AN ECONOMIC ACCOUNTING FRAMEWORK

Forests interact with the economy in more ways than any other resource. This means that a system of economic accounts taking the SNA as its starting point must make multiple adjustments related to them. It also means that there is a significant risk of double-counting if those adjustments are not made carefully.

The appropriate starting point for considering how to account for the benefits of forest resources is therefore at a conceptual level. One must: (i) identify the broad classes of adjustments that need to be made, and (ii) determine whether those adjustments represent adjustments to the current accounts or the asset accounts and, if the former, whether they are adjustments to gross or net value added (GDP or NDP if the adjustments are made for the entire economy instead of just forestry). This process must be grounded in economic theory, to ensure internal consistency and coherence. Intuition might not be a reliable guide in such a complex undertaking.

We expect that the economic backgrounds of readers of this report will vary considerably. Consequently, we will keep the details of economic theory to a minimum in the main body of the report, concentrating on rationale and results instead. We present full details in the appendices. The models presented in the appendices draw upon models in the literature by Mäler (1991), Hartwick (1990, 1992, 1993, 1997), Hamilton (1997), and Vincent (1997), as well as upon our own original research conducted for this report.

Economic contributions of forest resources

We develop a model to make our assumptions about forest/economy interactions explicit. The model includes the following aspects of forest resources:

(i) Forests as a source of timber harvested by logging companies and used as a production input by wood-processing industries.

(ii) Forests as a source of tangible nonmarket forest products collected and consumed by households but not bought and sold in markets, for example fuelwood, poles, bamboo, fruit, game, medicinals, and perhaps some portion of the timber harvest.

(iii) Forests as a source of less tangible forest amenities consumed by households. These values include such things as existence values associated with biodiversity and the pleasure gained from scenic vistas.

(iv) Forests as a source of environmental services (ecological functions) that benefit other industries in the economy. The prime example is watershed protection, which benefits downstream agriculture (among other industries).

(v) Forests as a disposal site for airborne pollutants damaging to forest health (e.g., acid deposition).

(vi) Forests as a sink for and a source of carbon dioxide, which potentially damages other industries through global climate change. If CO₂ concentration affects forests themselves, either directly through "CO₂ fertilization" or indirectly through climate changes that alter species composition, such impacts are equivalent to the one described in (v).

(vii) Forests as a source of land for other purposes, in particular agriculture. The flip side of this benefit is, of course, deforestation, which results in the long-term loss of forest-related benefits.

(viii) Forest management as an activity involving the use of both variable inputs (labor, materials) and fixed factors (human-made capital).

Although other contributions can be imagined, this is a quite comprehensive list. It spans the full range of adjustments that existing studies have tried to make. Although it might appear to omit one aspect of forest resources that (appropriately) receives much attention, i.e. biodiversity, in fact it includes biodiversity values, implicitly if not explicitly, in categories (ii)-(iv).

Adjustments to current accounts

In a model containing these elements, several forest-related adjustments to the national accounts are theoretically justified if one wishes to convert production measures to welfare measures. Before presenting details of the model, let us summarize the main findings. Consider the current accounts, GDP specifically, first. As noted in the introduction, GDP can be calculated either by adding up value added across all industries, by adding up expenditure on final uses (final consumption and savings) of goods and services, or by adding up primary incomes from production activities. The same adjustments to the level of GDP must be made in all three cases. These adjustments relate to the consumption-related nonmarket benefits of forests described in (ii) and (iii) above. Specifically, one must add to GDP the value of household consumption of tangible, but nonmarket, forest products like fuelwood and less tangible forest amenities. Consumption of these types of goods contributes to welfare, but it is not ordinarily recorded in GDP because the goods are not bought and sold in markets (in the first instance) and fall outside the production boundary (in the second).¹¹

Although adjustments to the overall level of GDP are required for the consumption values described in (ii) and (iii), they are not required for current production values of the types described in (iv) and (v). That is, even though forests provide beneficial environmental services to other industries and suffer damage from pollution discharged by other industries, no adjustments must be made to GDP to reflect these current production linkages. The reason is simple: GDP already reflects the impacts. For example, if watershed protection services provided by forests positively influence current output of agriculture (due, for example, to reduced flooding), and forest cover declines, then agricultural output will also decline. GDP will be lower as a consequence. Similarly, if pollution damages the current stock of commercial timber, then timber production, and thus GDP, will be automatically lower.

