The Timeline of Presidential Elections
Christopher Wlezien
University of Texas at Austin
What Happens on Election Day?
The “Internal” Fundamentals
• Most voters tend to vote for the same party year
after year.
– Party identification.
– Match between interests/preferences and
candidates’ positions.
• Even true despite the growth in Independents, that
is, because what has grown are the Independent
“leaners,” who vote very much like partisans, and the
percentage of “pure” Independents actually has
shrunk.
Referendum Judgments
• The “floating voters,” who support candidates
for president from different parties in different
years, mostly decide on evaluations of the
incumbent president.
• Referendum judgments matter even if the
incumbent is not running, as in 2016.
“External” Fundamentals
• The Big E: Economic conditions.
• Aggregate trends – “sociotropic” voting – not our
personal pocketbooks per se.
• Especially the slope of the economy – the direction and
magnitude – not so much its level. That is, are things
getting better? Are they a little better or a lot? Large
positive growth benefits the incumbent.
• Other factors
• National security.
• “Cost of ruling.”
• Reflected in the president’s approval rating, that is,
public evaluations of presidential job performance.
The Choice also Matters
• Ideological/issue proximity.
• Voters in the middle tend to reward candidates closer
to the middle, and punish those off left or right.
• The Bush-Gore election as an example.
• Other characteristics, e.g., competence.
• But, basic referendum judgments tell most of the
story.
• Is all of the attention to the nominations much ado
about relatively little – the choice voters will face in the
general election?
The Timeline of Presidential Elections
• Imagine the timeline of a presidential election.
– E.g., from last election (2012) through Nov 8, 2016.
• We have “trial heat” polls of voter preferences between pairs
of candidates at different points in the timeline.
• What do these poll results tell us about the eventual outcome
of the November election?
• What happens to change voter preferences?
• What role does the election campaign play?
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
-20-10 0 10 20
-20-10 0 10 20
-20-10 0 10 20
1952
-300
2000
2004
2008
-20-10 0 10 20
Democratic Vote in Polls (as deviation from Actual Vote)
Trial Heat Preferences for Final 300 Days of Campaign, 1952-2008,
-300
-200
-100
0
-300
-200
-100
0
-300
Days Until Election
Graphs by xyear
-200
-100
0
-200
-100
0
46
48
% Obama, of 2-Party Vote
50
52
54
56
58
Trial Heat Preferences During 2012
-300
-200
Days Until Election
-100
0
Polls and the Vote, by Days Until the Election
-200
-150
-120
-90
-60
-30
-10
-1
25
50
75
25
50
75
25
50
75
-300
25
50
75
25
50
75
25
50
75
Dem. % in Polls
Dem % actual 2-party vote
Graphs by Days before Election
Dem. % in Polls
0
Adjusted R Squared
.25
.5
.75
1
Polls-Vote Prediction by Days to Election for the
Final 300 Days: From Low to High Predictability
-300
-200
-100
0
Days Until Election
Adjusted R Squared
Lowess Fit
Polls-Vote Prediction by Days Before and
After the Party Conventions
Weekly Vote Intentions by Lagged Weekly Vote
Intentions, Final Four Weeks of the Campaign
-3
-20 -10
% Democratic minus 50%, Previous Week% Democratic minus 50%, Previous Week
-1
0
10
20
-2
-20 -10
% Democratic
0
10
20
-4
-20 -10
0
10
20
-20 -10
0
10
20
% Democratic minus 50%, Previous Week% Democratic minus 50%, Previous Week
Graphs by Week
Proportion Stable, Interview to Election
.9
.92
.94
.96
.98
Stability of vote choice over the campaign
(National Election Study 1952-2008)
-60
-40
-20
Date of interview, Relative to Election Day
Lowess Fit
Observed Stability
0
% for President's Party, Election Day Vote
-5
0
5
10
15
20
From the Final Week to Election Day:
Vote for the President’s Party
1972
1964
1984
1956
1988
1996
2004
2000
1960
1968
1976
1992
1980
2008
1952
20
5
10
15
-5
0
% for President's Party minus 50%, Final Week's Polls
Predicting the Polls and the Vote
• Two different economic indicators.
– Objective: Per capita quarterly income growth,
with each quarter weighted 1.25 times the
previous one in the election cycle (Hibbs)
– Subjective: Percent saying “Business Conditions”
have improved over the past year minus the
percent saying they have gotten worse (on a 0-200
scale, from Survey of Consumer Finances,
University of Michigan)
Income Growth and the Polls and the Vote,
April to November
Election Day
30
Mid-April
% for Pres. Party (minus 50%)
0
10
% for Pres. Party in Polls (minus 50%)
-10
0
10
20
20
30
1964
1956
1992
1996 1972
1980
1984
1960
2004
2008
19761988
2000
1972
1984
1956
1996 1988
2004
2000
1976 1968
1960
2008
1992
1980
1952
-10
1968
1964
-20
-20
1952
0
.5
1
Cumulatieve P/C Income Growth
1.5
0
.5
1
Cumulative P/C Income Growth
1.5
Perceived Economy and the Polls and the Vote,
April to November
Election Day
30
Mid-April
1956
1996
1972
1992
1980
1984
2004
2008
1988
1960
1976
2000
19721964
1984
1956
1996
1988
2004
1960
2000
1968
1976
1992
1980
2008
-20
-20
-10
1968
% for Pres. Party (minus 50%)
10
0
% for Pres. Party in Polls (minus 50%)
0
10
20
-10
20
30
1964
0
50
100
Economic Perceptions
150
0
50
100
Economic Perceptions
150
.8
.6
.4
.2
0
with Presidential Approval
Correlation
1
Correlations between Presidential Approval and Vote
Intentions and the Election Day Vote
-200
-150
-100
Days Until Election
Vote Intentions
-50
Election Day Vote
0
0
Regression Coefficient
.2
.4
.6
Predicting the Vote from Vote Intentions
and Presidential Approval
-200
-150
-100
Days Until Election
Vote Intentions
-50
Presidential Approval
0
Candidate positions and the Public
• Recall that it’s not just a referendum.
• Public preferences for policy matter.
– Public shifts to the left (right) benefit Democrats
(Republicans);
• Candidate positions matter.
– Candidate shifts to the left (right) benefit Republicans
(Democrats).
• This really registers with voters during the
convention season.
On Election Campaign Effects
• Conventions matter, and quite a lot.
• Debates not so much, if at all.
• Other events matter, but small effects, though
ones that add up to influence the final vote.
• But most of the effect of the election
campaign is to “deliver” the fundamentals.
– Internal, mostly partisan dispositions.
– External, mostly referendum judgments.
Questions to Ponder as the 2016
Timeline Unfolds
• How will the economy evolve?
• What about presidential approval?
• Which candidates will win the nominations?
• What positions will they take?
• Will the campaign deliver the “fundamentals”?
• Time and the information it reveals will tell.
Thank you