Pocketbook Predictions of Presidential Elections, 2012

The Anderson-Geckil "pocketbook" model is estimated using data from a variety of sources covering economic conditions and institutional factors in the United States from 1916 to 2008, and predicted economic conditions for 2012. The econometric techniques used include standard linear regression as well as more advanced weighted, panel, and robust regression techniques. An extensive description of the model, including summary statistics, is contained in the 2004 article in the journal Business Economics

The model is primarily intended to explain the decisions of voters rather than to predict the outcome of elections. As with any other econometric model, many variables are not included due to lack of data, and because models necessarily present a simplified explanation of complicated decisions. Furthermore, any prediction using this, or any other econometric model, relies upon factors that cannot be known in advance. The authors noted in their 2004 article, and repeat again today, that voters should consider economic and other factors in their decisions.

This model has been presented at the Hauenstein Center for Presidential Studies, first in 2004 and again on February 7, 2012. The 2012 presentation can be viewed here

Anderson and Geckil also describe the model in the article Pocketbook Predictions of Presidential Elections (2004), which was published in the journal of Business Economics. The article earned the Edward A. Mennis award for the authors. 

Please click below for detailed exhibits. 

AEG’s Patrick Anderson and Ilhan Geckil intend to update their election analysis for the 2016 race, as economic data becomes available. These updates will be available at


View PDF