Of course, the impacts of forest degradation might persist beyond a single period or might not be felt immediately. Reduced harvests from lower timber growth might not be observed until several decades after the initial pollution damage occurs, and increased flooding might not occur until several years after significant amounts of forest cover are lost. Adjustments for such future impacts must be made, but they should be made to NDP, not GDP. For the same reason, we do not need to adjust GDP for the depletion of timber (item (i)), for net sequestration of carbon in forests (item (vi)), or for deforestation (item (vii)), even though all have persistent economic consequences. We will discuss adjustments for these long-lasting impacts later.

In sum, just two adjustments, both related to nonmarket household consumption values, must be made to GDP: one should add the value of household consumption of tangible nonmarket forest products, and also add the value of household consumption of less tangible forest amenities. In effect, one must add the value of direct consumption of private and public goods that do not pass through markets. One does not need to adjust GDP for current production-related links between forestry and other industries in the economy, even when those links involve nonmarket impacts (as they usually do). Hence, we have:

The foregoing discussion pertained to economy wide GDP. Although adjustments for current production-related impacts are not justified at that level, they are for GDP at the industry level, i.e. for industry-specific value added. What is needed is to reallocate some value added from other industries to forestry in the case of environmental and pollution disposal services provided by forests. As this is simply a matter of reallocation, the level of overall GDP does not change.

The reallocation is most easily understood by working with GDP as defined by the sum of value added across industries (i.e., GDP from the production account). Assuming a four-industry economy with balanced trade, conventional GDP is given by:

Ignoring taxes and subsidies, value added is the sum of employee compensation and operating surplus, with the latter defined as the value of output minus employee compensation minus expenditure on intermediate inputs. From a social standpoint, operating surplus in manufacturing is overstated because that industry receives free pollution disposal services from forests. Similarly, operating surplus in agriculture is overstated because agriculture receives free watershed services from forests. For both reasons, operating surplus, and therefore value added, in forestry is understated.

The necessary adjustments are as follows:

No adjustment is needed for the logging industry.

NDP, the sum of final consumption and net investment, provides a truer measure of income than GDP, the sum of final consumption and gross investment, because it deducts the consumption of capital. That is, it deducts depreciation. Conventional NDP includes net investment in just human-made capital. As demonstrated in Appendix 1, NDP should also include net investment in forest capital (and other forms of natural capital, and human capital as well, for that matter).

But net investment is not necessarily adequate for analyzing economic sustainability. Net investment is the value of the change in quantity of capital. It equals the change in value of capital — which is what matters for sustainability — only if no holding gains/losses occur. In the presence of holding gains/losses, net investment does not equal the change in value of capital. Then, NDP defined as the sum of final consumption and net investment does not provide an accurate long-run welfare measure, even if final consumption includes nonmarket goods and even if net investment is defined across all forms of capital. For example, if the value of the timber stock falls due to a permanent decrease in log prices, then long-run welfare prospects fall as well. Excluding this holding loss from NDP would cause NDP to overstate long-run welfare prospects (Asheim 1997).

There are two ways of addressing this problem. One is to continue to define NDP as the sum of final consumption and net investment, and to define a new measure that is the sum of NDP and holding gains/losses. The alternative is to broaden the definition of NDP to include holding gains/losses. Which approach is chosen is purely a matter of convention. In the end, the sum of final consumption plus net investment plus holding gains/losses is what matters, regardless of what we call it. For the sake of simplicity, we will define "adjusted NDP" as including not just the sum of conventional NDP and net investment in natural capital, but also holding gains/losses. That is, we define adjusted NDP as final consumption (including consumption of nonmarket goods) plus the change in value (not just the value of the change) of the total capital stock. In keeping with national accounts terminology, we refer to the change in value as net accumulation.

