This study evaluates the usefulness of different modifications of empirical models estimating the so-called Boone indicator for capturing changes in the intensity of competition. We use as “natural experiments” in this evaluation data from three cartel cases: i) international elevators and escalators cartel in various European countries during the years 1995-2004, ii) Finnish raw wood cartel during the years 1997-2004, and iii) Finnish construction cartel 1994-2002. The findings support our argument that particularly when the primary interest is to evaluate the effectiveness of a certain competition policy action, the empirical model should properly take into account a possible structural break in data due to the policy action. Furthermore, our data hint that the methodological choice of prior empirical studies to use data only from one industry at a time may lead into the false conclusions when the Boone indicator is used for evaluating the effectiveness of sector-specific competition policy actions.
