Chapter 3
Economic considerations on
instruments and institutions

William F. Hyde

INTRODUCTION

This chapter reviews the policy instruments and institutions affecting the production of both commodity and non-commodity outputs of forests. We want to account for interactions between market and non-market valued forest resources, such as timber harvests that damage critical habitat for resources like biodiversity, and also interactions between activities in other, non-forest, sectors of the economy that influence the non-commodity and non-market products of the forest. The conversion of forest land to agricultural production is the common example, but activities in other sectors that indirectly affect the forest through their influences on timber harvests and agricultural land conversion can also have important effects on the non-market outputs of the forest.

We are interested in the policy instruments and institutional mechanisms that reinforce the favourable interactions or mitigate the effects of negative interactions - including both instruments and institutional mechanisms that affect forestry directly and those that create spillover effects from other sectors of the economy. The source of our interest is the increasing global concern for long-term sustainable resource use and the recognition that changing demands over time (and with economic development) increase the importance of maintaining all future options for resource use. With changing demands, some resources that were previously unknown or of low value take on new and greater value. Therefore, long-term sustainability requires that we protect the future availability of all resources. Those resources that have little or no current market value and those that are easily affected by inadvertent spillovers from other sectors of the economy are often the easiest to overlook and the most difficult to protect for future generations.

Unfortunately, forests and forestry occur at the margin of most other economic activity and many forest resources and forest-based environmental services are low-valued and dispersed. This means they are easily affected by inadvertent spillovers from macroeconomic planning, from institutional change and from activities in higher-valued geographically adjacent sectors like agriculture. Furthermore, monitoring and enforcement to restrict undesirable activity in the forest is expensive relative to the values at risk - because forest resources are often dispersed and undetected access to them is often very easy. Nevertheless, undeveloped natural forests are the only remaining sources of some fundamental resources and the social valuation of some of these apparently grows with time and with economic development. Therefore, our vigilance in protecting the forest from inadvertent spillovers is both difficult now and critical for future social well-being.

The consensus opinion seems to be that spillover effects in general have important negative effects on the forest. However, the magnitudes of these effects have seldom been measured and there are mixed opinions, for example, regarding the potentially negative effects of agricultural policy or the potentially more favourable effects of transfers of forest management from state to local institutions. Furthermore, those few attempts to quantify spillover effects on the forest have generally focused on the entire forest. They do not usually consider differential local impacts on the specific non-commodity and non-market forest products and forest resource services. It is important that we take those differential effects into account in this chapter.

The chapter is divided into four basic sections. The first introduces a forest organization or taxonomy that features three characteristics - location, labour (a variable that is related to population for some economic assessments) and local institutions - that determine which forest areas are the sources of most forest resource uses and, therefore, which forest areas are under greater risk and are of greater importance for immediate policy action. The second section incorporates this taxonomy in a discussion of taxes, incentives and regulations that alter the different forested areas and their important commodity and non-commodity outputs. The third section considers adjacent sectoral policy and institutional spillover effects including infrastructure and property rights. The fourth section of the chapter discusses five key forest products and environmental services.

Conclusions summarize those policies and institutions that are most effective in mitigating negative or enhancing positive effects on the commodity, non-commodity and non-market values of the global forest. Various selective policies can have favourable consequences for the protection of particular forest resources. However, the discussion in this final section notes that the general problem of rural poverty may be the greatest source of deforestation and loss of the various non-market valued forest resources. Therefore, general economic development may have a greater positive effect on the protection of these resources than any set of selective and specific policy or institutional improvements. As a corollary, policies that are detrimental to the rural poor may have the greatest negative influences.

A TAXONOMY OF FOREST DEVELOPMENT

This section reviews the fundamental market economic characteristics of the forest and develops an organization for identifying the most likely spillover effects from other sectors of the economy. This organization is also useful for discussing the non-market values of forests, as it assists in judging where non-market values are under greatest risk - as well as for identifying the source of risk and the means for its mitigation.

Consider how economic activity in general affects the structure of forests. A common pattern emerges from observations taken almost anywhere around the world. This is a pattern of new settlement, followed by deforestation and increasing scarcity of forest products eventually combined with regional development, rising prices and the forest investment that limits further depletion of the remaining natural forest. This pattern describes development of any region over time, but it also describes a cross-section of the world today. All regions of the world fit within one stage or another in this pattern. Indeed, many countries contain regions that fit into different stages of the pattern, with some of their regions containing only natural forests on the frontier of economic development while other regions of the same country may contain aggressively managed modern forest plantations.

Figures 3.1-3.3 capture the basic elements of this pattern. They conform to the general pattern of economic geography first proposed by von Thunen in the nineteenth century. They also provide the key reference points necessary for further reflection on investment timing, institutional constraints and the markets and policies affecting forest products and their values.