Assuming that the economically most significant aspects of forest resources are the ones listed at the start of this chapter, net accumulation of forest capital can be decomposed into four elements. The first is net accumulation of timber. In the absence of holding gains/losses, the change in the value of the timber stock depends on the balance between timber growth on the one hand and harvest and other factors reducing the stock on the other. The second element is net accumulation of carbon. If we assume that sequestered carbon is proportional to the standing timber volume, then, in the absence of holding gains/losses for either, net accumulation of carbon is proportional to net accumulation of timber. This does not imply double-counting, as the timber and carbon stocks are valued by different procedures (discounted flows of stumpage value in the case of timber, and discounted flows of avoided economic damages in the case of carbon).

The third and fourth elements are also closely related. The third is net accumulation of forestland. Deforestation reduces the stream of future benefits provided by forests: timber harvests, consumption of nontimber products and forest amenities, environmental services, and carbon sequestration will all be less. On the other hand, expansion of forest area increases the future provision of those benefits, so the capital value of forestland can appreciate as well as depreciate.

The fourth element is the net accumulation of converted land. One must not simply subtract the loss of forest capital due to deforestation from conventional NDP; one must also add the capital value of that land in its new use. Otherwise, the accounts will overstate the negative economic impacts of deforestation. If markets function perfectly, then the appreciation of developed land due to forest conversion should at least offset the depreciation of forestland due to deforestation. If, in addition, land conversion is costly, as it typically is (one must not only harvest all the commercial trees but also fell and dispose smaller trees, remove stumps, etc.), then the former will exceed the latter. That is, deforestation should result in a net increase in land-related capital values. Because markets do not work perfectly in practice, however, some of the costs of deforestation are typically ignored when land-use decisions are made by either private parties or the government. Hence, the decumulation of forestland could well exceed the accumulation of converted land. Whether it does or not is an empirical question, whose answer depends on the magnitude of nonmarket forest values and on the characteristics of property rights and other institutional factors.

In effect, the first two elements, net accumulation of timber and carbon, reflect changes in the value of the land area that remains under forest cover during the accounting period, while the second two elements, net accumulation of forestland and converted land, reflect changes in value due to changes in forest area that occur during the accounting period. This observation raises two issues. The first is the issue of how to account for changes in timber and carbon stocks on areas deforested during the accounting period. Should they be included in net accumulation of timber and carbon, or in net accumulation of forestland? There is an obvious risk of double-counting. We recommend the following. Begin by calculating the present value of future timber production (the discounted sum of stumpage value from future harvests; more on this in the next chapter) from the deforested area at the beginning of the accounting period (before it is deforested) and at the end (after it is deforested). These present values equal the asset value of the timber stock. In making these calculations, assume (hypothetically, obviously) that the area will remain under forest cover, i.e. that it will indeed produce future timber harvests and not be deforested. The difference between the two present values, ending value minus beginning value, is by definition net accumulation. Include it in net accumulation of timber. It gives the change in the value of the beginning timber stock, on the assumption that the land stays in timber production. Include the negative of the ending value in net accumulation of forestland. It gives the change in value of forestland due to the loss in future timber harvests caused by the deforestation that actually occurs. Calculations for the carbon stock are analogous.

The second issue has to do with forest degradation, i.e. changes in quality as opposed to changes in quantity or holding gains/losses. The four-element framework includes degradation related to timber and carbon, in that pollution and other forms of degradation (e.g., improper logging) damage timber growth and thus future values associated with timber and carbon. All other values of forests are assumed to be simply proportional to forest area and not subject to degradation. This is obviously unrealistic. This shortcoming can easily be rectified in theory: one simply adds a fifth form of net accumulation, for changes in the value of forests as a source of nontimber, noncarbon values. This fifth item would be calculated as the difference between the present value of future values of forest-related benefits, other than timber and carbon (which are already accounted for separately), at the end of the period, minus the present value at the beginning of the period. This calculation could be done for the forest as a whole, or, if data permit, for individual values (several additional net accumulation entries). Developing a separate net accumulation entry is probably most important to consider for the case of fuelwood, which is the most important nonmarket product of forests in most developing countries and an important cause of degradation (through over harvesting). As in the case of timber and c arbon, net accumulation calculations for forest degradation must be done carefully to avoid double-counting when deforestation is also occurring.

