A multilevel model of early retirement decisions among autoworkers in plants with different futures

During the period of their 1986-1989 General Motors (GM)-United Auto Worker s (UAW) contract, about 17% of all GM autoworkers who were eligible to elec t early retirement did so. Those who did were distinctive in theoretically expected ways, with expectations defined by individual characteristics such as age, physical health, and pension wealth. But some of the workers were employed in plants that GM had decided to abandon. Did that difference in o rganizational context make a difference in individual workers' decisions ab out early retirement? Would workers who chose to take early retirement and who were employed in plants scheduled to close have made the same decision had their plants not been selected for closure? If the rate of early retire ment was higher in plants scheduled to dose, and it was, how did that diffe rence relate to the process by which individual workers reached their decis ions? These are some of the questions asked and answered through multilevel analyses of data from a probability sample of GM's autoworkers. These anal yses generate findings not detected in single-level analyses of the same da ta.