We use the cost-of-carry model to investigate the extent of market efficiency in the EU futures market for carbon dioxide allowances over the period of June 2005 to December 2007. We reject the cost-of-carry hypothesis for the entire data sample, but find some evidence of improvement in market efficiency over the period. Recursive estimates of some cost-of-carry model parameters start approaching their theoretical values when estimated on progressively smaller and more recent sub-samples.