Production Function Estimation: Biased Coefficients and Endogenous Regressors, or a Case of Collective Amnesia?

The possible endogeneity of labor and capital in production functions, the consequent bias estimated elasticities, has been discussed addressed literature different ways since 1940s. This paper revisits an argument first outlined 1950s, which questioned function estimations. is that output, capital, employment are linked through a distribution accounting identity, key point recent overlooked. identity can be rewritten as form resembles (Cobb-Douglas, CES, translog). We show this happens because data used empirical exercises value (monetary) data, not physical quantities. clear predictions about size factor elasticities what commonly interpreted elasticities. To test these predictions, we estimate typical Cobb-Douglas using five estimators that: (i) responsible for fact must shares; (ii) (i.e., departure from shares) is, reality, caused by omission term identity. However, unlike standard omitted-variable problem, here omitted known; (iii) estimation method second-order issue. Estimation methods theoretically deal with endogeneity, including most ones, cannot solve problem. conclude use monetary values rather than poses insoluble problem functions. consequently, far more serious any supposed problems.