A new frontier, stage I: Figure 3.1 describes a simple landscape of agriculture and forests at the time of the first permanent settlement. New settlement is generally associated with a level of agricultural use. The value of agricultural land is a function of the net farmgate price of agricultural products - which is greatest when the farmgate is nearer the local market at point A. Land value in agricultural use declines with more limited access (which is closely related to increasing distance) as described by the function V_a. That is, the periodic crop value per hectare minus the costs of growing that crop creates a net value function that declines as we move to less and less accessible land farther and farther from the value centre at A. Households and communities absorb a transactions cost C_r for establishing and maintaining secure rights to this land. This is not a part of the net value calculation represented by V_a. This transactions cost increases as the levels of public infrastructure and effective control decline and the cost of excluding trespassers increases as we move farther and farther from the value centre at A.

Figure 3.1 - A new frontier, stage I

The functions explaining agricultural land value and the cost of secure property rights intersect at point B. Farmers manage land between points A and B for permanent and sustainable agricultural activities. They use land between points B and C (where agricultural land value declines to zero) as an open access resource to be exploited for short-term advantage. Local households and communities may protect some lands beyond point B to a limited degree – such as having children shepherd their grazing livestock. Nevertheless, transactions costs C_r continue to rise after point B until eventually no reasonable number of shepherds or other resource guards can fully exclude illegal loggers and other open access users of remote forests.

Local consumers harvest the products that grow naturally in this region between B and C, crops like fodder for their grazing animals, native fruits and nuts and fugitive resources like wildlife. They will not invest even in modest land improvements in the region between B and C because the costs of protecting their investments would be greater than the return on these investments. Their use of this open access region is unsustainable except for periodic removals from pulses of naturally growing vegetation.

At the time of initial settlement, the mature natural forest at the frontier of agricultural development at point B has a negative value because it gets in the way of agricultural production and its removal is costly. The first settlers remove trees whenever the agricultural value of converted forest plus the value of the trees in consumption exceeds the cost of their removal. In fact, farmers in some frontier settlements farm in and around the trees they have not yet removed in preference to absorbing the costs of removal. Therefore, the function V_f describing forest value must begin below the agricultural value gradient V_a and, in this initial stage of development, it does not extend as far as the intersection of the agricultural value gradient with the horizontal axis.

A developing frontier, stage II: Both market and subsistence household demands justify the removal of some forest products and they will continue to justify additional removal at each new moment in time. Therefore, the forest frontier must gradually shift outward. The most accessible forest resources are always removed first. This is true whether those resources are timber, fuelwood, bamboo, fruits, nuts, latex, or whatever. The forest value gradient continues to shift upward and outward over time until it intersects the horizontal axis at some point like D in Figure 3.2. The price of the forest product in the market at A is now just equal to the sum of its cost of removal plus its delivery to the market. Products will be removed until their in situ value at point D is zero. Therefore, the value of forest land at D is also zero. The area of unsustainable open access activities now extends from point B out to points C or D, whichever is farther. The costs of obtaining and protecting property rights ensure that the resources within the area described by BD will remain open access resources.

Forest degradation, a feature of the open access region: The area of open access will not be fully deforested. Rather, the forest on these lands will be degraded to the level where the expected return from lower quality products is equal to the opportunity cost of the labour and capital used in their extraction. The picture below Figure 3.2 illustrates the effect on the forest. The remaining degraded resource tends to be larger and the forest is more fully stocked (degradation is less) at point D than at point B because the opportunity costs of labour and capital used in resource extraction are higher at the greater distance represented by point D.

These opportunity costs are important explanatory factors for the levels of deforestation and forest degradation. We can anticipate that both the extent of deforestation and the severity of forest degradation are greater in markets where the opportunity costs of extraction are smaller. Consider two regions that are similar in agricultural and forest values, in applications of property rights and in forest characteristics. They differ only in the opportunity costs of extraction from the forest. Loggers and harvesters of other forest resources extract from the forest out to the point where their compensation equals their opportunity cost. The lower their opportunity costs, the farther they are willing to go to extract products of the same value. Therefore, the forest value gradient in the higher wage and capital cost market will intersect the horizontal axis closer to the market at point A, while the forest value gradient in the lower wage and capital cost market will intersect farther to the right. (Point D will be farther to the right.) Deforestation will be more extensive in the lower opportunity cost market and the alternative labour and capital opportunities are critical determinants of deforestation.

Forest degradation will also be more severe in the lower cost market because smaller and lower quality products still offer sufficient rewards to offset the opportunities foregone while extracting harvests from the open access region.

