On the cost of reducing CO2 emissions

I read in the newspaper that the Japanese were turning down the air conditioning to conserve energy this summer. Their government estimated that raising office temperatures from 79.2 to 82 degrees would reduce carbon dioxide emissions by 2.9 million tons between June and September -a typical household releases 0.5 tons of the gas in one month. (WSJ article; read it here if you're not a subscriber.)

To stay cool, office workers were tossing suits and ties, waving paper fans, and applying cooling pads to their foreheads. Some people just couldn’t stand it -82 degrees “is an energy-saving environment –not a human environment.”

I admire Japan’s commitment to save the planet, but this measure strikes me as foolish. I’m sure it's less costly to improve the efficiency of some crumbling power plant in Russia instead. Is it?

Vattenfall, a Swedish utility, has estimated the average cost of reducing one ton of CO2 emissions in different regions of the world. (Yes, what is relevant are the marginal costs, but I don’t have data on those.)

It turns out that the costs of reducing pollution are lower in rich countries than in developing ones (see chart). The differences are large. For example, it costs almost twice as much to reduce emissions in China as it does in North America.


Vattenfall doesn’t disclose the methodology of its study, but others have confirmed the result. David Maradan and Anatoli Vassiliev of the University of Geneva report in a 2005 research paper that marginal abatement costs, in terms of Gross Domestic Product, are 4.5 times higher among the poorest countries than among the wealthiest ones.

Jayant Sathaye and Amol Phadke of the Berkeley National Laboratory calculate the cost of reducing emissions in India trough the use of liquefied natural gas (LNG) plants in place of coal plants –natural gas is cleaner than coal. They find that the cost of producing electricity with LNG is higher in India than in the US, but that running coal-fueled plants costs the same in both countries. Therefore, abating pollution by switching to natural gas is cheaper in America.


There are multiple reasons for this finding. For example, differences in the cost and heating value of fuels, and tariffs on imported equipment. Costs of capital and financing are less likely to play a role, since capital is so mobile.

Whatever the reasons are, if the chart above is more or less correct, then the richest countries in the world are emitting too much CO2, and developing ones are releasing too little.

A global cap-and-trade system would greatly benefit the US. In such a system, each participant receives a limited number of emission rights, and is allowed to buy and sell those in the market. Because abating emissions is cheaper here, the US would sell its unused allowances to developing countries and thus receive big transfers from them.

Greg Mankiw says that a cap-and-trade system would result in a “large transfer of national income [from Americans] to the Chinese.” He’s assuming, correctly I think, that allowances would be allocated based on current population or on historical emissions. Mankiw concludes that, because the US now emits more CO2 per person than China, Americans would have to buy allowances from them. But the question is not how much we are polluting now, but how much CO2 we would be emitting in a world where countries can trade emission rights. And in that hypothetical world, the US would be a net winner.

The vision of coal-fired power plants in China spewing pitch-black fumes without control is dreadful. But it seems like it’s in the world’s best interest to keep them running, and pay the Japanese to turn down their air conditioners.

Does it seem right that abatement costs are lower in developing countries? Why do you think that is? Would the US benefit from a global cap-and-trade system, as opposed to a carbon tax? Leave your answers, or whatever you have to say, in the comments.

Technorati tags:
economics, environment, CO2 emissions, abatement cost, cap-and-trade

Raising the gasoline tax will not curb CO2 emissions (very much)

Fuel taxes in the United States are among the lowest in the world. In 2006, the average American paid $0.39 a gallon in fuel taxes, including federal and state rates. People paid $2.04 in Japan, $3.81 in France and $4.23 in Britain, according to the Federal Highway Administration. Moreover, the federal government has not raised the tax since 1993, making an increase politically justifiable.

Motor gasoline plus diesel fuel accounted for 1,604.6 million metric tons of CO2 emissions in 2005, second only to power generation as the leading source of carbon dioxide emissions (see chart, created with data from the Energy Information Administration). An increase in the gas tax is also easy and cheap to implement, unlike the cap-and-trade system, another solution to CO2 emissions that the media are discussing .


