For the past two US presidential elections I’ve compared the results of the Iowa Electronic Markets two days before election day.

In 2008 and 2012 I posted what the market predicted and waited to see how it would do. In each of those elections the market predicted accurately within a few percentage points of the actual vote.

In fact the IEM has an excellent track record going back to 2000 when it accurately forecast Gore winning the popular vote and Bush winning the election.

The dynamics of this year’s election have been unusual and many polls in the primaries proved to be inaccurate. Can the IEM keep its streak even in 2016?

Let’s review the past numbers and then see what the IEM says today, two days before election day.

Two days before the 2000 election

Who will win
Al Gore – 27.3%
George Bush – 74.9%

Vote Share Prediction
Al Gore – 49.6%
George Bush – 49.4%

Actual result
Al Gore – 48.4%,
George Bush 47.9% – Winner

Two days before the 2004 election

Who will win
George Bush – 55.2%
John Kerry – 45.3%

Vote Share Prediction
George Bush – 51.7%
John Kerry – 48%

Actual result
George Bush – 50.7% – Winner
John Kerry – 48.3%

Two days before the 2008 election

Who will win
John McCain – 11.2%
Barack Obama – 88.3%

Vote Share Prediction
John McCain – 47%
Barack Obama – 53.5%

Actual Result
John McCain – 45.66%
Barack Obama – 52.92% – Winner

2 days before the 2012 election

Who will win
Mit Romney – 29.1%
Barack Obama – 71.8%

Vote Share Prediction
Mit Romney – 48.0%
Barack Obama – 52.8%

Actual Result
Mit Romney – 47.2%
Barack Obama – 51.1% – Winner

2 days before the 2016 election
Who will win
Donald Trump– 29.3%
Hillary Clinton – 70.5%

Vote Share Prediction
Donald Trump– 45.3%
Hillary Clinton – 57.5%

Actual Result
Donald Trump- 46.1% – Winner
Hillary Clinton- 48.2%
[[For the first time only one vote share was correct, and winner not correctly predicted]]