<p>The consensus view among economists is that carbon prices, in order to be efficient, must be the same across the globe. But when there are inefficiencies in the allocation of capital so that countries are subject to different discount rates, we show that efficient carbon prices must be different across countries. This is a consequence of Hotelling’s familiar argument on the price of a non-renewable resource: it must grow at the rate of the next best use of marginal funds, which is equal to the country’s discount rate. If different countries discount at different rates, their carbon prices ought to grow at different rates as well. If they grow at different rates, they can’t be the same all of the time, as first-best carbon prices are. The computational climate policy literature has so far avoided this conclusion by altering time preferences in a country-specific way through time-varying Negishi weights. We show that the use of such weights causes inefficient policy prescriptions and, furthermore, has the particularly undesirable consequence of incorrectly discounting future consumption more in countries with high growth rates. The existence of inefficiencies in the savings process—causing differences in discount rates—is well-known and should be acknowledged head on in climate policy analysis. Doing so results in global mitigation policy with carbon price paths for different countries growing (efficiently) at different rates.</p>