The gasoline tax can accomplish several good things, and do it very quickly. It raises funds to build roads and bridges. It can spur economic growth, by stimulating saving and investments, if we compensate it with a reduction in income taxes. But there is one thing the gas tax will not be able to do, and that is curbing carbon emissions in the near future.

When the government increases the fuel tax, everybody pays more for their gasoline. But not everyone is equally able and willing to cut back on their driving. UPS has no way to deliver packages to your door other than with a truck, but city dwellers can take the bus downtown instead of driving. A lot of people would take the train downtown on a weekday, when the service is frequent, but it would be a pain to do so on a Sunday, when the train comes once every hour.

An economist would say that different car users, or a given car user under different circumstances, have different elasticities to the price of gasoline. In the examples above, city dwellers’ demand for gasoline is more elastic than UPS’s. In theory, as well as intuitively, more elastic users should be taxed more than inelastic users, because the former reduce their gas consumption more than the latter, for a given increase in gas prices.

We cannot charge different gasoline tax rates to different people or at different locations. Arbitrage would render that policy ineffective. So federal and city authorities should implement mechanisms that increase the effective cost of driving, not of gasoline. It is much easier to tax driving at different rates than it is to make the gasoline tax vary by location or time.

Here are some examples of how to tax driving, while at the same time taxing elastic drivers more heavily:

1. Charging a fee to access the city by car. The intention here is to make commuters switch to public transportation once they reach the metropolitan area. To make this measure effective, it might be necessary to build parking lots near train and bus stations in peripheral neighborhoods. Cities might also need to increase the frequency of bus and train services. The city of London has implemented a city access toll, with great success, and New York and San Francisco are thinking about it.

2. Variable price of city stickers. City stickers should be more expensive in neighborhoods where driving is less necessary.

3. Price of parking. The price of parking should vary according to the availability of public transportation connections. Parking spots in areas which can reached by public transportation or on foot should be more expensive than those in isolated neighborhoods. Likewise, the price of parking should vary during the day and over the week, becoming more expensive during business hours, when subways and buses have more frequent services, and cheaper on weekends and late at night. The city of San Francisco has experimented with dynamic pricing of parking.

4. Taxicab surcharges. Cities could apply a surcharge to cab rides within the metropolitan area covered by subway and bus lines.

5. Taxing commercial and private vehicles at different rates. Fuel taxes and city access fees should be lower for commercial trucks. This may come across as counter-intuitive: after all, trucks pollute more than cars. But remember, the principle is to charge more to more elastic drivers. Trucking companies have no other way to bring goods to the city core; you can do your shopping on the internet when gas gets very expensive. (By the way, Austan Goolsbee has found empirical evidence that people shop more online in states where shopping in physical stores is relatively more expensive.)

Many cities in the US lack the basic public transportation networks to make the proposed measures work. Those places should go ahead with the measures nevertheless. The taxes and fees I propose will raise large tax revenues quickly, which city governments can then use to create public transportation services. Because creating that infrastructure will take time, cities could borrow money upfront, to be repaid with toll and parking revenues, and run temporary shuttle and bus services from the periphery to central areas (bus lines are faster and cheaper to set up than trains and subways). Additionally, or while those cities develop adequate public transportation networks, businesses should be given incentives to have their employees work from home, say one day a week.

The measures I call for have the added attractiveness of reducing traffic congestion. If the government’s only purpose is to cut carbon dioxide emissions though, I would propose exempting electric cars from city access tolls, surcharges, etc. and reducing them for fuel-efficient cars.

An undiscriminating, flat-rate gasoline tax is very effective at raising revenues, but it is a blunt instrument for the purpose of curbing carbon dioxide emissions. We should come up with fees, taxes, regulations and, in general, clever mechanisms that effectively reduce those emissions, as cheaply and quickly as possible.

Technorati tags:
gasoline tax, carbon tax, CO2 emissions, carbon dioxide emissions