A global three-dimensional chemical transport model is used to investigate several possible causes for the observed decrease since the end of 1990 in the tropospheric CO abundance. Several sensitivity calculations are performed to assess the impact of potential recent changes in the industrial CO source, in the emissions from biomass burning, in the stratospheric ozone column and in tropospheric temperature. Only a combination of these factors could potentially account for the observed changes, provided that sufficiently large changes occurred. Most of the specified changes applied to the model should be regarded as hypothetical and need to be confirmed by experimental evidence.