The four-element decomposition of net accumulation is therefore meant to be illustrative, not definitive. Alternative decompositions are possible and probably desirable in many countries. Whatever decomposition format is selected, three points are critical. First, asset values should equal present values of future returns. The present values should reflect quantity and quality changes and holding gains/losses. Second, net accumulation should equal the present value at the end of the accounting period minus the present value at the beginning of the period. Third, in the presence of changes in forest area, net accumulation of elements of the forest capital stock that are treated as separate assets must be accounted for carefully to avoid double-counting.

Assuming that the four-element decomposition is appropriate, the following four adjustments must be made to conventional NDP:

Written out in full, we have:

This may be viewed as the master relationship between conventional current accounts and accounts adjusted for current and future economic contributions of forestry. To recap, GDP needs to be adjusted for economic benefits associated with two items, which carry over into NDP. NDP needs to be adjusted for changes in value (net accumulation) of four categories of natural capital associated with forests. Current values of environmental services provided by forests as inputs into other industries do not require any adjustments to GDP, but they do require offsetting adjustments to value added in forestry and the receiving industries if one wishes to measure social value added accurately at the industry level. The present value of future services should, however, be reflected in adjusted NDP through the net accumulation terms. For example, watershed services are one component of the value of forestland, so net accumulation of forestland should reflect the loss in future value of such services when deforestation occurs.

So far, we have not mentioned any adjustments to either GDP or NDP related to item (viii), forest management activities. That is because no adjustments are needed. In the production account, GDP already includes value added in forestry (employee compensation plus operating surplus); in the use of income account, it already includes forestry-related investments in fixed capital (e.g., construction of forest roads). These accounts correctly exclude forestry’s use of intermediate inputs. As NDP is a forward-looking measure, there is no need to take into account historical expenditures on forest management, despite the fact that the benefits of these expenditures (enhanced timber production) are typically not realized until many years after the expenditures are made. All one needs to do is to account for the depreciation of human-made capital in forestry, as in any industry.

Adjustments to the asset accounts

In simplified form, SNA balance sheets contain three entries:

• Opening value of asset

• Net accumulation (= change in value of asset)

• Closing value of asset.

Opening and closing values are defined as the present value of future returns generated by the asset. For example, the asset value of a timber production forest equals the discounted sum of total resource rent (volume harvested times average net price, or stumpage value) associated with future harvests. If the forest is also an important source of nontimber benefits, then the asset value should also include the discounted sum of net nontimber benefits. Given that discounting is involved, the asset value depends on the discount rate, the frequency between harvests (and production cycles for nontimber benefits, if not annual), and the age of the forest.

If asset markets are efficient, then the market price of an asset should equal the present value of future returns, making it unnecessary to calculate the asset value by the present value method. In practice, the market price of forestland seldom reflects the full asset value, due to the presence of nonmarket benefits like nontimber products, amenity values, and environmental services. Market price usually (though not always) reflects only timber values. Moreover, many countries, especially ones in the developing world, do not allow private ownership of forests, so there is no market price that can be used as even a lower bound on asset value. In such cases, there is no choice but to use the present value method to calculate asset values.

Once the asset values of the opening and closing stocks are determined, through either market-determined asset prices or present values, net accumulation can be calculated by subtracting the value of the opening stock from the value of the closing stock. In theory, net accumulation can also be calculated directly, instead of by taking the difference in asset prices or present values. We review methods for directly calculating net accumulation for timber in the next chapter. Those methods are also applicable to other forest benefits provided on a cyclical basis (e.g., fuelwood). In practice, sometimes one approach is easier, and sometimes the other. Which is the case depends on the type of forest asset under consideration. Direct calculation of net accumulation is likely to be easiest when timber is the principal benefit provided by forests. In that case, however, market prices for the asset (the prices of commercial timberland of various species and various ages) are also more likely to be available. Generally, it seems that when direct calculation of net accumulation is easiest, it is also least essential.

Whether net accumulation can be decomposed into separate elements like the value of increases in the volume of the asset (e.g., timber growth), the value of decreases in volume (e.g., timber harvesting, catastrophic losses related to fires and pest outbreaks, etc.), the value of degradation of the quality of the stock (e.g., reduced growth and excess mortality to pollution damage or a change in species composition due to ecological thresholds being exceeded), and holding gains/losses — that is, whether detailed accumulation accounts can be created — is also an empirical matter. Although such decomposition is not necessary for examining aggregate issues related to the economic contributions of forests, it might provide useful information for policy analysis and planning. For example, if a forestry department wished to assess the potential benefits of an increase in the budget allocation for, say, fire-fighting, a set of accumulation accounts that gave information on current and past losses in timber asset values due to forest fires would facilitate the needed analysis.

