In this paper we present preliminary experiments that aim to reduce lexical data sparsity in statistical parsing by exploiting information about named entities. Words in the WSJ corpus are mapped to named entity clusters and a latent variable constituency parser is trained and tested on the transformed corpus. We explore two different methods for mapping words to entities, and look at the effect of mapping various subsets of named entity types. Thus far, results show no improvement in parsing accuracy over the best baseline score; we identify possible problems and outline suggestions for future directions.