Total factor productivity and the measurement of technological change

TFP is interpreted in the literature in different, mutually contradictory ways. Changes in TFP are shown to measure not technological change, only the super-normal returns to investing in such change - returns that exceed the full opportunity cost of the activity. Thus, in the limit, technological change can proceed with unchanged TFP. Measuring the effects of technological change instead requires counterfactual estimates. Reasons why changes in TFP are imperfect measures of super normal returns are also studied - reasons connected with the timing of output responses, the treatment of R&D in the national accounts, the omission of resource inputs, and two types of aggregation.