Election 2016 Forecast / Identity Politics: Youth Vote
The youth vote has been a consistent voting block for the Democrats. Especially in the last two presidential elections, the youth vote clearly helped the Democrats to victory. As part of the ‘Coalition’ that elected Obama, many expect the youth vote to support Clinton to an equal extent in 2016. In fact, polls using prior election data to help forecast things like turnout are incorporating this data pretty much by default.
Looking at data from 2016, however, the youth vote is a real wild card. Depending on how you want to interpret some data points, turnout of the youth vote could plummet. Additionally, depending on how you organize the data, the percent of the vote that Democrats will receive will vary greatly. When looking at a head-to-head race, Clinton will likely attract a slightly higher percent of the youth vote than Obama did in 2012 but slightly lower than he did in 2008, which seems very positive. But when including other candidates like Johnson from the Libertarian Party and Stein from the Green Party, the percentage going to Clinton in comparison to Obama’s previous victories falls dramatically.
My impression is that if there is a surprise in the youth vote during 2016 it will be that turnout will be relatively lower as compared to other age cohorts. This is mostly based on the fact that as a group the youth demographic shows exceptionally low satisfaction with the two main candidates. Sanders appears to have been the candidate that excited this age cohort the most and it further appears that he has not been truly successful in passing on this excitement to Clinton. The youth vote will still disproportionately help Clinton, but to a far lesser extent than it did Obama. In this sense, it will be a marginal negative for Clinton during 2016.
Starting with how the youth demographic leans, we can see that traditionally it has voted Democrat in presidential elections.
Chart 1: Adjusted Percent of Youth Vote Voting for Democratic Presidential Candidates
Source: Cornell University
Note: this data is adjusted so that we can compare it across elections. We are most interested here in comparing the two major parties and how different groups voted for them. The adjustment is to take the actual percent voting for Democrats and dividing by the sum of percent voting for Democrats plus Republicans. This takes out the impact of third parties.
The trend is definitely in the Democrats favor. Note however, this data is adjusted as it takes out the impact of third party candidates. The adjustment is that the percent of vote gained by the Democrats is divided by the sum of the percent vote gained by the Democrats and Republicans. In this way, we can compare across elections when there are strong third party candidates.
Using the same adjusted standard, the forecast for 2016 based on the average of seven polls (Reuters, Quinnipiac University, Monmouth University, PPP, PEW, McClatchy-Marist, and YouGov) is 67% or just slightly below the all-time high from 2008. The data from these polls seem less believable however as they vary to more of a significant degree than for other demographics. A wider range implies a more volatile outcome. So, although the average appears in an acceptable range it seems like the actual result might very well vary significantly from this point.
Looking just at the Democratic / Republican split, it does look like Clinton will hold the advantage in 2016 for the youth vote. However, 2008 and 2012 did not have potentially strong third party performance. In 2016, Johnson from the Libertarian Party and Stein from the Green Party are polling strongly with the youth demographic. In this sense, much of the benefit that Clinton could have received will likely go to the Libertarian and Green Parties.
Chart 2: Percent of Age Group Vote Intending to Vote Libertarian and Green in 2016
Source: PEW Research
The youth demographic is more attracted to third party candidates than older cohorts. Further, as the youth demographic normally votes disproportionately Democrat during ‘normal’ elections, this will more directly hit Clinton. This data is taken from PEW Research but other sources show similar voter intention breakdowns – the youngest voters more heavily support third parties.
In terms of preferred candidate, the youth vote prefers Clinton over Trump. But, once you include the Libertarian and Green Parties, much of this age group gets siphoned off. Depending on the source, it could be anywhere from around a quarter to a third of the age group could vote for a third party candidate. This is something that Obama simply did not have to contend with and something that will marginally pull votes disproportionately from Clinton.
Turning to turnout, the situation seems equally as questionable. My impression is that turnout for the youth demographic will likely decline in relative terms versus the voting populace, but the range of possible scenarios is actually quite large.
The youth demographic is actually extremely interested in the election, as are every age cohort. In terms of interest, as measured by how much thought they have given to the election, the youth demographic in 2016 seems to be much more interested in absolute terms and about the same in relative terms versus other age cohorts.
Chart 3: Percent of Registered Voters in Age Cohort who Have Thought Quite a lot about the Election
Source: PEW Research
Judging from the previous chart you might expect turnout to increase significantly. However, the youth cohort, though apparently following the election closely, is not satisfied with the candidates. This by itself might not be a red flag as many in 2016 are not very satisfied. However, the youth group seems exceptionally unsatisfied as can be seen in the following chart.
Chart 4: Percent of Voters Very or Fairly Satisfied with Candidates for President
Source: PEW Research
The decline is staggering in both absolute and relative terms. The youth demographic went from the most satisfied age cohort from 1992 to 2012 to the least in 2016. This is a very important point, so let’s repeat it again. In the previous six US presidential races youth voters were the most satisfied of all the age groups – then in 2016 they have become the least. This swing is very unusual and there must be consequences for such a wild swing. A forecast that does not include this massive swing will be difficult to accept.