Illegal logging, a second feature of the open access region: Some governments protect some lands past point B but they must absorb the increasing protection costs - and even then trespass and de facto open access activity occurs. For example, forested parklands around the world suffer encroachment and timber reserves around the world experience illegal logging. An almost unlimited number of well-trained and well-motivated forest guards could not prevent these activities entirely. For example, illegal logging is an issue of serious policy concern in developing countries like Bolivia and Indonesia, where it may be the source of 80 and 51 percent, respectively, of all wood harvests, but it occurs in the US as well (Smith, 2002). For example, some local citizens illegally harvest Christmas trees from the well-managed national forests in the eastern US. The US Forest Service does not expend greater effort to prohibit this theft because the costs of enforcement would be greater than the potential gain - which is another way of saying that the function C_r is above the function V_f.

A mature frontier, stage III: Eventually the margin at D extends far enough - and delivered costs and local prices become great enough - to induce substitution. This occurs when the costs of removal at a point like D plus the costs of delivery to the market equal the backstop cost of some substitute. Substitution may take the form of new consumption alternatives to forest products (for example, kerosene or improved stoves as substitutes for fuelwood or concrete block as a substitute for construction timber), or it may be production-related (as in planting and sustainable forest management on some land closer to the market). Very clearly, the evidence of planting and sustainable forest management is not ubiquitous - but the physical presence of sustainable management is not trivial either. FAO estimates that forest plantations cover more than 187 million hectares worldwide - more than 4 percent of the world’s total forest area and more than 1.4 percent of the total area in all land uses. The land area in plantations is growing by almost 4.5 million hectares annually and plantations now supply approximately 22 percent of the world’s industrial roundwood (FAO, 2001). For specifically tropical and sub-tropical plantations, Sedjo (1994) observed that the share of industrial wood production doubled from 1977 to 1992. Undoubtedly, that share has increased since 1992. The forest value gradient rises with the increase in delivered costs (from the dashed line to the new solid forest rent gradient in Figure 3.3) until, at some moment in time, it intersects the agriculture rent gradient to the left of agriculture’s intersection with the cost function for secure property rights.

We might call Figure 3.3 a description of the “mature” frontier of primary forest. For communities at mature frontiers, forest product prices will be sufficient to justify the substitution of managed forests for the resources of the open access natural forest. The new sustainably managed forests occur in the area described by B'B” of Figure 3.3. They may take the form of industrial timber plantations or of agroforestry or of just a few trees planted around individual households or in home gardens. The latter are excluded from most measures of the forest stock but their economic importance can be large. In the semi-arid Denver metropolitan area in the US, for example, trees comprise only 7 percent of the land cover, but they act as a $44 million regional storm-water management system, remove $2.2 million pounds of air pollutants annually and save $50 per household in annual air conditioning costs (Stein, 2001). Throughout south and southeast Asia, trees that grow in plots that are too scattered and too sparse to meet the measurement standards for inclusion in official measures of forest stock account for 65 to 85 percent of all market timber and fuelwood production (FAO RWEDP, 2000). They are major sources of fuelwood in Malawi (Hyde and Seve, 1993) and timber production in Kenya (Scherr, 1995). On the densely populated island of Java, they may account for tree cover on 47 percent of the land area - in addition to the 24 percent that comprises the official estimate of forest land.¹ The 47 percent is not contained in the official estimate because it is composed of scattered trees and stands rather than continuous forest and because its value is for shade, fuel, fruit and estate crops like coconut and palm oil rather than commercial timber.

In all cases, removals from the mature natural stock are concentrated in the neighbourhood of point D. Mature natural stocks from more accessible lands were removed in earlier times because they were open access resources. In most cases, a mature natural forest of no market or subsistence value exists beyond D. Sometimes the total area of this land beyond point D is negligible (e.g. in Ireland or Cape Verde). Sometimes it extends well beyond the frontier of economic activity at D (Siberia, Alaska, northern Canada, much of the Amazon, a large component of the Rocky Mountains in the US and much of Kalimantan in Indonesia) until it becomes the largest share of reported physical stocks.

Implications for forest and environmental policy analysis: The three figures trace an inter-temporal progression but, at any moment in time, they also describe static snapshots of local conditions and all three local conditions may exist simultaneously in different regions of some countries. They can exist simultaneously because most basic commercial forest products are either bulky or perishable and do not transport well before they reach the location of their next level of processing. Their markets are geographically contained. Therefore, standing natural forest reserves remain in some regions of a country (e.g. Siberia, Alaska, northern Canada, the eastern Amazon, Kalimantan in Indonesia) while the forests in other regions of the same country are depleted and some landowners may have begun to plant trees on their own lands (e.g. the Caucasus, the southern US, southern Ontario in Canada, the developed part of Brazil's Paragominas, central Java in Indonesia).

