The short-run relationship between employment and output has generated a wealth of literature since the early 1960's (Hultgren (1960, 1965), Kuh (1960, 1965a, 1965b), Neild (1963), Raines (1963), Wilson and Eckstein (1964), Brechling (1965), Ball and St Cyr (1966), Soligo (1966), Brechling and O'Brien (1967), Dhrymes (1967), Ireland and Smyth (1967, 1970), Masters (1967), Smyth and Ireland (1967), Fair (1969), Nadiri and Rosen (1969, 1973). A substantial part of this work has involved the construction and estimation of employment equations the composition of which is derived primarily from an (inverted) short-run production function. The choice of statistical surrogate for employment has strong implications for the structure of the estimating equations and yet, remarkably, relatively little attention has been paid to this aspect of the problem. In most of the studies employment has been approximated either by the number of men in employment or by man-hours. Both formulations assume implicitly that men and average hours are perfect substitutes for one another; that is they are functionally related to the same set of exogenous variables, adjust at an equal rate to their own lagged values and exhibit similar reactions over the cycle to each and every influence. Such assumptions, however, are bold and several writers have seen the need to provide a separate specification for equations depicting the demand for the number of workers and their rate of utilization (see Kuh (1956b), Brechling (1965), Fair (1969), Nadiri and Rosen (1969, 1973) from above and also Brechling (1974) and Ehrenberg (1971)). Three major reasons may be given which would warrant such a separation: (i) Hours may be viewed as comprising the principal short-run means of adjusting labour to output changes while men are adjusted to meet longer-term movements in output, capital stock, etc. (ii) Men and hours may themselves be interdependent (recursively) given the different time scales mentioned above. (iii) Certain exogenous influences may affect the demand for men in different (and possibly opposite) ways from their effect on hours. The main purpose of this paper is to present and test a simple theory which takes explicit account of the interdependence between the employee and utilization decisions. The functional form of our model is developed in Section 2 with particular variable constructions and results presented in Section 3. Where possible, it will be interesting to compare our results with those of Brechling (1965) since the early development of our model follows closely along lines suggested by Brechling and, like him, our data refer to British manufacturing industries. Brief conclusions appear in Section 4.
Read full abstract