Two business cycle models with endogenous firm and product entry are estimated by matching impulse responses to a monetary policy shock. The 'competition effect' implies that entry lowers desired markups and dampens inflation. Under translog preferences, where the substitutability between goods depends on their number, we find evidence of such an effect. That model generates more countercyclical markups than monopolistic competition model, where price stickiness is the only source of markup fluctuations. In contrast, a model with strategic interactions between oligopolistic firms cannot generate an empirically relevant competition effect and is statistically equivalent to the Dixit-Stiglitz model.