DOES the economic situation prior to a presidential election affect the outcome? It sure does, says Roy C. Fair, a Yale University economics professor. He has constructed an econometric equation for predicting presidential elections, using data on inflation and economic growth, that he says has proved ``remarkably accurate.''
Indeed, this forecasting device, when applied to the economic data prevailing before presidential elections back to 1916, makes only three errors. In each of those instances the elections were extremely close. What does the equation say about the fall election? Well, it depends upon your economic assumptions. But unless the economy comes up with a major surprise in the second and third quarters - say a full-fledged recession - the presidential race is still very much of a horse race.
``Both parties [Republican and Democratic] ought to work very hard, because they both have a very good shot at it,'' says Dr. Fair.
Fair's equation employs two numbers that indicate how the economy is affecting the welfare of voters. One is the growth rate of per capita real gross national product (GNP) at an annual rate in the second and third quarters of the election year. The second is the inflation rate for two years before the election.
To look at the forthcoming election, Fair assumes the inflation rate will run about 3.4 percent. Because most of the two years is already past, that rate is not likely to change much, says Fair. He uses the Blue Chip consensus of about 50 economists for the GNP forecast. That is 1.9 percent. When corrected for population growth, the number comes out to 1 percent per capita GNP growth. With those numbers, the equation predicts a vote for the Democrats of 49 percent of the two-party vote. (Third parties are ignored.) That would mean a narrow Republican victory. But the equation has an average error of 3 percentage points. So the election could easily go either way.
Assuming that inflation doesn't change much, Fair's equation calculates that it would take a per capita economic decline at a dramatic 6 percent annual rate in the next two quarters to ensure the Democrats of victory. A 3 percent drop would give the Democrats a ``good chance.''
On the other hand, if per capita growth goes up to 6 percent, the Republicans would have victory pretty well sealed, Fair reckons. Fair himself guesses the economy might grow at an annual rate of half of 1 percent or 1 percent faster than the consensus. But this still isn't enough to ensure Republican victory.
Fair isn't saying that the economic scene is the only factor determining presidential election results - only that it is highly important. His equation made its biggest mathematical error in 1964, when it predicted that President Lyndon Johnson would win with 54 percent of the vote over Republican Barry Goldwater. The actual victory was at 61 percent. The equation didn't account for the extremist charges laid against Mr. Goldwater.
The three close elections where the equation actually forecast the wrong winner was the Carter-Ford election in 1976, the Nixon-Humphrey election in 1968, and the Kennedy-Nixon battle in 1960.
For journalists the closeness of the election, according to the Fair equation, may be a good thing. ``Otherwise,'' he says, ``you could write one article and forget about it for the rest of the year.''