The Gollop and Jorgenson [118] study focused on U.S. productivity growth by two-digit industry between 1947 and 1973. Accordingly, their production function explicitly included intermediate inputs in order to capture, among other things, changes associated with input substitutions. Gollop and Roberts [119) extended the Gollop-Jorgenson model to account explicitly for domestically produced and imported intermediate inputs within the production function. They examined 1948 to 1973 data from the U.S. manufacturing sector. One important conclusion is that an indirect (as well as direct) contribution by inputs to the rate of technological change exists: "The indirect effect on technical change is a function of substitution possibilities and the factor-using/factor-saving nature of technical change" [119, p. 173). In three-fourths of the manufacturing industries studied, the indirect effect of more costly imported intermediate inputs (in particular, oil after 1973) was found to retard the rate of technological change (and thus had an impact on the slowdown in productivity growth during the mid-1970s).