Generated regressors in linear and nonlinear models

The need to use estimates from separate studies as regressors often arises. Such regressors need to be modelled as variables measured with error. We demonstrate the use of generated regressors in linear and nonlinear models that arise from the predictions in Grossman and Helpman (American Economic Review, 1994, 84 (4), 833–850) and Grossman and Helpman (Journal of Political Economy, 1995, 103, 675–708).