This paper deals with modeling total factor productivity (TFP) growth in a
flexible manner using panel data. Several competing parametric models are u
sed to explore whether there are ant similarities in the estimates of TFP g
rowth and technical change among these models. Using a primal approach, we
decompose TFP growth into different components. The models are then used to
measure productivity and technical change in the Swedish cement industry.
In general, the results are found to be model dependent and often conflicti
ng, although much less so for returns to scale and overall productivity gro
wth. JEL classification: O30; C33.