The median value of realized changes for surveyed firms expecting "no change" is zero. I suggest to modify the Carlson-Parkin procedure using this conditional absolute null (CAN) regularity. The adjustment solves the zero response problem and improves the variance estimates.

We investigate how firms' market power affects the price level. In our small macro- model we show, that firms - in addition to hypothesised structural mark-up pricing power - may take advantage of favourable business cycle fluctuations. The paper provides empirical evidence for both these...