Summary
The 3 winners of the 5th Democratic debate were Yang, Sanders, and Buttigieg. Social Media Influence (SMI), our main election forecasting tool, put them in the top three, as did the on-line post-debate surveys of who won the debate, as highlighted in Table 1.
Gabbard showed mixed results, but certainly performed better than consensus expectations, which are fairly low.
The remainder logged questionable performances. Specifically, Biden and Warren, two of the recognized leaders in polls and betting markets, posted disappointing results.
Democrats still do not appear to have a clear front-runner. Traditional forecasting tools such as betting markets and polls have pointed to Biden and Warren, while our data show them struggling especially at debates. In contrast, our data has shown Yang, Sanders, and Buttigieg performing well during the last few debates.
Making matters more complicated are new entrants into the race. Though not participating in the debate, new candidates such as Deval Patrick and perhaps Michael Bloomberg and even Hillary Clinton have created a new dimension.
This changing campaign landscape as well as the aforementioned divergences between debate performance and poll numbers has produced one of the more difficult forecasting environments.
The Data
There are three different types of data in Table 1 (below).
SMI uses social media data to measure on-line influence. The next two entries, Heavy and Washington Examiner, are on-line surveys asking who won the debate. We specifically chose a left-leaning (Heavy) and right-leaning (Washington Examiner) survey sources to see if there was any difference. The last two refer to Google Search popularity during the debate and during a peak period from the next day to see how trends might have shifted post-debate when people search to find out more about a candidate.
It is important to state that SMI’s rankings are completely independent of the other forms of analysis — and that SMI provides similar yet unique insights.
Additionally, we should highlight that SMI made early and correct forecasts for the 2016 US Election, 2017 French Election, and 2018 Brazil Election – all considered extremely difficult elections to correctly forecast.
Comparative Rankings
The following table lays out the ranked finish of each of the twelve candidates according to a variety of methods.
Table 1: Ranking by Candidate’s Debate Performance by a Variety of Methods
| SMI | Heavy.com (Survey) | WashingtonExaminer.com (Survey) | Google Search, during debate | Google Search, next day | Avg. | |
| Andrew Yang | 1 | 1 | 1 | 6 | 5 | 2.8 |
| Bernie Sanders | 2 | 2 | 2 | 9 | 4 | 3.8 |
| Pete Buttigieg | 3 | 3 | 3 | 2 | 3 | 2.8 |
| Kamala Harris | 4 | 9 | 9 | 8 | 7 | 7.4 |
| Cory Booker | 5 | 10 | 8 | 7 | 9 | 7.8 |
| Tulsi Gabbard | 6 | 4 | 4 | 1 | 2 | 3.4 |
| Amy Klobuchar | 7 | 7 | 5 | 4 | 6 | 5.8 |
| Elizabeth Warren | 8 | 8 | 6 | 10 | 8 | 8 |
| Joe Biden | 9 | 6 | 7 | 3 | 1 | 5.2 |
| Tom Steyer | 10 | 5 | 10 | 5 | 10 | 8 |
Source: ZettaCap, Google Trends, Washington Examiner, Heavy
Focusing on SMI data, Andrew Yang once again prevailed on debate night. This makes an extremely impressive three debate wins in a row according to SMI for Yang. In almost any other campaign, such a string of debate wins would amount to a breakout moment. The 2020 Democratic race, however, has proven more difficult to predict due to the Democrats focus on ‘electability’.
Bernie Sanders has also posted impressive debate performances of late. According to SMI, he has finished in the top three positions in the last three debates. Also, on-line debate surveys have him in the top few. His improved performance could be linked with ‘The Squad’ officially supporting his campaign.
Using an average of all the indicators listed in Table 1, Pete Buttigieg tied for first place. His relatively strong Google search figures pulled him into the lead. Overall, Buttigieg is a strong performer and one who has received more traction of late in polls and betting markets.
The most disappointing debate performances, especially in relation to high consensus expectations, were from Biden and Warren. They are just simply not posting the type of debate performances that are expected from front-runners. SMI shows them near the bottom of the pack. On-line surveys of debate performances have them under the mid-point. Google search trends show Warren at the bottom. While Biden looks strong in terms of Google search, if you look closer at highlighted search terms they are related to his ‘punching at it’ gaffe — in other words people searched for Biden to laugh not for support.
The 5th debate showed us that those who seem to be performing well during debates are not necessarily those who are doing well in polls and betting markets. This is a trend that has been apparent for some time, and not one having started in the most recent debate.
Debate Performance Trend
Comparing the ranked debate finishes, by SMI, is instrumental in that it highlights the strength of a few non-traditional candidates as well as the weakness of a few perceived front-runners.
Table 2: Social Media Influence (SMI), Ranked Finishes of Democratic Candidates at Debates
| Debate 1 | Debate 2 | Debate 3 | Debate 4 | Debate 5 | Avg | |
| Yang | 3 | 2 | 1 | 1 | 1 | 1.6 |
| Sanders | 8 | 4 | 3 | 2 | 2 | 3.8 |
| Buttigieg | 2 | 3 | 2 | 4 | 3 | 2.8 |
| Harris | 1 | 6 | 4 | 8 | 4 | 4.6 |
| Booker | 5 | 7 | 7 | 9 | 5 | 6.6 |
| Gabbard | 4 | 1 | 3 | 6 | 3.5 | |
| Klobuchar | 7 | 9 | 8 | 6 | 7 | 7.4 |
| Warren | 6 | 5 | 6 | 5 | 8 | 6 |
| Biden | 9 | 8 | 5 | 7 | 9 | 7.6 |
| Steyer | 10 | 10 | 10 |
Source: ZettaCap
Now, just looking at SMI debate rankings during the five debates, we would assume that Yang, Sanders and Buttigieg would be currently battling it out for leadership.
Yang has finished in the top three spots in every debate, has a trend of improving finishes and has won the last three debates. In a standard election, Yang would be doing extremely well, likely leading in betting markets and/or polls.
In a similar vein, Sanders has shown a strong trend of improving finishes and has ranked second in the last two debates. We would expect him to be tightly competing with Yang for leadership.
Buttigieg has been one of the most consistent debate performers. In terms of the average of the finishes, he places second to Yang. We would also expect Buttigieg to be near the top by most metrics.
In contrast, we would expect Warren and Biden to be struggling to stay relevant, having finished in the middle or bottom of the pack during each debate.
Debates are, quite obviously, important facets to modern elections. It is almost unheard of for candidates to perform poorly during debates but pull off dramatic election-day victories.
Further it is not just SMI highlighting these trends. On-line surveys taken post-debate show that Biden and Warren have mostly logged mediocre to poor performances. During the most recent debate, in fact, they finished near the bottom of the pack.
These same on-line surveys have shown Yang, Sanders, Buttigieg, and Gabbard as being strong performers. Such data tends to confirm the debate SMI trends outlined in Table 2.
Electability
The main variable explaining the divergence between debate performance and implied performance via traditional forecasting tools (polls and betting markets) appears to be electability.
Democrats seem to be so focused on electability that those candidates who appear the most electable (presumably Biden and Warren), do not need to perform well at debates — they just need to survive them.
We’ll see how this strategy plays out. As of right now, it looks like the Democrat’s level of enthusiasm going into the general election could be diminished if the focus remains on electability.
Our impression is that this focus as well as the mediocre to poor debate performances by Biden and Warren have left the door open to a potential shake-up. If this takes the form of late entrant candidates who will radically change the landscape or of a late surge by someone not currently seen as purely ‘electable’, it is difficult to say.
What we can say, however, is that consistently poor debate performances rarely makes a strong general election candidate.