Economics and Election-Year Politics

Long before the coining of the infamous political catch-phrase "It's the economy, stupid," professional forecasters have tried to link the economic climate to election outcomes. Now that the candidates of the two major parties are chosen well before their summertime political conventions, economists can put forth their election forecasts side-by-side with their economic forecasts and revise them with each new round of monthly economic statistics. The economists at the Dismal Scientist website (a project of RFA, Inc., an economic consulting firm based in Westchester, PA) offer their exclusive economic model and its predictions for the November presidential election. For discussion of their current forecast and periodic revisions you can visit their website at www.dismal.com.

As of June, 2000 the Dismal Scientist model was calling for a Gore win in November with an electoral vote count of 356 to 182 for Bush. Although still substantial, this lead was down dramatically from the site's April forecast which predicted a Gore win with a margin of 411 to 127. Why the change?

The Dismal Scientist model is built on state-level data from the last six Presidential elections. The authors explain that their model, like most economic models of political outcomes, is based on the premise that a favorable economy gives an incumbent a significant edge. Their model includes a variety of economic variables such as growth rates of GDP and personal income as well as measures of employment, inflation and stock prices. The really distinctive feature of the Dismal Scientist model, according to its designers, is its state-level focus. In past presidential elections, especially 1992, they found that national economic statistics mask great variation in local economic conditions. Incorporating those local variations significantly boosts the model's predictive power. Moreover, since the presidential election is determined by votes tallied in the Electoral College, the winner-take-all approach of allowing state-level variables to determine which candidate collects the state's votes is consistent with the actual election process.

So, what caused Gore's margin to slip since April? The declining stock market, rising interest rate environment and the author's forecast for general cooling of economic activity across the states account for the decline. The authors quickly point out that a number of states are "swing"states because the predicted vote margins are quite small. Consequently, small shifts in economic conditions can generate large movements in electoral votes from one candidate to the other. Segmenting electoral votes into solid vs. firm categories, the June forecast breaks out as follows: Solid Gore, 246 votes; Leaning Gore, 110 votes, Leaning Bush, 68 votes; Solid Bush, 114 votes. Two hundred seventy votes are required for victory. Major swing states leaning toward Gore include Ohio, Pennsylvania and Florida. Virginia, Indiana and Tennessee lean toward Bush. The model puts California, New York and Illinois squarely in the Gore column. Similarly, Texas and Georgia are in the Bush column.

Ironically, the last time a Bush ran for President, it was an economy that wouldn't take off that probably cost him the election. Recall that in 1992, the invisible recovery (which had actually started in the second quarter of 1991) revved up within weeks after the election. If the Dismal Scientist model is correct, then an economy that stubbornly refuses to slow may thwart the hopes of this year's candidate Bush. Stay tuned.