This dissertation analyzes expense preference behavior by managers of Nevada casinos. Annual aggregate data is obtained from the Nevada Gaming Control Board from 1990 to 2010. Using ordinary least squares regression, three models are run to account for various payroll related expenses. The dependent variable in the first model is number of employees. The second model uses salaries and wages as the dependent variable and the third model uses total payroll. All models include predictor variables of total revenue as an indication of size, a dummy variable for market concentration and a dummy variable for recessionary periods. The finding show that larger casinos spend more in payroll related expenses. Results indicate that for each 1% increase in revenue, number of employees increases 0.84%, salaries and wages increase 0.96% and total payroll increases 0.99%. The significant positive relationship between revenue and payroll related expense may be an indication of expense preference behavior by managers since larger firms typically have more layers of management. The additional layers also cause a separation between owners and managers and the more layers, the more separation. The market share variable is nonsignificant in all three models. The recession variable was significant and negative in all models. During recessionary periods, Nevada casinos decrease number of employees 35.5%, salaries and wages 7.7% and total payroll 5.2%. The negative significant relationship between the recession variable and payroll related expenses may also be an indication of expense preference behavior by Nevada casino managers during non-recessionary periods. Since managers are able to decrease payroll related expense after controlling for the change in business volumes, they are most likely operating inefficiently during non-recessionary periods.