Is total factor productivity growth measured correctly?

Total factor productivity growth has been on a downward trend in OECD countries, despite increase s in research spending, corporate modernisation and the length of studies. We can look for economic explanations for this paradox (declining efficiency of research, corporate concentration). But we also need to look into the risk of a problem in the measurement of total factor productivity, for two reasons: Q uality effects may be overestimated in the measurement of investment deflators, leading to an artificial increase in the real capital stock; The l abour and capital cost shares of value added have not been constant over time; the cost of capital share has increased, which reduces the usual measurement of total factor productivity. Our estimates show that the measurement of total factor productivity is highly sensitive to change s in the relative price of investment used and to the labour and capital cost shares . Its measurement is therefore fragile.