Our willingness to embrace climate change policies depends on our perception of their benefits and costs. Evaluation of these costs and benefits requires cartful economic analysis. Yet the standard tools for such assessment - computable general equilibrium (CGE) models - are inadequate on several grounds. Their underlying theory suffers from well-known logical difficulties; in general, their equilibria may be neither unique, stable, nor efficient. Moreover, real-world phenomena such as increasing returns to scale, learning, and technological innovation are neglected in CGE models. These phenomena make the resulting equilibria in the models inefficient; in the real world they can lock society into sub-optimal technology choices. They introduce uncertainty and path-dependence, annihilating the concept of a single efficient allocation produced by the unfettered market. Yet conventional economics assesses the cost of policies solely on the basis of their departure from a purportedly efficient equilibrium -ignoring deeper structural changes that are often decisive in practice. New socio-economic theories and models are emerging that allow for bounded rationality, the limiting and enabling character of institutions, technological change, and the complexities and uncertainties in economic evolution. Meanwhile, existing models should be modified to better reflect real-world phenomena and to abandon unfounded assumptions about the inherent "inefficiencies" of government intervention in the market.