Accounting for Sustainable Development?

Jesse Alan Gordon, 1992 and 1995

When we calculate GDP, we do not account for the depletion of natural resources. In normal business practice, depletion of capital assets is subtracted from the gross product to get a net product. That's done as "Net Domestic Product" (NDP) for the depreciation of capital assets, but not for environmental assets. The result is that many developing countries have a GDP which is artificially high, because they extract a lot of natural resources, count the sales price in their GDP, but never discount the value of the loss of the natural resources themselves.

For example, when Indonesia sells its oil, the selling price of the oil adds into its GDP. Since a large part of Indonesia's economy is based on oil, the government can decide how much their GDP will rise by setting an extraction rate for oil. The politics of that situation is that when an election comes up, the government makes sure the GDP rises, so there's no "recession" to hinder the incumbent. The economics of that situation is that the country should not be used as a model for growth in other developing countries, since those without a large natural resource base cannot do the same tricks with their GDP.

The solution is to account for the depletion of environmental assets, by subtracting from GDP the value that the resource had before it was extracted. That is, the GDP is decreased by the value lost when the resource was extracted, since that value is no longer owned by the country. The value that a natural resource has when it is unextracted is called its "scarcity rent" (because the unextracted value reflects how scarce the resource is) or its "marginal utility cost" (or MUC). The MUC also applies to environmental amenities other than tangible natural resources, such as clean air, the recreation value of lakes, or the biodiversity provided by forests. When those amenities are lost, by polluting the air, spilling wastes into the lake, or cutting down the forest, the loss in value should similarly be subtracted from GDP. I'll discuss below how to evaluate losses to intangible resources.

If all of the MUCs are summed up, it would represent the total amount of natural resource depletion (NRD, where natural resources includes environmental amenities as well). The NRD is an adjustment to GDP which accounts for the losses of environmental assets, as capital depreciation accounts for the loss of capital assets. The definition of "sustainable development" is that the total amount of the NRD must be reinvested in reproducible capital (assuming that none of the depletion causes irreversible events, such as species extinction).

The MUC for natural resources is the price minus the marginal extraction cost, minus any other distortions which shift the price from its "shadow price," such as monopoly rents, tariff distortions, etc. The NRD for environmental amenities is based on the amount of value lost when the amenity is destroyed. There are a number of means of valuing environmental amenities (discussed below); the NRD can be simply defined as the value before the loss minus the value after the loss. More specifically, the NRD should be the environmental value without the activity causing the loss, minus the value with the activity. The distinction from the before/after definition is that we must also measure what would have happened had we not done the activity which caused the environmental degradation. For example, if we do not extract oil and burn it to create electricity, we would generate power by some other method, which would cause some other environmental degradation. This finer distinction comes into play when doing cost/benefit analyses; it is not so significant here.

To evaluate the losses in environmental amenities, we either look at the effects on marketed goods, or we use a proxy for goods which are not marketed, or we hypothesize a market where none exists. The three examples above correspond to those three levels of market evaluation, as detailed next. Of course, by using the market to evaluate environmental amenities, we are placing a human value on the environment, that is, we are assuming that the only value that matters is that which affects people. "Deep Ecologists" claim an inherent value to environmental amenities independent of their value to humans; that leaves economics and enters ethics, so I won't consider that here.

For tangible natural resources, we measure the value lost because we no longer possess them after we have sold them. The method suggested above is to sum up their scarcity rents, which indicates the value of the unextracted resources. A popular alternative method is to sum up the total asset value at the beginning of the year, subtract it from the total asset value at the end of the year, and consider the difference in value to be the loss due to extraction. The first method is difficult because one must estimate the scarcity rent in markets which are often highly distorted by monopolies (especially for oil), by export subsidies, by import tariffs, etc. Distortionary policies are especially common in developing countries, where natural resources make up a large component of the economy, so this method is most difficult where it is most important to estimate accurately.