In sum, this characterization identifies three stages of forest development (Box 3.1) and the three categories of forest that exist today: managed forests (including industrial forest plantations, more scattered household trees and agroforestry plantings) over the land area described by B'B”, depleted forests from point B (or B”) to point D and an unmarketable mature natural forest beyond point D. For most commercial uses of the forest, we could identify a fourth category: the general area of current harvests from mature natural forests in the neighbourhood of D. Commercial timber and fuelwood, the most common wood products of the forest, generally originate from the first (managed forests) and fourth categories. Most forest policy and management is concerned with effects on the first three forest categories. Environmental and aesthetic concerns feature either the last category (e.g. biodiversity and natural preserves) or select locations within any of the first three (parks, erosion control, sustainability).

The critical points are that:

• we must differentiate the forest into these three (or four) categories before we discuss specific policy or market impacts, and

• the appropriate measure of the forest depends on the policy objective.

The differentiation is crucial because the same policy can have opposite effects on different categories of forest. Consider a few examples: timber price incentives are an inducement to improve and expand the managed forests for regions in stage III, but they are also an incentive to expand the deforestation of natural forests. For regions in stages I and II, price incentives have only a negative impact on natural forests. There are no managed forests in these regions to take advantage of the price incentive.

Similarly, forest cost-sharing and technical assistance programmes are inducements to improve and expand the managed forests in regions in stage III, but they have no effect on regions in stages I or II which only contain natural forests. Agricultural incentives, another example, induce land conversion from the forest to agriculture for regions in new frontiers, stage I. They cause some agricultural conversion from previously degraded lands for regions in stage II - but whether these degraded lands are included in the official forest inventory depends on the local measurement standards. Agricultural incentives have no effect on the natural forest for mature regions in stage III, but they do cause agriculture to out-compete managed forests at their intensive margin (B’).

The importance of the particular measure of forest inventory can be illustrated in a comparison of policies intended a) to improve carbon sequestration or b) to protect endangered habitat. All trees sequester carbon. Therefore, an accurate measure of the effect of a policy designed for this objective must include those managed trees contained in orchards, backyards, city parks, windbreaks, along roadsides and in agroforestry cropping systems. These trees are excluded from most official national forest inventories. Overlooking them would grossly misestimate the total amount of carbon sequestered.

On the other hand, most forest-based endangered habitat is contained in the remaining natural forest past point D in our three figures. Measures of the forest that include all three categories of forest would not be good indicators of the remaining natural habitat, and policies designed to affect forests in general but which concentrated their effects on extending the area of managed forests and/or improving the condition of the depleted open access area have little effect on endangered habitat.

In conclusion, the three-stage classification of forest development and the three (or four) categories of forest that emerge from it create a means for tracing both policies and spillovers from commercial activities within and outside the forest onto the various components of the forest. We will rely on this classification in our assessment of the effects of market and policy spillovers and institutional modifications on the various uses of the forest in the next sections of the chapter.

TAXES, INCENTIVES AND REGULATIONS

The general economics literature separates government intervention in the market into “standards and charges”. Standards refer to physical limitations such as regulations that restrict the harvest and shipment of logs or environmental regulations that restrict timber harvest activity in riparian zones. Charges refer to economic instruments that permit managerial discretion in achieving publicly acceptable levels of resource allocation. The common charges are taxes or fees for licences or permits. Financial incentives and other government assistance that reduce the costs of production are a second group of economic instruments that are widely recognized in forestry. We will review the effects of forest taxes, incentives and standards or regulations in that order.

Taxes

Governments impose three general categories of tax on forests in one place or another: income taxes, property taxes or their substitutes and severance taxes. Some taxes are primary sources of public revenue. For these, the objective is to collect revenue without altering the use of inputs or the level of output.

Income taxes: Income taxes are taxes on individual income or corporate profits. They are a primary source of revenue for many central governments. When applied equitably across productive activities, they tax equal proportions of income or profits from all activities and the tax burden is the same on all activities. They are neutral with respect to the allocation of land and other inputs between competing productive activities.

Income taxes shift the forest value function V_f downward an equal proportional amount throughout its profitable range - from the solid to the broken lines in Figure 3.2. An equal proportional shift means a greater absolute shift at the left extreme of the function, but no change whatsoever after point D where the net return on land becomes zero. Income taxes have no effect on land use at the frontier. They also shift the agricultural value function V_a downward by the same proportional distance. Therefore, these income taxes do not alter the critical land use margins at B’, B” and D (Figure 3.3).

However, the tax codes of some countries include provisions that do not apply equally to all productive activities. Some countries tax income derived from capital gains at a lower rate than other income. Capital gains are the appreciation in asset value that occurs during the period an individual holds the asset. Most timber is held for long periods and it does appreciate over time. Therefore, a lower tax rate for capital gains favours investments in timber (and some other assets like real estate, minerals and oil) in preference to activities like agriculture, retail sales and manufacturing whose production periods are shorter. The effect is not trivial. One estimate for the US suggests that the favourable treatment of capital gains provided 20 percent of the forest industry’s after-tax profits in 1984 (Russakoff, 1985).

