In a recent post on the Institute for Energy Research website, Jordan McGillis, the IER’s deputy director of policy, worries about the implications that scientific and economic uncertainties pose for sound climate policy. He warns that if we reason that “climate change causes harm, therefore we should blunt that harm even if we aren’t sure on the magnitude,” we “risk causing more harm than we resolve.”

This is not a small matter. In that post and elsewhere, McGillis is especially critical of a “target-consistent” approach to climate policy that begins by setting a specific goal and then works backward to devise policies to reach it. World leaders have, in effect, already opted for target consistency. In the Paris Agreement, they set a goal to limit global warming to well below 2 degrees Celsius, compared to pre-industrial levels. To reach that goal, greenhouse gas emissions would need to fall to net-zero by about 2050.

McGillis and his organization are cautious on climate policy, but there are critics of target-consistent climate policy on the activist side, too. In an article for Science, Joseph Aldy and several colleagues express the view that a push to swap the economists’ approach, which drives emissions decisions by internalizing the social cost of carbon, for a target-consistent approach that works backwards “could set back climate policy, just as the United States is poised to take meaningful climate action.” As they see it, “the target-consistent approach replaces scientific assessments of damage … with subjective judgements about political targets and choices.” Because “the anchor for the target-consistent price is fundamentally political, not scientific, it is therefore subject to arbitrary change.”

In my view, these critices of target-consistent climate policy misunderstand the relationship between science and politics in climate policy, as well as misunderstand the alternative to target consistency. Here’s why.

The Social Cost of Carbon 

First, we need to understand what the critics of target-consistent policy hold up as the scientific alternative—an alternative that centers on the “social cost of carbon,” or SCC. The SCC is a measure of the harm done, not just immediately but through all of the future, by the emission today of one unit of carbon dioxide or its equivalent in other greenhouse gasses.

Economists have long argued for putting a price on carbon emissions equal to the SCC. The idea is that emitters would then curtail pollution when the cost of doing so was less than the price, but continue business as usual if the cost of cleaning up was greater than the price. This is often called the “Pigovian” approach after the British economist Arthur Cecil Pigou, who first proposed it more than a century ago.

The pricing approach could be implemented by taxing emissions or by creating an emissions trading system (aka “cap-and-trade”) that capped total emissions and allowed polluters to buy and sell rights to emit. Some economists argue that a carbon price alone could do the job of optimally controlling emissions. Others see a carbon price as part of a broader approach that also includes regulation and subsidies.

In that case, the SCC could still serve as a benchmark in testing whether a proposed regulation or subsidy would be cost effective. In any event, the outcome would be climate mitigation policy that was effective without being so aggressive as to (in McGillis’s words) “cause more harm than we resolve.”

All this sounds good, but it turns out to be devilishly difficult to pin down an exact value, or even a narrow range of values, for the SCC. It’s not that economists haven’t tried. One of them, William Nordhaus, even won a Nobel prizefor his efforts.

The most sophisticated estimates of the SCC rely on integrated assessment models. An IAM is a system of equations that spans economics, demographics, climatology and other disciplines. Basic assumptions regarding population growth, sensitivity of global temperatures to CO₂ concentrations, savings rates, and so on are fed into the IAM, and an estimate of the SCC comes out the other end.

The problem is that every single number that goes into the IAM is fraught with uncertainty. Many of the uncertainties stem from science or economics. How much will the planet warm in response to a 1 percent increase in atmospheric carbon? How much will it cost to install a gigawatt’s worth of solar panels in 2051? But although uncertain, objective analysis can at least determine plausible probability distributions for these factors. The really troubling uncertainties in SCC estimates concern “unknown unknowns” that are much harder to quantify.

The most important of those uncertainties concerns what economists call the discount rate. This is a measure of how much we value, say, $100 of reduced harm in the future compared to $100 of costs incurred today. For example, during the Obama administration, an influential interagency working group on the SCC used a discount rate of 3 percent. At 3 percent, it would be worth spending $97 today to avoid $100 in damages next year. But as the harm moves further into the future, its present value decreases. Discounted at 3 percent, the amount worth spending today to save $100 in harm at the end of this century would be just $10. 

SCC calculations have a long time horizon, reaching to the end of the century and beyond. As a result, the choice of discount rate fed into the IAM has a huge effect on the result. Reducing the discount rate from 3 percent to 2 percent more than doubles the present value of benefits enjoyed in the year 2100, whereas raising the discount rate to 7 percent cuts that present value to just 50 cents. Indeed, by using a 7 percent discount rate (among other tricks), the Trump administration was able to lower its estimate of the SCC to a trivial level at which no mitigation efforts at all would pass the cost-effectiveness test.

