Abstract : We show in this paper that a strong correlation exists between the performance of chunk parsers and the quality of the tagging task in input. This dependency becomes linear when the size of the chunks is small. Our demonstration is based on an experiment conducted at the end of the Passage 2007 shared task evaluation initiative (De la Clergerie et al. 2008). The performance of two parsers which took part in this evaluation has been investigated. The results indicate that the chunking task, for sufficiently short chunks, is similar to a super-tagging task.