The favourable treatment of capital gains, with its positive impact on profitability, is an incentive to shift investments in land, labour and capital away from non-favoured activities like agriculture and into favoured activities like forestry. This shifts the competitive agriculture-forestry land use margin (point B’) to the left and thereby increases the total land area in managed forests and also increases total production from the managed forests. Favourable treatment of capital gains has no effect on the extensive margin at B” because income taxes are proportional to profits. They do not directly alter the use of the land beyond B” that is unprofitable for forest management.

Property taxes: Property taxes are taxes on the value of real and personal property - land, its capital improvements, equipment and intangible assets such as stocks and bonds. Property taxes are a primary source of revenue for many local governments. Like income taxes, they are collected annually (or sometimes biennially).

Property taxes are applied to the assessed value of each property. This is generally some proportion of the property’s market value in its “highest and best” use. If all lands within a locality were assessed at the same proportion of their true market values and taxed at the same rate, then the property tax on land would shift the agriculture and forest land value functions downward by the same proportional amount throughout their profitable ranges. A diagram of the effect of the tax would be identical to Figure 3.1 for income taxes. However, standing timber and various other assets, as well as land, are also subject to the property tax and their taxation introduces biases into resource allocation. Since timber stands grow in place for long periods of time before they are ready to harvest, each annual collection of the property tax repeats the taxation of unharvested growth from all prior years. The final accumulation of annual property taxes paid on the timber by the time of harvest is much greater than the accumulation of taxes would be if timber production were an annual activity and each period’s growth were only taxed once. This effect of the property tax is known as the “time bias” against forest management. It encourages landowners to harvest their timber at an earlier age - in order to avoid some of the repeated and accumulating taxation.

The accumulated property tax on timber makes some land at the extensive margin (B”) unprofitable for forestry. The combined effects of the time bias on managed forests and the conversion of the extensive margin into unprofitable forest land were a major reason that many firms in the southern US forest industry “cut and got out” in the 1920s and moved to the American west. The southern industry harvested and then abandoned tens of millions of hectares of cutover lands rather than reforest and pay the property taxes. The forests in the Lake States of the US shared this experience to such an extent that, as the abandoned cutover lands reverted to state ownership between 1910 and 1940, they took on their own name, “the new public domain”.

The effect of the property tax on timber management has caused some policy-makers to re-examine local tax systems. In some cases, they have lowered the appraisal of forest lands. In others, they have introduced taxes that are paid only once, at the time of harvest (yield taxes). In others still, they have begun assessing forests according to their current use rather than their highest and best use. This reduces the tax bill on those forests that would have greater value in more developed uses and preserves managed forestry’s competitive advantage at the intensive margin. The preservation effect is temporary, however, if the value in alternative developed land uses continues to rise. The preservation effect lasts only until the developed land value exceeds the land’s current value in forestry plus the difference in tax bills between the two land uses. The total impact of these “green space” taxes on forest land has probably been small (Boyd and Hyde, 1989).

Severance taxes or royalties: Severance fees or royalties are charges on mature timber at the time of harvest. They are comparable to the stumpage fee paid by loggers to owners of managed stands, except that severance fees usually refer to natural timber in the neighbourhood of the frontier at point D in our three figures. Governments tend to be the nominal owners of timber at the frontier and beyond. Therefore, severance fees provide revenue for government treasuries or for the operation of forest ministries.

Severance fees act as disincentives for legal harvests. Therefore, raising them shifts the frontier at D inward, decreasing deforestation and saving natural forest environment. Whether they increase or decrease government revenue collection depends on the elasticity - just as revenue collection from any commercial activity depends on the price elasticity of the commercial good. Raising severance fees also increases the incentive to avoid the fee and log illegally. This is a critical incentive for forestry. Monitoring and enforcement are difficult in the best circumstances simply because the resource is so dispersed across the rural countryside. Therefore, while higher severance fees are a disincentive for legal harvests, they are also an incentive for illegal harvests and the net effect on the natural forest frontier will depend entirely on local conditions.

Severance fees are an important issue of debate for developing countries and many external advisers argue that governments could increase their revenue and improve the environment by increasing the fee. The government of Indonesia, for example, increased its severance fee in 1999 in response to pressure from external advisers. Government revenue collection has decreased and most argue that illegal logging has increased since then. However, many other factors have changed as well since the east Asian financial crisis of 1997 and the change in Indonesia’s government. We still await an empirical examination of the full relationships between severance fees, illegal logging, government revenue recovery, harvest levels and the forest environment - for Indonesia or anywhere else.

Incentives

Forest incentives come in the form of direct financial aid for forest management, free seedlings and advice and technical assistance for forest managers. The primary objective of all three is increased wood supply and all three generally target supply from small landowners. Financial incentives are more common in developed countries. They can afford them. The US, some Canadian provinces, the UK and the Nordic countries all provide financial assistance for small landowners. Even Chile’s economy, well-known for its free market, has been modified with a financial assistance programme for forestry, and the success of Chile’s forest sector has led policy-makers in some other countries to recommend these subsidies as ingredients in the development of successful forestry sectors. Finally, free or discounted seedlings and technical assistance are common to forest policy and rural development programmes around the world, regardless of a country’s development status.

