Did the Green Party decide the United States presidential election of 2016?

This paper models what voters do in the absence of their preferred party. In this context, some hypotheses are formulated with respect to the role of Green Party voters during the 2016 US presidential elections. To test these theoretical predictions, voting data at county level is analyzed with a Fixed Effects OLS model. Utilizing state differences in party's ballot access, state borders are used as a source of exogenous variation. Provided that voter preferences are single peaked and that the Democratic Party and Green Party are substitutes, we find that the presidential elections could have seen a different outcome had the Green Party not been on the ballot in just three states.