Taking the same data and manipulating it differently might make the trend more obvious. Next, we take the difference between the level of satisfaction with the candidates of these two age groups, the youth or 18 to 29 and the early mature group or 30 to 49.
Chart 5: Difference between two Age Cohorts, (18 to 29 years old) minus (30 to 49 years old), Percent of Voters Very or Fairly Satisfied with Candidates for President
Source: PEW Research
The data from 2016 depicts a sharp contrast to the fairly tight range from 1992 to 2012. This type of inversion is very unusual. When a dataset is very consistent over a number of decades, it will only dramatically change if there is something equally dramatic impacting it. Additionally, it would be a novice mistake to believe that such a large swing would not significantly impact turnout. Essentially, this age group is saying that, though they normally are more positive than the rest of the voting populace, this time they are the most negative.
This is where we need to make some guesstimates. It is more difficult to forecast absolute turnout for this age group due to the aforementioned swing. It is however within reason to forecast relative turnout. From these figures, it is safe to assume that relative turnout for the youth age group will decrease in relation to the rest of the society. This statement has various implications. Third parties will not do as well as expected as their strongest age cohort will have a relatively lower turnout. Also, on the margin, Trump will benefit from a lower youth turnout as he polls much better with older demographics.
Much of the somewhat contradictory and more volatile data regarding the youth vote could be explained by Bernie Sanders popularity within this age group and the fact that he has dropped out of the race. The former rival to Clinton focused much of his effort on college campuses and the general youth cohort. By general consensus he was wildly popular in this demographic. For instance, according to Civic Youth, Clinton only won the youth vote in 2 of 27 Democratic primary contests for which data is available. Of those primaries, she won 20 of the 27 in terms of overall vote. In short, Sanders pulled the youth vote to a tremendous degree but lost much of the older demographics. Clinton appears to have had serious difficulty obtaining the youth vote in the primaries and it is questionable if they will come out in great numbers to support her in the general election.
We can assume that much of the increase in the youth demographic ‘following the race’ and being ‘more interested in politics’ is due to the Sanders-effect. He inspired this generation to become active in US presidential politics. With him falling out of the race, the youth may still be ‘following the race’ but in fact are not ‘satisfied’ with the remaining candidates. In fact, their satisfaction is much less than the rest of the country – perhaps because their preferred candidate is no longer in the race.
A counter argument to this in favor of Clinton is that the youth vote greatly prefers Clinton over Trump, so a decline in ‘satisfaction’ will not really mean much. In other words, as long as the youth relatively approve of Clinton to a greater extent, they will vote for Clinton. In some ways, this tends to be confirmed in polls as the youth do tend to declare their preference for Clinton over Trump in head-to-head polls. However, this does not mean they will actually vote for Clinton.
The youth might prefer Clinton to Trump, but if there is little inspiration to vote, it will negatively impact turnout. Support for this idea is that ‘satisfaction’ within Democratic youth is actually lower for Clinton than ‘satisfaction’ within Republican youth. According to PEW Research, only 26% of Democrats / Leaning Democrats from the age group 18 to 34 are very or fairly satisfied with candidates for president. This is actually slightly lower than the same data for Republicans / Leaning Republicans, which is 29%.
Summarizing what we know, the youth vote: (1) prefers Clinton to Trump in a head-to-head vote, (2) has a much higher probability to vote for Libertarian and Green candidates than other age groups if given the choice, (3) is ‘following the election’ closely in absolute terms as well as in relative terms against other age groups, (4) is far less ‘satisfied’ with the candidates in absolute terms but also in relative terms against other age groups, (5) greatly preferred Sanders who was extremely popular within the Democrat / Leaning Democrat youth group, and (6) the Republican / Leaning Republican demographic is more satisfied with Trump than the Democratic / Leaning Democratic is with Clinton.
Filling in the blanks, we assume that: (1) the youth vote became very interested in the election due to Sanders, (2) Sanders dropping out caused their satisfaction to fall off a cliff but did not negatively impact their interest in the race, (3) as most of the Sanders’ supporters were Democrat / Leaning Democrat the satisfaction with the candidates fell most within this group, (4) the excitement from within the youth demographic for Sanders has not transferred to Clinton, (5) many of the youth demographic have migrated to the Libertarian and Green Parties and those who remain supporting Clinton are not very inspired, (6) youth turnout will likely decline relative to other age groups and on the margin will mostly hurt Clinton.
Generally, the youth demographic has been a highlight for Democratic candidates. In 2016, however, there will likely be some negative surprises in store. Although there is contradictory data, once we pull it all together and make some informed assumptions, it seems like Clinton’s inability to fill the void within the youth demographic after Sanders dropped out will result in a decline in relative turnout for this demographic.