Recent evidence show that factor shares, if properly measured, are far from constant. Moreover, the shares of natural resources and raw labor seem to be negatively correlated with income per capita while the share of human and physical capital is positively correlated with income per capita. Now, if factor shares are not constant then (i) growth accounting exercises rely on a false assumption and (ii) there is a measurement problem. The effect that change s in factor shares ha ve on output depend on the relative abundance of factors and, fo r this reason, it is necessary to have correct measures. We propose an empiri cal methodology to solve the measurement issue and estimate TFP growth.

---