Direct financial incentives, seedlings and technical assistance offer forms of cost savings in forest management. Therefore, their impacts are largely restricted to regions in the third stage of forest development and to the managed forests where cost savings can have a direct effect. Policy-makers and programme managers often overlook the fundamental point that incentive programmes in regions characterized by the first two stages of forest development will not be successful and they waste public resources when they attempt to offer forest management incentives in these regions. Landowners in these regions will not be receptive to the incentives because forest resources from the frontier are still less expensive than managed forest resources. The economically productive lands in these regions yield greater returns in other non-forest activities.

Financial assistance: Direct financial assistance or forest incentive payments (FIP) are monetary inducements to participate in forest management. The US, for example, has administered various financial assistance programmes since the 1920s. The current programme began in 1974. It provides government cost sharing for up to 75 percent of all reforestation and management expenses for landowners with fewer than 210 hectares (Boyd and Hyde, 1989). Chile, as another example, began its FIP programme in 1974 with the objective of improving the international competitiveness of its forest sector. Chile’s programme returns 75 percent of costs to the landowner one year after successful afforestation. Reforestation without compensation is compulsory for successive timber rotations. In 1992 the programme was restricted to landowners with fewer than 500 hectares.

The impacts of FIP programmes on land use can be traced through Figure 3.3. Financial assistance decreases the private management costs and, therefore, allows the forest value function V_f to shift upward by the per hectare value of the assistance. Land at both the intensive and extensive margins (B’ and B”) shifts away from competitive uses and into forest management and total production from the managed forest also increases.

The effect of the US FIP programme is likely to be small because many landowners fail to take advantage of government assistance and because many other factors are important for those landowners who do obtain government assistance. Some accept the assistance without ever intending to harvest their forests (Boyd and Hyde, 1989). The effects of Chile’s programme may be greater. The non-industrial private landowners who receive assistance account for nearly 40 percent of the timber supplied to Chile’s large forest and wood products sector, its fifth largest source of GDP and third largest source of export earnings. Many of them would not have afforested their lands without the initial promise of government assistance. Therefore, government financial assistance may be a source of significant increases in forest land use and timber production in Chile.

Distributive arguments are sometimes used to justify FIP programmes. These should be accepted with caution. In developed countries, even small forest landowners are not among the poorest or most disadvantaged for whom we design our redistributive policies. Moreover, in the US, Boyd and Hyde (1989) observed that those who do take advantage of the FIP programme are not among either the poorer or the smaller private landowners. Furthermore, it is an open question whether FIP programmes are of greater benefit to the landowners or to the wood processing industry that obtains its timber for lower prices because of government assistance for forest management.

Free seedlings: Various public programmes around the world donate seedlings to smallholders and communities, or build nurseries that provide seedlings at discounted prices. These have the same general effect as FIP programmes. They decrease the costs of forest management.

Like FIP programmes, the effects of free or discounted seedlings are restricted to managed forests within regions in the third stage of forest development. Also like FIP programmes, the effects of seedling distribution programmes can be traced from private financial savings for the seedlings through its increasing effect on the net forest value function V_f. The final effect on land use and forest production, even in the third stage of forest development, is probably small in many cases, although there are numerous examples from countries of various levels of economic development, where farmers willingly accepted seedlings and tended them to maturity.

Technical assistance: Technical assistance programmes are also known around the world, often in the form of forestry extension programmes. These programmes are designed to bring the latest information on modern technologies to local forestry operations. Like FIP programmes, forestry extension generally targets smaller operators. Occasionally, it includes advice on improved logging techniques. In this case, it has an impact on timber harvests in all three stages of forest development - leading to additional recovery of usable material from the residual open access forests and potentially extending harvests farther into the remote frontier.

More generally forestry extension, like agricultural extension, is designed to assist landowners improve their land management. Therefore, like FIP programmes and free or discounted seedlings, the effects of most technical assistance programmes in forestry can be traced through their impacts on the forest value function and on managed forests in the third stage of forest development. The greater questions for many forestry extension programmes have to do with obtaining rapid and widespread adoption of the preferred new technologies. In fact, these are important questions for FIP and seedling programmes as well. The experience of agricultural extension is relevant. In agriculture, better informed landowners and landowners who can afford the uncertainty inherent in trying a new variety of seed or a new production technology are the obvious targets for initial assistance. Other landowners will follow rapidly as they observe the successes of their neighbours (Feder et al., 1985).² Of course, the greater the management cost saving due to the new technology (the larger the positive impact on the forest value function V_f), the more rapid and widespread the rate and final level of adoption.

