Using nominal fishing effort to control fishing mortality and using cpue data from commercial fisheries as abundance indices require ability to correct fishing power for temporal development. It is often assumed in ICES stock assessments that fishing power Is constant over time. However, experience has suggested that this assumption may be false. This study investigates the time dynamics of an Index of Fishing Power (IFP). This index is based on the fleets cpue. relative to the cpue of a subset of vessels from the same fleet. The primary characteristic of the reference vessels is that their fishing power has not varied much over time. IFP is calculated for some of the Danish cod fisheries in the Baltic Sea. IFP appeared to be independent of the vessel composition of the reference sub- fleet and for the fleets fishing in the Eastern Baltic Sea. but less so for the Western Baltic Sea fleets. Variations in IFP are analysed by a GLM (General Linear Model). Results suggest that fishing power has developed in the Eastern Baltic Sea cod fishery at an annual rate of 2% and 6% for trawlers and gillnetters respectively. Mechanisms of fishing power creeping may include increased technical efficiency and also changes in vessel composition of the fleets over the period of analysis. (C) 2001 International Council for the Exploration of the Sea.