| |
Global Insight’s 2004 Presidential Election Model
Will The Economy Help or Hurt President Bush in the 2004 Elections?
Model Explainer, Current Projections and Alternative Scenarios
October 2004 Update
By Nariman Behravesh and Nigel Gault
Overview
Global Insight has been using economic drivers to predict Presidential election results for many years, following the pioneering work of Professor Ray Fair of Yale University. Predictions are based on an equation fitted over the 14 Presidential elections since the Second World War, which models the percentage share of the two-party vote won by the incumbent party. The key drivers are a combination of economic factors (income growth and the unemployment rate) and pre-determined political factors (incumbency, party fatigue and party affiliations). See the Key Election Model Factors section below for more details.
Two major themes have emerged over the years from statistical attempts to explain U.S. presidential election results. First, Americans tend to vote their pocketbooks. If the economy is growing strongly, the incumbent party has a very good chance of holding on to power. When the economy falters, U.S. voters will more likely vote for change. Second, Americans tend to favor an incumbent President running for re-election. However, if the economy is weak enough, an incumbent President can lose, as the senior George Bush found in 1992.
In the 2004 Election, President Bush will have the advantage of running as an incumbent, but his re-election hopes will hinge on the same pocketbook issues that sank his father in 1992. How will these factors affect his re-election chances? What role will they play in the overall landscape of issues, most notably foreign policy and public sentiment? Based on historical relationships, and given Global Insight's relatively upbeat outlook for the economy for the remainder of 2004, we find that Bush would normally be expected to win comfortably. However, the equation does not take into account non-economic factors such as foreign policy issues, which could prove decisive in this election. In addition, voters' perceptions of the strength of the economy might be less favorable than the equation assumes.
Election Model Track Record
The equation explains 92% of the variation in the vote going to the incumbent party, and gets most elections right. It errs in the close election of 1968 by predicting a narrow win for Humphrey rather than a narrow win for Nixon. This may reflect the turmoil in the Democratic Party over the Vietnam War. It also errs in 1976, over-predicting Ford's vote share by 2.3 percentage points. It seems likely that the Watergate scandal more than wiped out the "Republican advantage" built into the equation.
In the most recent election, for 2000, the equation correctly predicts a majority of the popular vote for Gore, but about 1.5 percentage points higher than he actually received. The equation predicts that Gore should have won the popular vote decisively enough not to face the risk of losing the Electoral College vote.
It is interesting that the equation accurately explains the defeat of the senior President Bush in 1992, through a combination of a bad economy and voter fatigue with the Republicans. Since the unemployment rate was 7.6% at the time, it finds that incumbency was a disadvantage rather than an advantage. Real income growth (weak) was also unhelpful. The "fatigue factor" also penalized Bush, since Republicans had been in office for three terms.
2004 Election Outlook: Election Model Projects Bush Victory
According to our equation, President Bush has key political factors in his favor.
- He is an incumbent Republican president running for re-election
- He is a first-termer without "fatigue factor"
Currently, the two economic factors potentially working against Bush are
- The long delay before the recovery helped to improve the labor market.
- An outlook for real per capita income growth that is positive but not impressive.
The economy is sufficiently strong, though, for the election equation to continue to project a win for Bush, using the economic assumptions of Global Insight's baseline forecast updated in October 2004.
Current Election Model Projection: Up Slightly to 55.7%
The predicted Bush vote edged up slightly to 55.7% in September from 55.5% in August. While the unemployment rate remained at 5.4%, a slight increase in forecasted income growth (from 1.0% to 1.2%) worked to Bush's advantage.
Alternative Election Scenarios
The election equation implies that this remains Bush's race to lose. There is now very little difference between the optimistic, baseline, and pessimistic cases. That is because the equation is based on economic variables as of Q3 2004. Since we are now deep into the third quarter and have a good idea of where these variables will come out, there is not much room for variation.
Pessimistic Scenario: A Somewhat Tighter Election
What happens if the economy performs worse than the Global Insight baseline? Using slightly more pessimistic assumptions (0.9% real per capita income growth instead of 1.2%, and an unemployment rate of 5.5% instead of 5.4%) reduces the projected Bush vote to 55.3%. Admittedly, a Bush defeat under these assumptions would require a bigger equation error than in any of the previous fourteen elections, but it is not unthinkable.
Bush Nightmare Scenario: Less and Less Plausible
It would require a combination of per capita income growth below zero and an unemployment rate of around 7.0% for the equation to deliver a Bush defeat—with a voting share of 49.9%. The election is now so close that economic numbers this bad are almost inconceivable.
The Impact of Non-Economic Factors
While the Nightmare Scenario is becoming ever more unlikely as the clock ticks and the election nears, Bush could still be vulnerable even under a less extreme outcome. A weaker lead from the economy combined with trouble on other key issues—the two most significant being foreign policy and public sentiment—could tip the balance against him.
Public Sentiment: Voters may not be impressed by an unemployment rate around 5.5%, since the rate was 3.9% in November 2000. A 5.5% rate is not high by historical standards, so it does Bush little damage in the election equation. But if voters react to the change in unemployment during his Presidency, then Bush will be more vulnerable. There is also an increased sense of insecurity about the labor market as manufacturing jobs have only recently turned higher, while new groups of workers in the service sector have discovered that their jobs too can be exported via outsourcing. If voters respond to these fears and to the loss of 1.2 million jobs since January 2001, then Bush may be more at risk than the equation suggests. Downbeat public sentiment certainly played a role in the senior President Bush's defeat and it could do the same for the current President Bush.
Foreign Policy: The situation in Iraq with the investigation into pre-war intelligence issues as well as post-war problems pose further dangers to Bush. The election equation suggests that President Bush is unlikely to lose on the economy alone, but it does not factor in foreign policy issues. The less impressive the recovery, the more vulnerable Bush will be to defeat over foreign policy issues.
More information on Global Insight's Presidential Election Model—including the model's forecasting track record—can be found in Global Insight's Presidential Election Model Explainer.
For more information, or a copy of the Presidential Election Model Explainer, please contact info@globalinsight.com.
Related Highlight:
Bush or Kerry? Regardless, Taxes Are Set to Rise
|