This is a revised version of our earlier paper which was prepared with the support from the Center for Slavic Studies of the University of California at Berkeley. We have benefited from the discussions with professors G. Grossman, T. Marschak, B. Ward, T. Rothenberg, P. Zarembka, J. Kmenta and A. Zellner.
The typical list of industrial branches is: (1) Electric Power, (2) Fuels, (3) Metallurgy, (4) Machine Building, (5) Chemicals, (6) Building Materials, (7) Lumber and Woodworking, (8) Paper, (9) Glass and Ceramics, (10) Textiles, (11) Clothing, (12) Leather and Shoes, (13) Printing, (14) Food Processing, (15) Other Industrial Branches.
Manhours in Czechoslovakia and Poland; Number of workers in Rumania and Hungary, and employees in Bulgaria.
Capital stock was adjusted in two steps: The average annual capital stock for each branch was first multiplied by the branch “shift coefficient” and then regressed on (1) the productive consumption of electric power in the given branch, (2) time, and (3) product of electrical power and time variables. The fitted values from this equation were then taken for capital adjusted for the degree of utilization. This method is apparently based on assumption that the utilization of the capital stock is proportional to be length of the ,working day’ as measured by the shift coefficient (one shift = 1) and at the same time the variation in the degree of utilization are connected with the short term variations in the ratio of the consumed electricity per unit of capital stock.
Here we use this term for simplicity, although we are fully aware of the fact that r is more a "measure of our ignorance" than anything else
See Balestra and Nerlove (1966), Mundlak (1966), Wallace and Hussain (1969) and Kmenta (1970, 1971).
The reported Durbin-Watson statistic is an arithmetic mean of the branch Durbin-Watson statistics calculated with the branch sums of squared residuals as weights.