This paper uses institutional level data collected by the American Association of University Professors as part of their annual survey of faculty members' compensation to analyze faculty turnover. Analyses of aggregate data over almost a twenty-year period highlight how remarkably stable faculty retention rates have been nationwide and how little they vary across broad categories of institutions. Analyses of variations in faculty retention rates across individual institutions stress the role that faculty compensation levels play. Higher levels of compensation appear to increase retention rates for assistant and associate professors (but not for full professors) and the magnitude of this effect grows larger as one moves from institutions with graduate programs, to four-year undergraduate institutions, to two-year institutions.