In his 2015 letter to shareholders, Warren Buffett addressed climate change risk. Asked by shareholders how his company was planning to address climate change, Buffett responded that climate change itself posed little threat to his insurance business. Premiums are adjusted annually, therefore risks that increase over decades can be managed by increasing prices. Knowing the price of risks allows the company to chart a reasonable and profitable course despite potentially increasing exposure. This is hardly an alarmist position.

But Buffett says that it seems highly likely that climate change poses a major problem for the planet. He says highly likely instead of certain, because he has no scientific background (though many scientists would probably share that opinion). Buffett recalls the advice of experts during the Y2K scare, when fears of calamity turned out to be overblown. Buffett, however, does not say that proof of damage is required to motivate action.

It would be foolish, however, for me or anyone to demand 100% proof of huge forthcoming damage to the world if that outcome seemed at all possible and if prompt action had even a small chance of thwarting the danger… if there is only a 1% chance the planet is heading toward a truly major disaster and delay means passing a point of no return, inaction now is foolhardy.

This perspective can inform people of any political stripe. Conservatives are willing to marshal action based on risk in non-climate contexts. Dick Cheney cited the same 1 percent to implore action for national security, saying a 1 percent chance of terrorists getting a weapon of mass destruction requires action as if it were a certainty. So while conservatives may be more risk-averse toward national security threats than climate change, such a calculation should not be unimaginable.

So how do we decide how to act when there is a diversity of risk preferences amongst political groups? The financial sector provides an example where political ideology gives way to the preferences of market actors, who are willing to pay a substantial premium on non-diversifiable risks (in this context, risks that are a major problem for the planet and cannot be insured). And in this context, that premium should be included in a market price on carbon emissions. A price on carbon will reduce emissions and offer a chance of reducing the risk of catastrophe.

Many climate skeptics argue that the most likely scenario for global warming is that human emissions of CO2 and other greenhouse gases will cause mild warming, a geographic mixture of winners and losers, and what problems arise can be met by adaptation. Innovation will drive down the cost of carbon-free energy on its own. Therefore, a carbon price to reduce emissions of CO2 will likely do more harm than good.

But this is the wrong way to look at the problem. Even if climate skeptics are right about the most likely scenario being a moderate warming, the possibility remains that the unlikely will occur. Thinking about the problem in terms of temperature increase for a doubling of atmospheric CO2 (which we will probably exceed with current policies and energy trends), even studies that reinforce the skeptical narrative of low mean climate sensitivity leaves some chance of warming greatly exceeding international goals and historical boundaries (say a 5 percent chance of warming exceeding 4 °C). How such a warming would impact the probability of irreversible changes to elements of the climate system (melting ice sheets, reversal or slowing of ocean currents, release of carbon in permafrost) is unknown. So are the economic implications of such a warming.

How the myriad uncertainties in the physical climate system, and the human and ecological response to warming, can be incorporated into carbon pricing is a matter of research. But the notion that the risks of climate change compel us to action— and therefore some carbon price— does not demand perfect knowledge of such risks.