Although a realistic characterization of the production function is critical to macroeconomic analysis, estimating the function's characteristics is hampered by both data limitations and methodological difficulties. In this paper, I develop a new empirical strategy for estimating the CES production function with biased technical change. The proposed method extends the control function approach to the CES specification to address endogeneity concerns and is able to retrieve sector-specific and time-varying estimates of technical change. Using data from U.S. manufacturing industries, I find evidence that (i) the production technology exhibits nonincreasing returns to scale, (ii) the elasticity of substitution between capital and labor is below unity, and (iii) technical change is generally labor-augmenting along the balanced growth path.
