We have carried out a meta-analysis of the costs of mitigating global GHG emissions over the period to 2100, with and without the effects of induced technological change. The literature reporting costs uses a variety of assumptions and modelling approaches and a limited range of economic instruments, usually carbon taxes or auctioned CO2 emissions trading allowances at a regional or global level. It reports a wide range of costs with confusing and overlapping choices of assumptions. The purpose of the study is to use regression and related analyses to assess what effect the assumptions about treatment of technological change have on the published estimates of the costs, measured as changes in welfare or gross world product, and of the required CO2 tax rates and emission permit prices. We report the results in terms of two sets of equations (one for gross world product, the other for the tax rates/permit prices) explaining most of the variance in the published results, covering the Innovation modelling Comparison Project’s 2006 study and the earlier meta-analyses done by the World Resources Institute for the US economy, 1997, and the IPCC post-SRES models for the global economy, 2002. In the full study covering some 1,500 observations, the major influences on the results for world product and growth (besides the extent of the reduction in CO2 required) are found to be assumptions made for (1) the treatment of technological change and (2) the use of revenues from taxes and permit auctions. When the models allow for induced technological change or when revenues are recycled, e.g. via investment incentives, growth is higher. Allowance for the Kyoto Mechanisms, climate and non-climate benefits, and a backstop technology all further reduce costs. The level of tax rates and permit prices is found to depend on the stringency of the CO2 stabilization target (raising prices), and the modelling of induced technological change and disaggregation of sectors (reducing prices). The overall conclusion from the modelling literature is that even stringent stabilisation targets can be met without materially affecting world GDP growth, at low carbon tax rates or permit prices, at least by 2030 (in $US(2000), less than $15/tCO2 for 550ppmv and $50/tCO2 for 450ppmv for CO2). However induced technological change is a relatively new topic in economic modelling and results are often experimental and controversial.