The second method is easier to calculate, since one must only tally up the total value of the existing stocks at market value, which is straightforward for tangible resources. The weakness of that method is that if the stock changes during the year, it is interpreted as a gain in value, or a negative NRD. For example, if a new oil reserve is discovered, or if the estimate of an existing reserve size increases, then the total stock of oil may increase even if the quantity of oil extracted was far above the sustainable level. In addition, if the market price of oil rises significantly, the value of the total stock of oil at the end of the year may be higher than the value at the beginning of the year, even if the quantity is much smaller. That problem could be resolved by using economic prices (i.e., with all market distortions removed), but then this method has the same difficulties as the first method. The stock quantity problem is inherent in the method, and makes this method less valid as a measure of sustainability.

To measure the value of intangible resources, we cannot use market prices, since markets for the goods themselves do not exist. From the examples above, air pollution uses an actual market price as affected by the unmarketable environmental amenity; lake recreation uses a proxy market price; and biodiversity uses an artificial market price. For each case, I'll describe the method of valuation, which includes both marketable goods and intangible goods.

For clean air, we measure the direct health benefits to the economy, by estimating the number of working days lost when the air is dirty and gets people sick. The marginal improvement in air quality causes a marginal improvement in health, and fewer employees call in sick; the losses from sick days is the cost of air pollution.

Since air pollution is typically an urban problem, measuring the human health cost is usually sufficient to evaluate the entire environmental cost. For air pollution or water pollution problems which occur outside of cities, we measure the commercial loss to livestock, fisheries, agricultural production, etc. With acid rain, for example, the cost is primarily in lost revenue to fisheries (because lakes are damaged by acidification), and to farmers (because soil becomes acidified and less productive).

We could also measure the loss in value due to limited visibility, for example, from urban smog. There is no commercial loss due to decreased visibility, and no proxy goods apply, so the "contingent valuation" method (discussed below) would apply. Once again, we only measure the effect on human welfare -- we do not measure the decreased welfare of the livestock or fish, only the value loss that their decreased welfare causes to humans. Deep ecologists have their own welfare decreased by smog, and that welfare loss is captured in the same contingent valuation that measures the value that people place on visibility. A deep ecologist would presumably value clean air more highly than other people, and that value would count -- but the loss in value of clean air to animals themselves, or to the earth itself, or to God, does not count.

For a recreational lake, there is a commercial loss as well as an intangible loss. The loss of cleanliness in a lake creates a direct loss to businesses which use the lake as part of their business, such as boat rentals, fishing guides, campsites, etc. The commercial loss is measured directly, by revenues lost, just as we would measure any other direct loss. The intangible component of the loss is how much value people place on their recreation itself, that is, their utility for their enjoyment of the lake, which they do not pay for in cash terms.

For a lake's intangible recreation value, we use a proxy market, by constructing a demand curve based on what people pay to travel to the lake. The "travel cost method" uses people's willingness to pay to travel to the recreation site as an indication of their valuation. With increasing distance from the recreation site, the cost of getting there increases, for both direct expenses (automobile operating costs, tolls, etc.) and opportunity costs (value of time spent travelling). By graphing the cost of travel versus the number of people from each area, a demand curve is constructed, and then the consumer surplus loss can be estimated for marginal changes in supply of the recreation site. This method has been in use for decades and is the strongest in terms of theoretical foundations, because it uses revealed preferences (actual spending) rather than hypothetical payments.

For a forest's biodiversity value, we use a "contingent valuation" method, where people are surveyed as to how much they'd be willing to pay to preserve the forest. People are shown pictures, for example, of a thriving forest, with many species, versus a managed forest with few species. Then they are asked how much they would be willing to pay, say, in increased taxes, for the difference between the two. A demand curve is thus established again, since many people would be willing to pay a small amount, and decreasing numbers of people would be willing to pay a larger amount.

