The economic analysis of global warming is dominated by models based on optimal growth theory. This approach can generate biases in the presence of positional goods and status effects. We show that by ignoring these direct consumption externalities, integrated assessment models overestimate the social return to conventional investment and underestimate the optimal amount of investment in mitigation. Empirical evidence on the influence of relative consumption on utility suggests that the bias could be quantitatively significant. Our results from a simple survey support this conclusion.