It has often been argued that voter turnout in the United States is too low, particularly compared with turnout in other industrialized democracies, and that a healthy democracy should have higher turnout. One proposal that has been considered by Congress to increase voter turnout is the creation of Democracy Day, making Election Day a national holiday. In this study I evaluate the likely effectiveness of an election holiday in increasing turnout by studying how state regulations making election day a holiday for state employees affects voter turnout among state employees in those states. I exploit these natural experiments in a difference-in-difference context, using various groups of non-state employees as controls. My analysis relies on data from Voting Supplements to the Current Population Survey in November 2004 and 2006. The results are clear. There is no evidence from the natural experiment of states providing an election holiday for state employees that such holidays significantly increase voter turnout. I conclude that having an election holiday, by itself, is not an effective strategy to increase voter turnout.