Testing the Proposed Model

Changes in NTV can be calculated from two independent data sources, first from the differences between the growth rates of employment and the volume of capital and secondly from the changes in the constituents of NTV. Their independence allows the model to be tested. The significant correlations between the two are evidence for the robustness of the model and thus for the validity of the conclusions derived from it. Due to data limitations, changes in NTV for companies are compared with those for the economy as a whole. The correlations are strong for profit margins and leverage throughout the whole period from 1946 to 2016 and for the period 1966 to 2016 are strong for all constituents of NTV except interest rates. The proposed model should therefore be preferred to the current consensus model. It is testable and robust when tested while the standard approach appears to be untestable.