In many cases, however, physical indicators from physical forestry accounts or the SEEA might be a more feasible, and perhaps entirely adequate, source of information on the causes of changes in forestry asset values. While recognizing this point, we note here one potentially feasible decomposition for timber production forests that is linked to the methods discussed in the next chapter:

• Decumulation due to timber harvesting in the current period

• Accumulation due to future growth of forests logged in the current period

• Accumulation due to growth of immature forests

• Decumulation due to fires, pest outbreaks, and other losses in mature and immature forests.

A further step would be to divide each of these elements between: (i) estimates based on constant prices (no holding gains/losses), and (ii) the estimated difference in present values due to differences between expected future prices and constant prices (holding gains/losses).

A final point is that, given the multiple benefits provided by forests, and the fact that those benefits have very different characteristics in terms of production cycle (constant flow or cyclical), market status (priced or unpriced), and so on, it is unreasonable to expect that a single format of accumulation accounts can be applied to all types of forest capital. Accumulation accounts need to be tailored to fit the particular types of forest capital that they cover. For example, the structure indicated in the previous paragraph might be perfectly suitable for tracking changes in values of timber stocks, but it is unlikely to be ideal for tracking changes in values of biodiversity stocks (if biodiversity is treated as a separate forest asset). While the lack of a "one size fits all" framework might concern national accountants interested in standardization across assets and countries, it is not a cause of concern when the purpose of the accounting system is to generate accurate and appropriate measurement of forestry asset values and their changes within a specific country.

National accounting matrices

One way to visualize the various adjustments to the current accounts and their relationship to entries in conventional accounts is to construct a social accounting matrix (SAM), which we prefer to call a national accounting matrix (NAM). Rows of a NAM show supplies of inputs (i.e., sources of income), while columns show use of those inputs (i.e., patterns of expenditure). The sum of entries in a given row equals the sum of entries in the corresponding column. That is, a NAM is structured to be in balance in an accounting sense.

Table 1 presents a NAM for a situation in which timber is the only value of forests. There are three primary factors, labor, capital, and forests, and one intermediate input, logs. Labor, capital, and logs are used in the production of consumer goods, investment goods, and logs themselves. There is no deforestation or depreciation of human-made capital. Under these assumptions, NDP according to the income approach is given by the sum of the last row: wages, capital earnings, and total resource rent (so-called Hotelling rent plus inframarginal rent (producer surplus)), minus depreciation of the timber stock. Households own the primary factors, are the source of savings for investment, and consume the consumer goods. We ignore the role of government as a resource owner and saver; in a sense, the "households" category can be thought of as a catch-all category that includes both private and public owners, savers, and consumers. According to the expenditure approach, NDP is the sum of the last column: expenditure on consumer goods and investment goods, minus depreciation. Appendix 3 provides a more detailed analysis.

Table 2 shows additional entries to the NAM when the forest provides amenity values. In the last row, forest-owning households now receive income from selling forest amenities to other households. The expenditure by the receiving households now shows up in the last column. If the amenities are nonmarket, which is the usual case, then monetary transactions do not occur, and these cells need to be filled in using estimates from nonmarket valuation methods. Appendix 4 provides a more detailed analysis of this NAM.

Table 3 presents the final and most complicated NAM. The value of the standing forest is again just for timber, as in 1, but now forestland has an alternative use, agriculture. Hence, there is an additional primary factor of production, agricultural land, which is generated by clearing forestland, a process that requires inputs of labor, capital, and logs. Both the income and expenditure versions of NDP now include additional terms related to deforestation and development of new agricultural land. Note in particular that the last column includes the difference between the price of forestland and the price of agricultural land, not just the former. That is, the NAM reflects both the depreciation of forestland due to deforestation and the appreciation of agricultural land due to development of cleared forestland. Appendix 5 provides a detailed analysis of the NAM in Table 3.