Topic 7
Voting and elections

We have discussed the ways in which electoral institutions affect party systems.
We have also discussed some examples of two party systems in US history, and analyzed one of the changes.
Now let’s address the question of why individual elections turn out as they do, focusing on presidential elections.

There are multiple causes of election outcomes, but in terms of individual voting decisions, the causes can be condensed into four main ones.

Party identification (See Light, pp. 221-226, box 6-4)
 

Individual loyalty to a party is a baseline predisposition that may endure from election to election. Some of you can already tell us who you will vote for in the 2004 presidential election.
Candidate appeal, measured by the sum of likes minus the sum of dislikes. It doesn’t take a rocket scientist to predict that being better liked than your opponent helps. However, consider
Nixon liked more than Kennedy
Ford liked more than Carter
Carter liked more than Reagan
Obviously election outcomes are caused by multiple variables.
 
Retrospective voting. As Reagan asked in his debate with Mondale in 1984, "Ask yourself, are you better off than you were four years ago?" Key examples of negative retrospective voting:

1932 reject Republicans (economy)

1952 reject Democrats (war)

            1968 reject Democrats (war)

            1976 reject Republicans (war, economy, Watergate)

            1980 reject Democrats (economy)

 Issues and the median voter. Why elections rarely revolve around issues (See S&B, pp. 83-91)

Three assumptions of median voter theorem:

Issues are all on a single dimension. (This is the same as the single-peakedness condition explained in S&B, chapter 5).
Each voter votes for the option closest to her ideal point.
Each of two candidates chooses an issue position with the goal of winning. Definition1     2     3     4     5     6     7     8     9
. Winset of x: the set of points preferred by a majority to x.
Winset of 1: all other points except 9
Winset of 2: 3,4,5,6,7
Winset of 3: 4,5,6.
Winset of 4: 5
Winset of 5: empty.
The position of the median voter is 5.
This is a majority rule equilibrium.
It is a Condorcet winner.
There can be no cycles on a single dimension.

Illustrations:
Election of 1960 (between Kennedy and Nixon)
Elections of 1964 (between Johnson and Goldwater) and 1972 between Nixon and McGovern.
Other elections.

Reprise:
What has Monica Lewinsky has taught political science?
Or why have Clinton’s approval ratings remained so high?

THE ELECTORAL COLLEGE

The president and vice-president are elected not by popular votes but by an absolute majority of votes in the "electoral college."

That is, 538/2=269 + 1= 270 votes.

It is still the case that the votes in the Electoral College are allocated by a formula wherein each state has a number of votes equal to its membership in the House and Senate, e.g. CA: 52+2; PA: 21+2;WY: 1+2.

This obviously violates the one person, one vote norm.

Almost all states allocate their electoral votes as a bloc to the winner of a plurality of popular votes in their state.  (e.g. if CA had a result of 48% Bush, 42% Gore, and 10% for Nader, Bush would get all 54 votes.  This would happen even if all Nader voters preferred Gore to Bush.)

It is obviously possible for a candidate to have fewer popular votes than another, but to win in the Electoral College.  This happened in 1888, when incumbent president Grover Cleveland got more popular votes than Benjamin Harrison, but Harrison got more electoral votes and won the office. It hasn't happened since.

How bad is this?  Consider the ways in which direct election can choose a Condorcet loser.

When there are more than two candidates, there is no method of election that is fully satisfactory.