Allocation of individuals to job levels under rationing

We augment an ordered response model for the job level at which individuals desire to work with allowance for individual reporting of a discrepancy between actual and desired job level, and we also investigate possible biases from restricting the observations to the employed, and omitting the unemployed. We find that the latter bias can be ignored and interpret this as evidence that the unemployed are not choosier than the employed. People claiming inadequate capability utilization are not markedly differently allocated than those claiming a proper match. Women's lower job levels cannot be explained from rationing by capability utilization.