Should John McCain and Barack Obama be keeping a close eye on the stock market? Recently, a friend told me that the price performance of the Standard & Poor's 500-stock index during the three calendar months leading up to the U.S. Presidential election was a good predictor of whether the President or his party would be re-elected or replaced. Basically, the idea is that a positive performance by the index signals a positive result for the incumbent or the party in power; a negative performance by the "500" means the opposite.
Intrigued, I decided to perform my own back test. I found this election-prognostication technique did an excellent job, recording a 79% accuracy rate in predicting the re-election of the party in power and an 83% success ratio in calling for a change of party.
Take a look at the accompanying table. There have been 20 Presidential elections since 1928; 2008 will mark the 21st. Based on the S&P 500's performance in the three-month period of August through October, this indicator correctly predicted re-election or replacement 80% of the time.
The Market Was Wrong in 1932
Digging a little bit deeper, I found the index's performance also did an excellent job of forecasting if the existing person or party would be re-elected. In these 14 occurrences, the model was correct 11 times, or 79% of the time. It was wrong in 1932, since the S&P 500 gained more than 14%, yet President Herbert Hoover was still voted out of office. Maybe investors were encouraged that Franklin D. Roosevelt would be elected and anticipated that "happy days would be here again."
The market was also wrong in 1968 when the market's action predicted that Hubert Humphrey would be the first vice-president elected to the Presidency since Martin Van Buren. Of course, third-party candidate George Wallace stole enough Democratic votes to make Richard Nixon pleased as punch.
Finally, in 1980 the model took one for the Gipper, as it pointed to a Jimmy Carter re-election only to defer to "voodoo economics" and the Reagan Revolution.
Everybody Still Liked Ike in 1956
The model's ability to identify changes in political parties that occupy the White House was supreme, as it was correct five of six times, for an 83% success ratio. The only time it incorrectly forecast a change in party was in 1956. The market's three-month decline of 7.7% did not cause the unseating of President Dwight Eisenhower by Adlai Stevenson, probably because everybody still liked Ike.
What about this time around? Obviously, it's too soon to tell since the three-month stretch has just started. Plus, there's no guarantee that what worked in the past will work again in the future. In general, however, it seems that Wall Street rarely supports the Democratic candidate. This is curious, since the S&P 500 has posted an average annual rise of 10.7% in the 28 years that a Democratic President has held office, vs. an average 7.6% gain in the 35 years under a Republican.
What's more, George W. Bush will likely go down in history as only the third two-term President since World War II whose tenure was book-ended by recessions. The other two were Republicans as well: Eisenhower, and Richard Nixon, whose second term was completed by Gerald Ford.
The Winner Must Still Face the Issues
Finally, the two mega-meltdowns in the stock market since 1945, in which the S&P 500 declined by more than 48%, occurred on the Republican Party's watch.
A frequently stated worry is that a Democratic victory in 2008 will likely result in: 1) a push in 2009 for a middle-class tax cut to help stimulate the floundering economy, which 2) will likely be paid for by a hike in the capital-gains and dividend tax rates that, 3) is expected to be retroactive to the start of 2009, and 4) as a result, trigger an yearend sell-off in equity prices in an attempt by investors to accelerate capital gains that will be taxed at lower rates.
We think that no matter which party is ultimately voted into the White House will regard the recession, oil prices, health-care costs and availability, and defense spending as major issues to be addressed.