Just as scientists labor to learn where global temperatures will go in the future, economists have worked to forecast the costs of climate policy.
So far, in the United States, the best guesses come in thick, complex documents riddled with equations and graphs.
These are the fruits of economic models, the field's best way of peering decades into the future to see how major trends affect wealth and the distribution of wealth.
Given the uncertainty, it's no surprise that there's a wide range of conclusions. Some studies of the House-passed climate legislation, H.R. 2454, have found burdensome costs for households, on the order of several hundred dollars a year. Others claim a much smaller cost. Still others propose that thanks to energy efficiency, a climate bill might even save Americans money.
As with most economic studies, what goes into it -- the assumptions -- has a lot to do with what comes out. That can make the studies political grist for groups that want to portray the bill's backbreaking cost or its potential to revitalize the economy with "green jobs."
One variable, in particular, has drawn researchers' attention for its potency: carbon offsets.
These widely critiqued projects play no small role in the House bill: It allows up to 2 billion tons' worth of offsets to count against the emissions cap each year. A fraction of these must come from projects in the United States, but a majority could come from abroad.
The secrets of computer modelers
Increasingly, research suggests that depending on how they are modeled, offsets can make the difference between showing a climate bill that is an economic bargain or an albatross.
In September, the Congressional Research Service published a report comparing seven studies of H.R. 2454's cost. Whether they were authored by industry groups, academics or government agencies, CRS found, all agreed that offsets had great potential to make the bill cheaper. Eliminate access to international offsets, and the price of a carbon permit in the United States would jump 60 percent.
"The cases generally indicate that the availability of offsets (particularly the international offsets) is potentially the key factor in determining the cost of H.R. 2454," CRS said.
To economists, the reason is simple.
In principle, a wide variety of projects could count as "offsets," from stopping deforestation to using new methods of plowing land. As the argument goes, these would be cheaper than technological options like adding carbon capture devices to smokestacks or buying electric cars.
The projects are thought to be cheapest abroad, where the "low-hanging fruit" of carbon reduction hang lowest.
Roger Sedjo, a senior fellow at Resources for the Future, used a forestry example. Trees capture carbon, so there is a climate cost when they are cut down. It's not free to convince loggers in foreign countries to stop or slow their work, but it's far less costly than technology, Sedjo said.
In theory, offsets purchased en masse could take the economic edge off the climate bill as U.S. firms, forgoing high-tech emissions cuts here, meet their obligations with low-tech cuts elsewhere.
As more firms do this, there's less demand for the fixed number of carbon permits in the U.S. system, and the price of a permit falls, making climate policy cheaper for those buying permits.
'Doing cheaper stuff abroad' drops costs
"Instead of making expensive reductions at home, basically you're doing cheaper stuff abroad," explained Sergey Paltsev, a researcher at the Massachusetts Institute of Technology and lead author of one study reviewed by CRS.
Paltsev's study found that the House bill would cost households an average of $400 a year. But if offsets became widely available right away, the cost would drop to $180. On the other hand, if they were less available and more expensive, the cost would jump to about $470.
Economists say the point has failed to get much attention on Capitol Hill.
Many of the debates raging in Congress have focused on other aspects, such as the renewable energy standard, who should receive free permits, and whom the bill should exempt entirely.
Nevertheless, both House and Senate legislation show some wariness of offsets. They each allow 2 billion tons of domestic and foreign offsets, but after a certain year, international ones become "discounted" -- to get credit for 4 tons of carbon under the U.S. regime, one has to buy 5 tons of offsets.
The parameters may reflect the broader dispute outside Capitol Hill: that international offsets may be plentiful and cheap, but they are tremendously untrustworthy.
In part, the claim stems from fears that the United States can't easily enforce its rules for offsets in the countries where the projects take place. But more generally, observers don't yet know how to guarantee an offset is bona fide.
An 'offset' that makes global emissions go up
To explain the difficulty, Robert Mendelsohn, a professor at the Yale School of Forestry and Environmental Studies, imagined a country that builds a natural gas-run power plant. This would have a carbon footprint, but the owners could argue that without their plant, a coal-fired generator would have gone up instead.
Subtracting their plant's emissions from those of the coal plant that doesn't exist, these owners could, in theory, sell an offset. But from the environment's perspective, global emissions have increased.
Another hairy issue, Mendelsohn said, is "leakage." When country A builds a natural gas plant, that may divert resources from country B, which wanted one, too. Country B may choose to build a new coal plant, instead, but if country A doesn't know that, it may sell an offset for its natural gas plant -- without accounting for the coal plant it induced.
"It's become largely an administrative question, because it's not purely an economic issue," Mendelsohn said. "What you count, what you don't count, what you assume as the baseline ... it's been a nightmare."
These "administrative questions" end up being why economists can't agree on how to model offsets. Without knowing the framework that will govern offsets, it's difficult to estimate where they will be -- and what they will cost, researchers say. So where there are question marks, they have to make assumptions.
Some project that international offsets will remain unreliable, so they will never make up more than a small share. Others assume offsets will come at a higher price than expected. Some assume that the United States would start using 2 billion offsets right away, while other studies ramp up to that number, to give governments time to come up with enforceable rules.
Tighten the rules and raise the costs
In the models, the stricter the rules on offsets, the less available they become, and the more it costs the U.S. economy. Some of the toughest forecasts come from the Heritage Foundation, which predicts a higher price of carbon than most of the other estimates in the CRS study.
David Kreutzer, senior policy analyst for energy economics and climate change, said that while offsets are a particularly difficult area to forecast, Heritage didn't find 2 billion offsets a realistic possibility.
"I don't think you'll find many people in town, hook them up to a lie detector, that think 2 billion tons is a reasonable goal," he said.
Kreutzer pointed to the barrage of tests that the House-passed bill applies to offsets. A project can't be something that would have happened anyway; it has to last for a guaranteed period of time, and it has to be verifiable.
With such delicate parameters, he said, it's hard to imagine 2 billion tons' worth of projects being found. So Kreutzer used what he thought was a more reasonable assumption.
In its study, Heritage assumed a pool of emission allowances, then added a 15 percent share of offsets, less than what the House bill allows. Combined with bearish projections for nuclear power and carbon capture and storage, Kreutzer said, Heritage ends up having a higher estimate of economic cost.
Output: ammo for the cap-and-trade debate
It's in these assumptions, he said, that all the major differences lie. In September, he said, Heritage held a panel with economists at the major institutions that have written such studies -- from the Congressional Budget Office to Charles River Associates, a consulting firm.
When the differing assumptions were removed from the studies, he said, "we got very close answers. They're not to the penny, but they weren't an order of magnitude difference you see in the numbers reported to the press."
There's another point of agreement among the researchers: Economic models can neither predict the future, nor can they give precise information about it.
The studies are frequently distilled to a signature finding, like a per-household dollar amount or a percent loss of gross domestic product -- statistics that quickly become ammunition in the debate on cap and trade.
But the researchers behind the studies say the specific numbers can't be trusted. There are simply too many variables -- what will happen to energy prices, how fast technology will move, and how consumers respond. And there's always the unpredictable -- events like 9/11 or the financial crisis -- which researchers can't forecast and never will.
What the studies do offer, the CRS said in its report, is a sense of what each policy lever does: for example, that allowing more offsets usually drops the cost of a climate bill.