Good energy policy can probably stabilize emissions at a 2050 GDP cost of less than 2% Studies project more like 1 to 1.5% but forecasting how it is all going to work out in 50 years is impossible. So 2% is just my speculation, based on my optimism about markets and innovation, and a recognition that humans can adapt and adjust policy as innovations and effects emerge.
What do I mean by good policy? Simple summary: a revenue-neutral carbon tax. Establish a predictable rising price on carbon from a low level of around $15/ton CO2, increasing to around $60/ton by 2050. If innovation is running ahead of plan, then that offers the possibility of reducing planned future increases. Inversely, if emissions are running ahead, increase the future carbon price. Something like Al Gore’s proposal could result in a GDP gain by 2050, not a loss [e.g., because energy not consumed is cost not spent]. Gore proposes rebates of the first dollars of payroll taxes — dollar for dollar with revenues from the carbon taxes.
For the details of how carbon tax revenues can be fully rebated and distributionally neutral, see e.g., “A Proposal for a U.S. Carbon Tax Swap – An Equitable Tax Reform to Address Global Climate Change” [Brookings Institution Hamilton Project Oct 2007] by Gilbert E. Metcalf, Professor of Economics, Tufts University.