units on employment data

Man-years of labor is one way to look at it. By definition, employment in the data base (and; therefore, the impacts) are part-time/full-time annual average jobs. This means it is a body count that is adjusted for seasonality. 4 workers working 3 months is one annual average job (whether or not they are part-time or full-time). If the majority are part-time the earnings per worker will be appropriately lower - it is the payroll that drives the induced effect.