Includes bibliographical references (leaves 54-55). / Using survey data from the third World Values Survey and democracy scores from Freedom House, I outline and test a supply and demand model of democratic change and stability. While "support for democracy" is a common concept in political science, the only empirical studies of the relationship between these regime preferences and democracy (Inglehart, 2003; Inglehart & Welzel, 2003; Welzel, Inglehart and Klingemann 2003) do not control for reciprocal causation and use poorly conceptualised and measured variables. They claim that deeply-rooted cultural orientations called "self-expression values" are a better measure of implicit support for democracy than overt expressions of regime preference. However, I find that once I control for the possibility that democracy is exogenous, there is little difference between the explanatory power of cultural values versus overtly expressed preferences. Furthermore, I argue that popular regime preferences (or demand) affect the change in the level of democracy, but do so only in relation to its current supply. "Net demand" is the driver of system change rather than absolute levels of popular support for democracy.