Unfortunately, there is no objective way, either in economics or climate science, to determine the “right” discount rate. For shorter-term projects like building a wind farm, which have clear financial costs and benefits, we can use market interest rates. However, there are no financial markets that operate over the hundred-plus-year time horizon of climate change.

Some IAMs try to get around this by positing infinitely long-lived individuals or by assuming that each generation cares as much about the welfare of the next as it cares about its own welfare. But those are only convenient fictions. The reality is, some people care a lot about the fate of future generations and others do not. We can refer to those diverse feelings as “discount rates,” but regardless of what we call them, they remain subjective.

What’s more, the subjective elements of SCC calculations do not end with decisions about discount rates. Many proponents of stronger climate action would like to add additional elements to the IAMs. For example, writing in Nature, Gernot Wagner, David Antoff, and colleagues argue that models should include risk-weighting of hard-to-quantify but potentially catastrophic climate tipping points. They also want to include equity-weighting to reflect the likelihood that a poor person would subjectively suffer more from $1 worth of climate harm than a rich person would. Finally, in their view, intangible ethical considerations should explicitly enter into the choice of discount rates.

Personally, I agree that ethics, equity-weighting and tipping points should be given due consideration in formulating climate policy. The problem, as Nicholas Stern, Joseph E. Stiglitz and Charlotte Taylor noted in a recent paper, is that doing so makes the plausible range of SCC estimates even wider than before. In becoming more inclusive, the whole idea of the SCC becomes more subjective, less scientific and less suitable as an anchor for policy.

Politics Is the Only Way 

It should be clear by now why it is misleading to say that the target-consistent approach to climate policy “replaces scientific assessments of damage … with subjective judgements about political targets and choices.” The SCC, at best, is only a partially scientific assessment of damage. The IAMs from which SCC estimates emerge are, by their nature, replete with subjective judgements — everyone who works with them knows that. The difference between a target-consistent approach and an SCC-based approach is that the former openly acknowledges its subjective components while the latter hides them behind inscrutable lines of computer code.

So, if we cannot rely on science for more than a supportive role in climate policy, where do we go?

Markets might seem one alternative. Markets are, in fact, good at dealing with subjective evaluations when the choices in questions are individualized. Choices of occupation, diet, lifestyle and fashion largely fit that pattern. Once goals are established, markets can help find low-cost ways of reaching them. That is the idea behind using carbon pricing to help find efficient ways to achieve climate targets. But markets are of little use in setting climate targets, since there is just one climate for everyone.

That leaves politics as the default method for making choices that are inherently subjective, yet cannot be individualized. But politics itself means different things in different countries. In some, political institutions feature elected executives and legislatures constrained by independent judiciaries. In others, they feature hereditary kings or self-appointed dictators. In still others, political institutions are weak or deadlocked so that decisions are made by default — and the default decision is often to do nothing.

Although there is no example I know of, it is possible to imagine a country in which power over climate policy is delegated to a council of scientists. Such a council could model the factors driving climate change and calculate probability distributions for sea-level rise, storm damage and so on. But those things alone would not make it possible to set an optimal carbon price or write an optimal set of climate regulations.

To finish the job, the scientific council would need to introduce relative weights for climate costs and benefits in the present or the distant future; weights for effects that occurred in different countries or that affected various economic or ethnic minorities in different ways; and weights attached to unlikely but potentially catastrophic tails of damage functions. There would be no objective way to assign those weights.

The council itself would have to resort to politics — to voting, to expulsion of dissenting members, to granting some members dictatorial power to override others. Or the scientific council itself could succumb to political paralysis and fail to agree to do anything — a result that would itself be a political decision.

• • •

The point here is that public policy decisions are inescapably political. In a democracy, there is nothing illegitimate in advocating a policy of going slow on climate action or even of doing nothing until various scientific and technological uncertainties are resolved. But it should be understood that advocacy of delay or inaction is itself a political position that, at least implicitly, reflects subjective preferences regarding issues such as time preference, risk-weighting and equity-weighting.

None of this is to say that science is irrelevant to good policymaking. Clearly, we should look for the best available estimates regarding the probability distributions of future global temperatures, sea-level rise, storm frequency and the rest. But there is no way to translate those estimates into policies without politics. In the end, an admittedly political, target-consistent approach to climate policy is more intellectually honest than one that represents fundamentally political and subjective choices as something different from what they are.