Hopes are high that removing fossil fuel subsidies could help to mitigate climate change by discouraging inefficient energy consumption and levelling the playing field for renewable energy. In September 2016, the G20 countries re-affirmed their 2009 commitment (at the G20 Leaders' Summit) to phase out fossil fuel subsidies and many national governments are using today's low oil prices as an opportunity to do so. In practical terms, this means abandoning policies that decrease the price of fossil fuels and electricity generated from fossil fuels to below normal market prices. However, whether the removal of subsidies, even if implemented worldwide, would have a large impact on climate change mitigation has not been systematically explored. Here we show that removing fossil fuel subsidies would have an unexpectedly small impact on global energy demand and carbon dioxide emissions and would not increase renewable energy use by 2030. Subsidy removal would reduce the carbon price necessary to stabilize greenhouse gas concentration at 550 parts per million by only 2-12 per cent under low oil prices. Removing subsidies in most regions would deliver smaller emission reductions than the Paris Agreement (2015) climate pledges and in some regions global subsidy removal may actually lead to an increase in emissions, owing to either coal replacing subsidized oil and natural gas or natural-gas use shifting from subsidizing, energy-exporting regions to non-subsidizing, importing regions. Our results show that subsidy removal would result in the largest CO2 emission reductions in high-income oil- and gas-exporting regions, where the reductions would exceed the climate pledges of these regions and where subsidy removal would affect fewer people living below the poverty line than in lower-income regions.