to gauge changes in employment opportunities, as measured by quantity of newspaper advertisements. Perhaps this measure does accurately reflect changes in employment opportunities from one period to the next for a given region. However, comparison of index values across regions is meaningless, unless the number of vacancies was identical in all regions in 1969. The job vacancy survey data compiled by Denton, Feaver, and Robb (1975) indicate that this is unlikely.2 Consequently, variation across regions in labour market conditions is probably inaccurately measured by ACSW. Since the measurement error occurs in one of the explanatory variables for their regression model, in principle it should result in estimation bias, thus calling into question the validity of their reported results (Johnston, 1972, 281-2). In a preliminary version of this paper, which was presented at the 1978 meetings of the Canadian Economics Association, the authors had results based not only on the regionalized help-wanted index, but also on regionalized job vacancy rates derived from the job vacancy survey data compiled by Denton et al. Given the conceptual problem associated with the regionalized help-wanted index in the context of this study, it would seem preferable to base conclusions on the estimates obtained using the job vacancy series.
Read full abstract