Local Business Activity Index: Its Construction and Uses ‐reply

Abstract

In my article [18] a derived monthly SMSA income series was employed a s a dependent variable to estimate an index of Gross Metropolitan Product. The multiple step-wise regression analysis involving three independent variablestotal nonagricultural employment multiplied by the unemployment rate, bank debits, and the industrial electric power usage in million kilowatt hoursand a "ratio-tomoving-average" method were used in the process of constructing the index. The data were adjusted for the observed lags and autocorrelations, Professor Pyun's comments pertain to: (1) the theoretical basis of the choice and spatial definition of multiple variables, and (2) the validity of results yielded by the step-wise regression analysis.To refute the theoretical logic of using multiple variable ,series, ProfessorPyun cites three separate studies by Adelman [l]. Koopmans [9], and Pyunand Heck 1151. The work of Adelman deals with the sampling errors associated with the construction of the Consumer Price Index or the Wholesale Price Index; it does not concern the problems of local multiple series. It is crucial to recognize the differences between the national and local data and the related ramifications. Koopman's article propounds the "fuller utilization of the concepts and hypotheses of economic theory . . . as part of the processes of observation and measurement,"