Observations on productivity and composition of building construction output in the United States, 1972-82

[Extract from introduction] This article extends by a considerable degree the analysis of changes in construction productivity as a function of the composition of outputs. The authors had access to F. W. Dodge contract construction data, 1972-82, and it was possible to develop cost functions to estimate changes in total factor productivity in building construction over the 1972-82 period based upon shifts in the physical composition of outputs and changes in the prices of those outputs. The focus of this study is admittedly narrower in scope than other studies which have addressed changes and measures of changes in real construction output. Here the analysis is confined to changes in the real cost of physical units of output (square feet of floor space) of building construction. It is assumed that any temporal changes in the average real unit cost of construction output are attributable to two general causation classes: (a) changes in the mix of types of construction outputs whlch have different unit costs; and (b) a residual category consisting of changes in total factor productivity, qualitative design and other unmeasured variables, including any statistical bias associated with the deflator indexes or the Lespeyres index system. The analytical framework employed in this study allows separate examinations of unit cost changes associated with these two classes of causes. The objective of the study is to provide estimates of changes in the real unit cost of construction output during the 1972-82 period and the portion of this change that is attributable to the mix or composition of types of outputs.