The big question this fall – now that Iowa State has beaten Iowa in football– is who will win the November presidential election.

Everyone expects the economy to play a deciding role, but seven weeks out, the signs are mixed.

At the close of the Democratic National Convention, National Public Radio reported that two econometric models were giving conflicting signals.

One model created by Moody’s Analytics was tilted toward a victory by President Obama, largely because many swing states, including Iowa and Ohio, are doing better than the country as a whole. Another model, created by retired academic Douglas A. Hibbs Jr., gave the edge to Mitt Romney based largely on recent changes in disposable income.

Just for fun, I looked at two economic indicators that are familiar to all and which might seem to be harbingers of voter sentiments, the stock market and the unemployment rate.

I looked at re-election campaigns beginning with Dwight Eisenhower in 1956. Specifically, I looked at what the Dow Jones Industrial Average and monthly unemployment rates were for three specific days: New Years Day, Labor Day and the day before the election.

The first set of numbers for 1956 shocked me, because the Dow fell a little more than 1 percent between Labor Day and election day when Eisenhower won by a landslide. A longer view showed a slight gain in the Dow between New Year’s Day and election day, but it was clear there was little connection between the market and that election.

Interestingly, Bloomberg News made this recent comparison between 1956 and today: “Economic growth this year is below normal for the post-World War II era, though stronger than when Dwight Eisenhower was re-elected in a 1956 landslide.”

The stock market, it appears, is not a good predictor of elections. If it was, Carter should have won in 1980 and George W. Bush should have lost in 2004. Still, it can be argued that market declines in 1976 and 1992 were harbingers of the losses by Gerald Ford and George H.W. Bush.

The employment rate is a somewhat better forecaster for elections, although it’s not perfect either.

If you look at the unemployment rate during every re-election from Eisenhower forward, you will find only one inconsistency. President Ford was the only incumbent to lose despite employment gains. Carter and George H.W. Bush fought rising unemployment and lost. For every other incumbent, including Obama, unemployment has trended down during re-election years.

If the economy is a fickle predictor of elections, is there another crystal ball that we could be looking at?

The answer is yes, and it’s right here in Iowa.

The Iowa Electronic Markets were created by the University of Iowa Tippie Business College during the 1980s as an online futures market where investors can buy and sell shares based on how they think an election or other events will turnout.

The market has traded shares in the current presidential election since July 2011.

Obama opened a lead last October that widened and then narrowed, but which has been widening again since mid-June, with 68 percent of the invested money now backing the president’s re-election.