Will Wilkinson worries about pricing carbon emissions:
In order to estimate the optimal pigouvian tax, we not only need a solid estimate of the net harm of warming, but we also need a good estimate of how much of that is the external effect of human activity. I don’t think there exists a good estimate, which I think gives us good reason to worry about proposed carbon taxes. Any tax, unless we are very lucky, will either be too low or too high. If it is too low, we’ll get too much carbon emission. But if it’s too high, we’ll get too little and I think that’s likely the more worrying scenario, especially if it slows growth for poor countries. And I worry that harm could turn out be larger than the harm the tax is meant to prevent.
I'm not entirely convinced the downside to slower growth is larger that the downside to worse-than-expected climate change, but I'm (a) not an expert in the field, and (b) pretty risk-averse in general. But even if one grants Wilkinson's point, this is an argument for erring on the side of a too-low price on carbon emissions (which is likely to come about anyway, snice the incentives for politicians are stacked against long-run environmental concerns), not an argument for having no carbon tax whatsoever. There's essentially no disagreement at all that there's some externality associated with carbon emissions, so the optimal carbon tax is certainly not zero. It seems to me the ideal solution here would be to take our best (admittedly rough) estimate for the optimal size of the carbon tax, then decide how sure we want to be that we're not erring on the high side (70 percent? 90 percent? 99 percent?), and set the carbon tax accordingly. It doesn't seem like it should be beyond the realm of econometric possibility.