We conduct Monte Carlo experiments to investigate the biases of assuming a misspecified demand model. We study continuous models (linear, log-linear and AIDS), and discrete choice models (logit) in the context of differentiated products and aggregate data. Estimating demand with the 'wrong'...