Estimating Duality Models with Biased Technical Change: A Time Series Approach

Abstract Technical change is an omitted variable in econometric models which estimate technical change biases with no direct measure of this variable. Modeling technical change as a deterministic time trend is a restrictive representation that may be inconsistent with the type of nonstationarity of the other model variables. We used a time‐series approach to estimate a cost function for central Canada and found that factor shares, prices, and output are cointegrated, implying that technical change is neutral. In contrast, estimating the system with a time trend as a technical change measure leads one to conclude inappropriately that technical change biases exist.