presumptive nominee (other than the south for George H.W. Bush in 1988), and represent both highly and sparsely populated states.

Figure1_20160922

Nine states voted after Donald Trump became the presumptive nominee on May 4th (CA, MT, NE, NJ, NM, OR, SD, WA, WV). At least five of the nine states that voted in the presumptive period in 2016 also voted in the presumptive period for each of the six earlier presumptive nominees.

Figure2_20160922

Voter Consolidation
I detail the proportion of the vote each presumptive nominee received in each state that voted in the presumptive period (see Table 1 above).

Ronald Reagan and George H.W. Bush both received over 80% of the vote in their respective presumptive periods. George W. Bush received over 77%. When restricted to states in common with 2016, the results are similar, with each of these three candidates in the high 70’s or just over 80%. Each of these candidates went on to win the general election.

Mitt Romney and Bob Dole each received a little over 67% of the vote (or in the low 70’s when restricting to states in common with 2016). Both lost the general election. Though John McCain received a greater share of the presumptive period vote, just under 77% (or just over 80% when restricting to states in common with 2016), he also lost the election.

Donald Trump has received just under 72% of the presumptive period vote. This share of the vote is more similar to the share Bob Dole and Mitt Romney received than that received by Ronald Reagan, George H.W. Bush, George W. Bush, and John McCain.

To be sure, there are limitations to this approach. Primary voting behavior is not a perfect approximation of voting behavior in the fall, though neither is polling data of registered or likely voters too far away from the general election. Though not all states voted in the presumptive period, at least 20% of states did so for each presumptive Republican nominee from 1980 to 2012.

Voters who choose to vote in the primaries may not be representative of general election voters, especially voters choosing to vote in the presumptive period when the primary is no longer competitive. However, primary ballots consist of more than presidential nominees. Down-ballot elections and ballot initiatives are also key elements of primary elections. In addition, some voters view the act of voting as a civic duty, regardless of the competitiveness of the race.

As well, primary voters that cast ballots in the presumptive period of the primaries can vote for their preferred candidate in the spring (who might not be the presumptive nominee), while still consolidating behind the nominee in the fall, particularly if the opposing Democratic candidate is especially disliked (http://nyti.ms/20XGbqM) or suffers from an ongoing scandal (http://cnn.it/29klxSV).

While a proportion of voters will follow this pattern, a distinct section of voters will not, choosing to never vote for the nominee or foregoing voting in the general election altogether. I am unable to observe and identify these two distinct groups of voters. For instance, I cannot distinguish voters in the presumptive period that cast a protest vote against Donald Trump in the primary but will vote for him in the fall from those voters in the ‘neverTrump’ camp.

In addition, any historical comparison of presidential elections suffers from a limited sample size. That said, the historical comparison of presumptive period states is informative, as it describes the candidate’s relative popularity as compared to previous presumptive nominees.

At the conclusion of the primaries, Donald Trump’s presumptive period share of the vote was relatively low as compared to previous Republican presumptive nominees. For months afterwards polling data and lack of support within the Republican establishment, such as former Republican Presidents (http://bit.ly/2cGRzEW) and Presidential nominees (http://bit.ly/2d84gNG) opting to skip the GOP convention, indicated that he was having difficulty expanding his popularity beyond his center of support.

In recent weeks, polling data suggests the race has tightened. Yet Mr. Trump has not reached a 50% threshold of support (http://bit.ly/1KI17z9), though neither has his opponent, Hillary Clinton, consistently done so. Their lack of popularity is driving a resurgence of interest in third party candidates, which could imply that many swing voters in key states may already be out of reach.

If so, this could explain Mr. Trump’s recent efforts to solidify support amongst his base in rural areas rather than court voters in larger population hubs (http://abcn.ws/2cppiBO). His ability to increase enthusiasm for his candidacy amongst this group of voters over these last two months may well determine who ascends to the office of the Presidency this January.

Robert Nathenson is a social scientist who applies quantitative methods to the study of health, education, politics, immigration, and sports.