By Ian Parry
With global leaders set to start signing the landmark Paris Agreement on climate change tomorrow—April 22 is Earth Day—at the United Nations in New York, countries will embark on the potentially difficult and contentious issue of setting prices for greenhouse gas emissions, most importantly carbon dioxide (CO2). Our back of the envelope calculations show that most large emitters will need to charge anywhere from $50 to $100 per ton or more (in current prices) by 2030 to meet their commitments to reduce carbon emissions.
These are quite sizeable numbers: for example a $50 per ton CO2 charge would add $0.12 cents per litre to the pump price for gasoline. And it would almost triple world coal prices.
As discussed in a previous blog, there is widespread recognition (just ask people in business and finance) that on average and over time, CO2 reduction pledges are best met through a robust and predictable emissions price. And the best way to implement CO2 pricing—and provide across-the-board incentives for investments in clean technology—is to charge for the carbon content of fossil fuels.
Some background: More than 190 countries came together last December to pledge to do their part to halt global warming. They also agreed to procedures for evaluating progress and updating these pledges. A typical commitment is to cut greenhouse gas emissions by around 30 percent by 2030 relative to emissions in some baseline year.
Level of carbon pricing
The above figures for carbon pricing come from some highly simplified calculations presented in the table below (assuming carbon pricing is the main mitigation instrument to implement the Paris pledges). Not all countries may need carbon prices of $100 per ton. For example, prices needed to meet carbon targets in Russia are lower than in other countries as Russia’s emissions have already fallen substantially relative to 1990, the benchmark year for its emissions pledge.
But we should be under no illusions about how far we have to go: only 12 percent of global emissions are currently covered by pricing systems and typically with prices below $10 per ton.
What about potential revenues from carbon pricing? Typically, these are well above 1 percent of GDP for the price levels (see table). These large revenues could allow, for example, substantial cuts in burdensome taxes on labor and capital.
A two-step approach
How do we assess needed carbon prices? First, we project a country’s future use of coal, natural gas, and petroleum products in the absence of new mitigation policies. This determines ‘business as usual’ (BAU) CO2 emissions and it depends on income growth, trends in energy efficiency, possible shifts in the future fuel mix from technological and price changes, and so on. Second, we infer the relation between CO2 prices and emissions using assumptions about the responsiveness of fuel use to CO2 pricing—a large empirical literature provides some sense of these fuel price responses based on previous experience.
Inevitably, there is a lot of uncertainty in these price estimates. So countries will need transparent and predictable rules for periodically adjusting CO2 pricing trajectories if energy systems evolve in unexpected ways. Emissions prices should also be phased in progressively to allow firms and households time to adjust, limiting risks of scrapping existing capital well before the end of its useful life. For example, a country requiring an estimated CO2 price of $75 per ton by 2030 might plan to increase the emissions price by $5 per ton each year starting this year.
In their own interests
The good news is that many countries are appreciating that carbon pricing can be in their own interests. The fiscal and domestic health benefits (from cleaner air) can both be substantial, before even counting the benefits for the global climate. And thanks to the Paris Agreement, governments will at least be under considerable pressure from their peers to demonstrate progress on meeting their commitments to reduce carbon emissions, which could create important momentum for carbon pricing.