In this paper we present two extensions that have been largely omitted in the recent literature on Bayesian estimation of DSGE models. First, we pay special attention to di?erent forms of complementarity between consump- tion and hours a?ecting the households preferences. Second, we allow for the presence of a fraction of non-Ricardian households —i.e. that have limited access to financial markets—. We show that exogenous changes in government transfers are crucial to distinguish between the two sources of comovements of consumption and hours in response to government spending shocks. The main conclusion from the estimated models is that private consumption increases af- ter a government spending shock, when either nonseparability, non-Ricardian behavior, or both, are introduced in the model. In addition, allowing for consumption-hours complementarity leads to a small, and stable over time, estimated fraction of non-Ricardian households.