This paper demonstrates both theoretically and empirically that there are industry-specific factors as well as the cyclical ones that affect quit and layoff incidences and that the industry-specific effects are positively correlated between the two incidences across industries while the cyclical effects are negatively correlated over time. We first set up a theoretical model to analyze how its parameters affect quits and layoffs through the corresponding change in the optimal wage for the employer and the worker, and then derive from the theoretical separation behaviors the two testable hypotheses - that quits and layoffs are positively correlated to each other across industries, and that quits move procyclically while layoffs move countercyclically. We analyze the two sets of data, BLS establishment data on turnover rates and PSID, to empirically test and confirm each of the two hypotheses.