Once, on a flight out of Albuquerque, I was reading a glossy magazine article criticizing gross national product when I encountered the following quote from a “young radical”: “Don’t tell me about your GNP. To me, it’s really Gross National Pollution.”
Actually, it is completely false. Our output measures do not count pollution. They include goods like cars and services but not carbon monoxide (CO) pumped into the air. The point is worth considering: measures of national output do not adequately correct for pollution or other spill-over effects of the economy. That is why a serious effort has been made to develop accounts that properly reflect these factors. “Green national accounting”, however, has turned out to be extremely difficult terrain.
Most discussions of national output refer to gross domestic product: the value of the goods and services produced by the economy, less the value of the goods and services used up in producing them. Gross domestic product (GDP) thus includes consumption goods like food, investment goods like new houses, production for government, and adjustments for foreign trade.
GDP has its critics. One elementary problem is that it includes gross investment and does not subtract depreciation. Hence, it includes all new houses built in a year but does not subtract the houses that are burned up by wildfires. A better measure would include only net investment as part of total output. It is also useful to focus on the income of residents, which would be represented by national product rather than domestic product. By subtracting depreciation from GDP and looking at the income of residents, we can obtain net national product (NNP).
If NNP is a sounder measure of a country’s output, why do national accounts rely on GDP? One reason is that depreciation is difficult to estimate, whereas gross investment can be estimated fairly accurately. Moreover, while NNP includes all the goods and services produced by residents of the country, it excludes important costs that are not produced and sold in markets. For example, it includes the electricity produced and sold by an electric utility but not the health damages caused by the pollution that the utility emits.
So, the problem with GDP and NNP, then, is that they do not include a subtraction for pollution. By contrast, a measure of green output would include important non-market goods, services and investments along with corrections for negative externalities.
Measuring the non-market
Most specialists would agree that it is important to correct for pollution, climate change and other non-market activities and externalities in the economic accounts. But how can this be done in practice? How could we figure out how to subtract the economic harm done by water pollution or carbon-dioxide (CO2) emissions from the value of food and shelter?
This seems like an impossible task, but the late Martin L. Weitzman of Harvard University showed the way. Weitzman’s approach, which has since been incorporated in green accounting (or full-income accounting), is actually quite intuitive. The idea is to extend the standard national economic accounts – which cover market transactions – to include non-market activities or processes. The standard accounts collect data on the quantity of production and prices (of apples, lumber, gasoline, cars and so on), calculate the values as the product of prices and quantities, and then calculate total national output as the sum of the values of final outputs sold to consumers and other sectors.
The Weitzman approach assumes that the harmful externalities are priced and then adds their value to the totals. But here, harmful activities have a negative price because they are “bads” rather than “goods”. If there are five million tons of air pollution in a year, and the damage from air pollution is US$100 per ton, US$500 million would be subtracted from national output.
This process would seem to be straightforward, except that the concept of the “price of pollution” can be puzzling. The price of potatoes is observable in the grocery store, but what is the price of the CO emitted from a truck? From the point of view of the firm and its commercial accounts, the price is zero, which is why there is no item called “sales of CO air pollution” in the national economic accounts. But the cost to people is not zero, because CO pollution damages human health.
According to the Weitzman approach, if each ton of CO does US$100 of damage, that is the appropriate price to use when subtracting the costs of pollution and other externalities in calculating green output. But, of course, actually calculating the costs of such pollution and other externalities is extremely difficult, because the data are sparse at best (if they exist at all).
Owing to this problem and other measurement difficulties, no comprehensive environmental accounts exist for any country. But we can use the sparse existing research to get a flavour of how environmental accounts have been or could easily be constructed to account for things like greenhouse gas (GHG) emissions and air pollution.
From a conceptual vantage point, the starting point is NNP. In developing an estimate, we can calculate both a level correction and a growth correction, with the level correction adding or subtracting the estimates of the negative externalities or other omissions from NNP. If these externalities are growing, this will reduce the green growth rate, whereas if they are shrinking, the green growth rate will increase.
Accounting for carbon
Let us now turn to some actual cases. The first example is the impact of the climate-change externality, particularly CO2. This is so simple to calculate that anyone can do it on a spreadsheet. The idea is to obtain estimates of the quantity and the price and then correct the accounts for the total. You would begin with a measure of GHG emissions (in this case, CO2) and then multiply the quantity by the price of emissions, as determined by the US government’s estimate of the “social cost of carbon”.
In 2018, the United States emitted 5.3 billion tons of CO2 at an estimated social cost of US$44 per ton, meaning that US$233.2 billion would be subtracted from the US$15.9 trillion of output that year – a level correction of 1.5%.
