Election 2016 Forecast / Identity Politics: African-American Community
Clinton will most definitely receive the lion’s share of the African-American vote. But can she actually replicate the success of Obama, the country’s first African-American president, with this demographic group? Because this community has been the most loyal of any major demographic group to the Democrats, any change in turnout and/or intention to vote Democrat will have a significant impact on the 2016 election.
Current indications are that relative turnout for the African-American community will likely decline in 2016 and a lower percent will vote Democrat than it did in 2008 or 2012. On the margin, this is obviously bad news for Clinton.
Starting by looking at the African-American propensity to vote Democrat shows just how powerful this demographic group is in producing electoral success for the Party.
Chart 1: Adjusted Percent of African-Americans 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.
There was a significant jump during 2008 which carried over into 2012. This appears to have been the Obama-effect. Eyeballing the chart, which is not scientific but works quite well at times, you can see that prior to Obama, the norm bounced around 90%. It would seem that post-Obama, we could expect it to return around this level, everything else remaining the same. Ironically, if you take the average of three well known polls (PPP, Reuters, and Economist/Yougov) which use more anonymous techniques to collect data, this percent actually is forecast at 90% for 2016. When taking the average of two polls (McClatchy-Marist and Quinnipiac) which use live telephone interviews this percent goes to a forecast of 98% for the 2016 election.
It seems like there could be social pressure on poll respondents to answer in a certain way, especially if Identity Politics becomes part of the equation. This topic will be covered later and is part of an issue of Social Desirability Bias that appears to have become excessive in 2016 as Identity Politics, increased personal transparency through social media, and media polarization have made many individuals change their answers on surveys and polls if they feel social pressure, real or imagined, to answer a certain way.
For African-American voting preference, it looks like Clinton will not receive the same positive boost that Obama received. Assuming historical pre-Obama trends take over and assuming more anonymous polling is correct, we can expect Clinton to receive something around 90% of the African-American Vote. On the margin, this is negative for the Democrats.
In terms of turnout, a similar pattern emerges. African-Americans broke multiple records for turnout during the 2008 and 2012 elections. Amazingly, African-American women became the demographic voting group with the highest turnout rate.
Table 1: Voter Turnout % in US Presidential Elections by Demographic Group, 1984 – 2012
| Black (Female) | Hispanic (Female) | White (Female) | |
| 1984 | 60.7 | 48.6 | 64.2 |
| 1988 | 55.9 | 46.3 | 62.5 |
| 1992 | 59.2 | 49.4 | 67.8 |
| 1996 | 58.1 | 48.4 | 60.6 |
| 2000 | 59.7 | 46.1 | 63.0 |
| 2004 | 63.4 | 49.4 | 68.4 |
| 2008 | 68.1 | 51.8 | 67.9 |
| 2012 | 70.1 | 49.8 | 65.6 |
Source: Rutgers University
African-American women smashed their own historical record in 2008 and became the US voter turnout leaders in 2012. But can this be replicated for Clinton? Or, did the African-American turnout spike higher due exclusively to Obama? Using the data from the previous table, the pre-Obama turnout for African-American women averaged somewhere around 60%, which is very much below the 70% recorded in 2012. Turnout will most certainly not be as high in the African-American community in 2016. The degree of the decline is still a question, but I would guesstimate that it would be somewhere between 5 and 10 percentage points. This will be a large blow for Democrats. Essentially, such a drop would bring turnout levels back down to around pre-Obama levels.
Additionally, as previously mentioned, most polls are likely modeled around the most recent election results. A 5 to 10 percentage point change in turnout for a group that was so instrumental in determining the last few elections will produce a lot of waves. In short, assuming such a drop is coming, current polls are over-inflating support for Clinton.
Lastly, it is of interest to look at potential relative turnout. Most demographic groups’ ‘interest in the election’ and ‘following the election closely’ measures have increased during 2016. These measures roughly rate enthusiasm which has been linked to voter turnout levels. We can more or less track longer term turnout trends of demographic groups by tracking these indicators. So, what would you expect to happen if one demographic group’s enthusiasm indicator makes a new high whereas another declines slightly? The most obvious conclusion is that on relative turnout, to say nothing of absolute turnout, will favor the group with increasing enthusiasm measures.
Chart 2: Percent of Registered Voters who have ‘thought a lot about the election’, Black and White Demographics Compared
Source: Pew Research
The longer term trends of how much thought each demographic group has given to the election is instructive. We clearly see that the Black and White demographics tend to move more or less in unison – with the exception of 2016. During this election cycle, the White demographic response increased, hitting a new high whereas the Black declined.
What would this imply for relative voter turnout? By taking the net, Black minus White, ‘thought a lot about the election’ and comparing it to the net voter turnout, we can see what to expect in 2016 for relative turnout.
Chart 3: Comparing (Black – White Women Voter Turnout 1992 to 2012) to (Black – White ‘Thought a lot about the Election’ 1992 to 2016)
Source: Pew Research and Roper Center
It seems like there is a reasonably good relationship. In five of the six historical cases, the net ‘thought a lot about the election’ and the net voter turnout moved in the same direction. In 2016, the net ‘thought a lot about the election’ has made a rather dramatic move by collapsing. The implied impact on voter turnout should be substantial with relative voter turnout likely seeing an inversion from what we have seen in the last two elections.
Summarizing, the results do not bode well for Democrats. The Party’s most loyal demographic likely topped out during 2008 – 2012 and such performance will be difficult to impossible for Clinton to replicate. A base scenario sees their propensity to vote Democrat and their turnout to return to around pre-Obama levels. These conclusions are backed by anonymous polls which in this emotional election cycle tend to provide more accurate information and by lower enthusiasm levels. Again, this does not envision a collapse in African-American pro-Democrat stance or in turnout, just a return to ‘normalized’ levels. But since many polls are almost certainly basing their models off of 2008 / 2012 turnout, it is already likely having a significant impact on the current race through overestimating Clinton’s support.