Estimating consumer behaviour in an energy-economy policy model

CIMS is a technologically explicit hybrid energy-economy model that forecasts effects of policy alternatives on technological change and greenhouse gas emissions. It strives towards realistically representing consumer behaviour to better forecast effects of policies. This study attempts to improve the behavioural realism of the model by calibrating the parameters representing consumer behaviour using historical data from 1990 to 2004. A statistical simulation called Markov Chain Monte Carlo generates a probability distribution over the behavioural parameters for three technology competitions. The calibrated model is then applied to a policy analysis forecasting the effects of a carbon tax on residential furnace emissions to 2050. Despite insufficient variation in energy prices over the historical period, uncertainty in model structure, and an absence of revealed preference data for emerging technologies, historical calibration can improve model credibility and thus usefulness for policy-makers, particularly when used in combination with other, stated preference parameter estimation research.