Regulations

Forestry has a long history of government regulation. The initial regulations were designed to protect the forest as a hunting ground for kings and large estate owners. Additional regulations were designed to protect the best trees for the kings’ navies. The emphasis on the rights of kings and large landlords changed in the latter part of the nineteenth century. First a sustained flow of timber and then, later in the twentieth century, a sustained flow of other environmental services from the forest became new justifications for government regulations. Table 3.1 traces the general pattern of forest regulations through the twentieth century. It lists the most common regulations in forestry in the order in which they have been introduced in most countries and it also identifies the general reason given for each regulation. In each case the fundamental reason is the protection of some broader social value.

Some North American and northern European countries began requiring reforestation following timber harvests as early as the 1930s. Today, some developing countries require a permit before allowing harvests even on private lands. Others restrict the shipment of timber across provincial boundaries. The objective has always been to control unlimited harvesting and to ensure a resource supply to local mills.

As the laws and policies of the developed countries began to focus on the environment, these countries introduced new restrictions on harvest systems. In North America, restrictions on clearcuts or regulations ensuring selective forest management became widespread. These were followed by restrictions on specialized forest practices as the public became aware of the detrimental effects of those practices. Herbicide and pesticide usage on forests were controlled, for example, once the human health effects of substances like DDT were recognized.

Water quality restrictions are another example. As water quality became a public concern, the laws and policies protecting it first focused on point sources like effluent from pulp and paper mills. By the mid-1980s the emphasis shifted to non-point sources like agriculture and forestry.

Table 3.1: Common regulations of private forest lands

Pulpmills are point sources because the exact point of their detrimental discharge, the pipe from which it spills, can be identified. Agriculture and forestry are non-point sources because their contributions to water pollution cannot be traced to a single discharge point. Rather, their water quality effects originate from activities like herbicide usage or general soil disturbance that are dispersed over large areas.

Finally, the most recent regulatory issues for private forestry refer to endangered species and carbon sequestration to protect against global climate change. All regulations are intended to encourage preferred behaviour or discourage the opposite (Box 3.2). As examples, the impacts of four categories of regulation will be examined: i) restrictions on harvests and shipment, ii) reforestation requirements, iii) environmental requirements such as restrictions on herbicide use, streamside management and clearcuts and iv) certification.

Restrictions on harvests and shipment: Restrictions on harvests are intended to ensure a long-term supply of the resource or, in some cases, to permanently protect the natural forest. Restrictions on shipments are intended to ensure a supply for local mills and thereby protect the welfare of local communities. Restrictions in many countries on shipments confine timber within state or provincial or some other local boundaries. Some countries have log export restrictions - which have the same effect, except that they restrict shipments from leaving the country rather than the internal region of a country.

Harvest restrictions occur in two basic forms. Some restrictions, known as logging bans, are intended to be absolute, halting all harvests on certain varieties of forest, such as remaining natural forests. Other restrictions are intended to control certain varieties of harvest or to limit harvests to approved levels. Logging bans are difficult to enforce because governments cannot hire enough guards to protect the entire forest and because it is difficult to prove that delivered timber came from the protected source - rather than from some farm woodlot. Logging bans quickly raise the spectre of expanding illegal harvests. More limited restrictions such as quotas for timber harvests create uncertainty in the minds of landowners regarding their right to make future harvests. Uncertainty encourages landowners to harvest earlier than they would without the restriction - while they can still be certain of an economic return. There may be a short-term increase in the harvest level as landowners hurry to obtain some certain return. However, the longer-term effect is negative as landowners transfer some forest lands into alternative activities where the economic returns are more certain and timber supply decreases as a result.

The costs of timber harvest permits, when available, also reduce returns to landowners. Sometimes these costs are small and their impacts may be inconsequential. In other cases, obtaining the permit consumes more time and financial resources and, even then, their availability may be uncertain. In these cases, the high costs make lands at the intensive and extensive margins of managed forestry (B’ and B” in Figure 3.3) less competitive as managed forests and they cause landowners to manage their remaining forests and trees less intensively. The standing timber resource and the long-term supply decline as a result - in direct contrast with the policy objective to ensure long-term supply.

India and Sri Lanka provide examples. India restricted all harvests of its high-valued but declining sandalwood in an effort to preserve the remaining resource. Some landowners responded by harvesting immediately, before the restriction became effective. Others have harvested illegally since then and the standing sandalwood inventory has actually declined since the restriction became official policy. Sri Lanka restricts all harvests from its remaining natural forest and requires permits and inspections before allowing harvests on private lands as one of the means of ensuring that harvested logs do not come from the protected natural forest. Permits have become such a serious constraint on supply that delivered log prices are as much as seven times the price received by producing landowners. The various costs associated with obtaining the permit absorb the difference. Therefore, the higher price has not become an incentive for additional management by private landowners and the managed forest inventory has declined. However, the higher price has been an incentive for illegal logging of the natural forest - also in direct contrast with the original policy objective. Timber production is now about one-fourth the level it was prior to the permit system and perhaps as much as half of all wood delivered to Sri Lanka’s mills is the result of illegal logging (Gunatilake, 2002).

