A meta-analysis is an analysis of existing empirical studies rather than another original primary study. The aim of a meta-analysis is to find out what is the average size of some parameter or effect in the existing literature - for example the average estimated damage from climate change or the elasticity of demand for gasoline - and what are the factors that cause there to be differences between the various studies.
The interfuel elasticity of substitution indicates how hard it is to replace one fuel with another when, for example, the price of one fuel goes up.* My meta-analysis of 46 empirical studies of this issue finds that at the level of individual industries it is probably not so hard to substitute between fuels (the elasticity of substitution is greater than one). At least some of the models, for example the G-Cubed model, used to assess the costs of climate change policy assume that it is a lot harder than this.
As far as I understand, this should mean that they overestimate the costs of climate change policy. The harder it is to replace fossil fuels with electricity or coal with natural gas the costlier it should be to adjust to a carbon tax or cap and trade scheme. But how sensitive are their results to the values of parameters such as the interfuel elasticity of substitution? Is there any research on this out there? (Yes, please point me to it!) Or is this a topic I should include in my research agenda? In fact I've been surprised in the past (back when Mick Common and I were debating ABARE in the run up to Kyoto) that the estimated costs of climate change policy aren't higher given the low substitutability assumptions built into these models.
* It's more complicated than that of course :)