This method is theoretically weak, and expensive to implement, but the US government now accepts contingent valuation in cost/benefit analyses. Problems with contingent valuation begin with the problems of any survey -- the selection process may be biased (only those interested in biodiversity answer the survey); the answers may be strategically high (since the subjects know they don't actually have to pay); the information presented biases the results (different pictures, in the example above, would yield different results). Additional problems arise because it's unclear what is being measured -- people's preference for unspoiled forests, or people's general environmental desires, versus their value for biodiversity per se.

Nevertheless, contingent valuation is the method in use to measure all forms of value which cannot otherwise be measured. Other methods can measure commercial value (losses to business) and use value (losses to recreation). But in many cases, people do not conduct business around environmental amenities, and do not use them personally, but want the amenities to remain accessible so that they have the option to use them, or so that their descendants may enjoy them. People also value unique resources, even in the absence of the option to use them, for their existence itself -- for example, we value the Alaskan wilderness areas because we know that they are pristine, not because we ever intend to go there. Contingent valuation measures the option value, bequest value, and existence value, which no other methods do.

Amorphous values such as those sound so vague that they seem to be beyond the scope of economics. But people's willingness to pay for the existence of an Alaskan park, indicates a human welfare increase. The willingness to pay for the existence of environmental amenities is demonstrated by the proliferation of environmental organizations, who earn their living from collecting donations to protect the existence of unusual species, unique natural resources, and other amenities. Amorphous though the values are, good economics must measure the increase in human welfare associated with protecting the environment.

Now let's look at how those techniques apply to GDP calculations. When oil is extracted, or timber removed from a virgin forest, or a lake sacrificed to development, only the value of the commercial goods produced is currently counted in GDP. To measure sustainable development, we should subtract from GDP all of the environmental losses which the development has caused. Extracting oil means that there is less oil in the ground -- the value that the oil would have had, if we had left it unextracted, should be discounted from GDP. Cutting virgin forests means that there is less forest, less recreational opportunity, and less biodiversity -- the loss of those environmental amenities should be subtracted from GDP as well. Development around an undeveloped lake means that the pristineness of the lake is destroyed -- the loss of people's value of the pristineness should be subtracted from the value of the development.

Environmental economists do not say that development in pristine areas should not occur, nor that non-renewable resources should not be extracted. We do say, however, that their full values should be accounted for. Failure to do so means that inefficient activities will occur, that is, activities which ignore part of their costs. Environmental degradation is a large "externality" of development. Internalizing the costs of environmental losses is necessary to properly assess which development is in the interests of people to do. Unaccounted external costs imply that the activity is inefficient, which means that people's net welfare is not maximized. Therefore, ignoring the loss of environmental amenities decreases people's overall welfare.

Accounting for the loss of environmental amenities is a key component of sustainable development. Extracting non-renewable resources at an unsustainable rate is relatively easy to measure and relatively easy to justify limiting the unsustainable activity. Destroying intangible resources is just as unsustainable -- there are a limited number of pristine lakes, and with the development of each one, the value of the remaining ones rises. Scarcity rent applies just as well to intangible amenities as it does to tangible resources. A country which spends its intangible environmental resources should discount the losses due to that spending, or they risk spending the resources unsustainably, and hence the resources will some day be gone.

The definition of sustainable development is that the loss of the limited resources is balanced by the creation of new reproducible capital. Development increases human welfare, but only if the damage caused by development does not cause more losses than gains. The means to determine if development is sustainable is to account for the environmental losses that development causes. Then environment and development will be seen as complements to one another, instead of as adversaries.

EJSCK Reprinted from "The Harvard College Economist", 1993.
All material copyright 1993 by Harvard University and Jesse Gordon.
Reprinting by permission only.


Jesse Gordon, 1770 Mass Ave., #630
Cambridge, MA 02140
Voice mail: (617) 354-2805

Article Index About Instant Web Page About WebMerchants Next Article