But now we need to look at the growth effect. As the table shows, US climate-corrected NNP actually grew between 1973 and 2018, reflecting the fact that emissions declined by 2.2% per year relative to output. Green NNP rose faster than conventional NNP. The negative growth effect is counterintuitive until we realize that it arises because CO2 emissions declined, making their effect on green output larger at the beginning than at the end. Thus, correcting for CO2 emissions lowers the level estimate of output but raises the growth rate of output by a tiny amount.
One pertinent question, then, is what the growth correction would be with more ambitious climate targets, such as the 2° Celsius limit on global warming set by the Paris climate agreement. This would imply a much higher social cost of carbon and therefore a much higher price of carbon in the calculation. One estimate is that the carbon price would be more than five times higher with the more stringent target. Using the same method as shown in the table, the level correction for the 2°C target is much larger, at 8% for 2018, and the growth correction is also correspondingly larger.
When environmental costs are larger, this implies that true output is also lower than conventionally measured output. But when environmental costs are declining, the growth correction is both positive and larger.
The price and cost of pollution
Now we can turn to a more complicated problem: air pollution. This includes some of the deadliest and costliest externalities, such as those associated with burning coal and other activities. Most of these activities are regulated in the US, but few are priced at a level that reflects their social costs.
In a 2011 study, Nicholas Z. Muller, Robert Mendelsohn, and I estimated air pollution damages in the standard manner by multiplying the price (damages per unit of pollution) by the quantities of five major pollutants (nitrogen oxides, sulphur dioxide, fine particulate matter, ammonia and volatile organic compounds) for 10,000 sources. What we found is that the total damages as a percent of NNP declined from 6.9% of output in 1999 to 3.4% of output in 2008. These corrections are clearly a substantial fraction of output and are also a much larger fraction of the output of the highly polluting industries.
Again, the growth effect was counterintuitively negative, because, as with CO2, the pollution subtraction at the end of the period was smaller than the subtraction at the beginning. The decline in pollution had the effect of raising total NNP growth from 2.03% per year to 2.45% per year – a substantial impact that has not been emphasized in discussions of the economics of pollution.
These two examples – greenhouse gases and air pollution – hardly exhaust the areas of interest in green accounting. Other relevant sectors would include forests, water, road and highway congestion, and toxic waste, but there are few estimates for those. Meanwhile, estimates of augmented accounts have been produced in other areas such as health, home cooking, family care, and leisure. But while these can have substantial effects on total output and on growth, they generally fall outside the purview of green accounting.
What green accounting shows
Here, then, is the summary of green national output: When we include impact estimates for resources and the environment that are currently excluded from the conventional national accounts, the difference in terms of the level of output can be substantial. A rough estimate is that including the impact of excluded sectors such as those reviewed here would subtract on the order of 10% of output from the US; but, because the research is incomplete, the total might be larger.
Correcting this omission, however, will tend to raise the growth rate of green output, at least for the US over the last half-century. The reason is that most measures of pollution have been declining relative to the overall economy – the result of cleaner power plants, factories and automobiles. It is the growth of pollution relative to other goods and services that affects the growth rate. The growth effect in the sectors examined to date is on the order of +1.5 percentage points per year – a substantial number that would add up considerably over the years. True, major sectors are missing from the estimates. But, while approximate, these numbers do cover some of the most important externalities.
The finding that US environmental policies are adding to genuine economic growth is important for debates about environmental policy. I would count this as a major victory for the green movement. The reason for this surprising finding is interesting. If we go back a half-century to the dawn of environmental regulation in the US, externalities such as air pollution reflected activities for which the marginal benefits of reducing pollution were far greater than the marginal costs. Environmental policy was, in effect, picking low-hanging and inexpensive fruit, reducing health and other damages substantially at minimal cost.
If we look only at the standard economic accounts, we will largely miss the improvements in economic welfare associated with picking the low-hanging environmental fruit, because the health benefits of environmental regulation are not counted in the standard accounts. But if we extend our horizon to include external benefits, the past half-century of environmental policies have actually improved growth substantially.
So, if the young radical was to come back today as an old radical, his attitude toward national accounts might be quite different. Those who claim that environmental regulations harm economic growth are completely wrong, because they are using the wrong yardstick. Pollution should be in our measures of output, but with a negative sign. If we use green national output as our standard, then environmental and safety regulations have increased true economic growth substantially in recent years.
This commentary has been adapted from The Spirit of Green: The Economics of Collisions and Contagions in a Crowded World (Princeton University Press, 2021).
William D. Nordhaus, a Nobel laureate in economics, is a professor at Yale University.
