As described earlier, the primitive data for our UK study required certain transformations in order to create exchange matrices E_IJ in the form recognized by the SFEcon system. Once the series of E_IJ matrices was available, the SFEcon algorithm was used to create a corresponding series of U_IJ matrices on the presumption that the exchanges represented a general economic optimum. These operations were carried-out in an MSWin Excel workbook containing the program used, together with instances of follow-up calculations required for some sectors.

The utility matrices U_IJ were then brought together in another workbook where they were sorted so as to create a time series for each element IJ of U_IJ. Series were created for U_IJ parameters 1) stated in terms of a constant monetary unit, and 2) in terms of parameters normalized as U_IJ/Z_J. Graphs were developed for the time series of each element IJ of U_IJ in two workbooks, here and here, – one workbook for each of the two suppositions about how the parameters might be expressed in the same unit from year to year. These suppositions being 1) utility parameters are constant, and 2) utility parameters change monotonically.

A comparable set of graphs was generated for ‘parameters’ created by simply inverting the exchange matrices E_IJ. The apparent lack of temporal stability in these inverse coefficients might be taken to indicate that the parameters referenced by more conventional Leontief analyses are not parameters in the sense of representing something that is stable in the dimension of time.

One last workbook was used to compute the R-bar-squared for the data’s fit to each of two naïve time series models: a simple average and an ordinary least squares regression against time. It is presumed that high correlations in these measures adequately indicate that hyperbolic production parameters reflect the gradual changes we expect in the economy’s underlying technical potential.