In the thrilling world of multinational industrial policy, it’s about as high-stakes a fight as you can get.

Every year, the world’s governments spend hundreds of billions of dollars making it cheaper to extract and burn fossil fuels. Almost as regularly, their representatives get together and beg everyone else to stop doing that. Then they go home and keep doing it themselves.

The pattern has worn on for more than two decades. Way back in 1997, the Kyoto Protocol—the first international treaty aimed at fixing global warning—called for governments to stop subsidizing all “greenhouse-gas-emitting sectors.” That didn’t happen, so, in 2009, the leaders of the G20 nations resolved anew to “phase out ... inefficient fossil-fuel subsidies.” Three years later, President Obama declared that “a century of subsidies to the oil companies is long enough.” In 2016, when G20 leaders met in China, they again “reaffirmed” the need to end subsidies.

Somehow, all those affirmations didn’t get the job done. Governments are still subsidizing oil extraction today, to the tune of about $400 billion per year. And climate advocates continue issuing unheeded proposals to cut those subsidies as a way of reducing greenhouse-gas pollution.

But maybe all the rigamarole isn’t worth it. A new study, published this week in the journal Nature, argues that withdrawing subsidies wouldn’t have as large an effect as anticipated. In both the world’s richest and poorest countries, canceling fossil-fuel subsidies would neither significantly reduce carbon-dioxide pollution nor increase the amount of investment in renewable energy between now and 2030.

Only in countries in a sort of middle tier—moderately wealthy places that export vast amounts of oil and gas, like Russia, Venezuela, or Saudi Arabia—would cutting subsidies lead to major declines in emissions.

And even on a global scale, slashing fossil-fuel subsidies would do less to help the climate than would universal adoption of the Paris Agreement on climate change, the study argues. That accord—which would only hold global warming to about 3 degrees Celsius, failing to hold off environmental devastation or dangerous sea-level rise—would nonetheless avert between four and eight gigatons of carbon-dioxide pollution every year. Killing subsidies would only prevent 0.5 to two gigatons of pollution annually.

“We’re not advocating keeping subsidies. We’re just advocating a more regionally differentiated discussion of them,” said Jessica Jewell, an author of the paper and a political-economy researcher at the International Institute for Applied Systems Analysis in Austria.

“A lot [of attention] has been focused on subsidy removal in [developed] OECD countries, whereas when you look at our results, the discussion politically should be on focusing on subsidy removal in oil- and gas-exporting regions,” she told me.

This finding is, she admits, diplomatically challenging. Countries in North America and Western Europe—except for the United States—have historically pushed for a more aggressive global climate policy. But it’s in the countries most resistant to reducing emissions—Russia, Saudi Arabia, Venezuela, and others near them—that slashing subsidies would have the biggest effect.

The new paper provides a useful global context for arguments happening in many world capitals. Most governments spend most of their subsidy money on the consumption side—that is, they help poor and middle-class people buy fossil fuels. Cutting that support can be ethically and environmentally tricky. In India, for example, cutting the subsidies sometimes led to increases in greenhouse-gas emissions—because the country’s poorest citizens, unable to afford kerosene, started burning even dirtier fuels like firewood or charcoal.

The situation isn’t any simpler in rich countries. Take the debate over subsidies in the United States.

In America, it’s not clear how much the public pays to cushion oil, gas, and coal companies. The Council on Foreign Relations, a nonpartisan think tank, estimated in 2016 that the federal government spends about $4 billion every year on tax breaks for fossil fuels. But Oil Change International, a progressive environmental group, looked at a broader set of federal and state policies last year and put the cost to taxpayers at $20.5 billion. (And even this number leaves out some subsidies, like the federal program that helps families pay their heating bills.)

No matter how you estimate them, would cutting these subsidies do any good? There, again, it also depends on whom you ask—and what assumptions they make. Gilbert Metcalf, the economist who arrived at the $4-billion figure, found that ditching the federal tax subsidies would only raise global oil prices by about 1 percent—the equivalent of at most two additional cents per every gallon of gasoline at the pump.

As such, he concluded the subsidies were a waste of money, because they didn’t make gas or electricity much cheaper for Americans. Their effect on global oil prices was just too small. But by the same token, he didn’t think canceling them would reduce greenhouse-gas emissions, either.

A paper published last year in Nature Energy arrived at a totally different conclusion. Though it mostly agreed with Metcalf about the domestic effects of oil subsidies, it looked at how their consequences accumulated over time. (Metcalf focused on their annual effects.) Suddenly, the subsidies seemed to have a gargantuan climate footprint: By 2050, the United States will have underwritten the drilling of an extra 17 billion barrels of oil, enough to emit over 6 billion tons of carbon dioxide.

Both papers, in other words, thought America should trash its subsidies—they just had different reasons why. They also had different ideas about how the world would respond to the change. The Nature Energy paper believes that killing U.S. subsidies would decrease emissions worldwide. Metcalf argues that other countries will just fill the hole that America left in the market.

Looking at the world as a whole in her new paper, Jewell takes the same view as Metcalf. If the United States or Europe were to kill their subsidies, she argues, then an oil- or gas-exporting country would just increase their production. “When you think about it, it makes sense because we’re operating on a globally liberalized market,” she told me.

Therefore, she proposes that climate advocates target killing subsidies in oil-exporting nations. “These countries are already facing budgetary pressures because oil prices are low,” she said. “In a place like Saudi Arabia, there’s an opportunity now. For them to decrease subsidies is kind of a win-win.”

Peter Erickson, a U.S.-based senior scientist at the Stockholm Environment Institute and one of the authors of the Nature Energy paper, said “their suggestion to focus on high-income countries is a good one, and is just good policy design if one is concerned about equity or simply efficiency.” But he takes issue with the idea that fossil-fuel subsidies don’t hurt the climate, even in the United States.

“Their conclusion that subsidy removal could reduce global carbon-dioxide emissions by 1 to 4 percent is substantial, and underscores why subsidy removal is an important climate solution,” he told me in an email. “One to 4 percent is only ‘small’ compared to the massive scale of the climate challenge at hand, a task that necessarily involves many complementary policy solutions.”

He also thought it was silly to compare killing subsidies to the Paris treaty. The national commitments “are generally economy-wide pledges that are designed to be attained through many individual policies and actions, including (in some cases) but not at all limited to subsidy removal,” he said.

Or maybe all the numbers are way too small. The International Monetary Fund calculates that fossil-fuel subsidies cost the world about $5.3 trillion in 2015—a number 13 times the size of Jewell’s $400 billion estimate and equal to about 6.5 percent of annual world economic output. But that study takes an especially magnanimous view of subsidies, including the global burden of air pollution and the future cost of climate change as well as the cost of specific government programs. “If you use that number, you get a totally different result,” Jewell said. “You get something totally different from what we got.”

They ignored that estimate for a reason. Jewell and her colleagues chose to calculate only the policies that governments already have in place—like tax breaks or drilling subsidies—and not include every possible cost to the public.

Ultimately, Jewell say her study should prompt environmental groups to be pickier about the kind of political fights they seek.

“A lot of the advocacy is coming not from governments, but from nongovernmental organizations, and there’s not a focus on this regional differentiation. It’s just a global message of subsidy removal for everyone,” she said. “Climate advocacy occurs at different levels, and climate advocates of course have to evaluate what makes sense for their context.”

“But what kind of political animal are you going up against?” she asked. By focusing on all subsidies everywhere, advocates may be stumbling into disaster—unintentionally targeting “a poor person who will wind up on the cover of a magazine because they can’t afford heating oil.”