Restrictions on shipment, including export restrictions, have a similar set of final effects on the forest. Their first effect is to reduce the number of loggers and mills eligible to compete for a resource. The US, Philippines, Nepal and China, for example, each restrict at least a share of some timber sales for local mills. These countries may succeed in protecting some less-competitive local mills in the short run. However, it is not clear that local community welfare is significantly enhanced because wood processing is generally a small share of local economic activity and loggers and mill workers have other employment opportunities. The long-term effect of shipment restrictions is identical to that of harvest restrictions. Landowners obtain lower prices for their timber and their incentive to manage and produce a long-term timber supply for local mills declines. Both margins of managed forest land shrink and this is certainly not in the interest of community welfare.

Reforestation requirements: Required reforestation following a timber harvest was originally introduced as another means of ensuring long-term timber supply. More recently, the objective has shifted to general environmental sustainability. Between 1903 and 1950, 16 US states introduced laws requiring reforestation to satisfy the first objective. Nine states revised or introduced new laws after 1968 to address the broader second objective. All the Nordic countries, several other western European countries, as well as Brazil, Chile and Ghana, for example, have similar laws with similar objectives. Forests regenerate rapidly and naturally on some lands. On these, the reforestation requirement imposes no real cost. On others, it adds the cost of reforestation to the timber harvest costs. This shifts the forest value function V_f downward. Where the requirement is enforced, harvests at the frontier decline and most would consider this an environmental improvement.

Required reforestation has no impact on the managed forests of the third stage of forest development because reforestation is already economically viable for these forests. Boyd and Hyde’s (1989) empirical examination of the reforestation component of Virginia’s state forest practice act supports this contention. Boyd and Hyde examined private forest management in the states of Virginia and North Carolina. These states are geographically adjacent and similar in their forest lands. Virginia has a reforestation requirement, while North Carolina does not. If the requirement were effective, then a greater level of reforestation should accompany it and, over time, standing forest inventories should also be greater. However, after checking for differences in site quality and regional prices, Boyd and Hyde observed no measurable differences in standing forest volumes between the two states.

Additional silvicultural prescriptions: A number of additional silvicultural and other environmental prescriptions have been imposed on timber harvesting and continuing forest management activities in several developed countries and a few tropical developing countries. These include limitations on clearcutting, harvest spacing restrictions designed for the regeneration of mixed forests, restrictions on the use of herbicides and pesticides and standards for streamside management and logging roads.

The added costs implicit in these environmental regulations shift the forest value function downward. Added management costs, such as those incurred for restrictions on herbicide use, pivot the function downward and affect only the range of managed forestry. Added harvest costs, such as those incurred for restrictions on clearcuts, shift the function downward over its entire range, reducing the land area in managed forests at both the intensive and extensive margins and also decreasing management intensity on the remaining managed forests. Harvests at the frontier decline as well.

The net effect is substantial. Sedjo (1999) estimated that these regulations add an average of 5 to 18 percent to the costs of forestry in the southern US, British Columbia and Finland - although the impacts on individual landowners vary with local land quality and enforcement conditions. If these costs are roughly comparable to a similar decrease in the stumpage price received by the landowner, and the supply price elasticity is greater than one (the common expectation), then they suggest an even greater percentage decrease in timber supply from the regions affected by these regulations.

Clearly, these environmental regulations alter forest production in important ways, decreasing production while improving the forest environment. However, the magnitudes of the costs and their effect on production raise two new issues, one for private land managers and the other for public environmental values. Private landowners challenge the “taking” of their land use rights that occurs when the government imposes new regulations that restrict them from making some decisions and benefiting from opportunities that were available before the new regulations were imposed. This concern is probably greatest in the southern US where non-industrial private landowners manage two-thirds of the forest land and account for 60 percent of annual timber harvests and where reforestation, water quality and endangered species restrictions on private lands have become more confining in recent years. Taking is also an issue of contention, however, in parts of Canada and in the Nordic countries. In fact, it is an issue in any place where expanding public environmental awareness threatens to affect private forest landowners - but promises them no restitution. As with uncertainty about harvest restrictions, uncertainty about increasing environmental regulation has caused some landowners to act pre-emptively. They have harvested early, before the policy became official, thereby ensuring to themselves some level of return on their timber investment, but foregoing any opportunity for greater return on the growing timber and also permanently damaging any environment that would have been protected by the new regulation.

The new issue for public environmental values has to do with shifts in the geography of production and environmental losses that must occur when the regulations are imposed in selective regions only. Consumer demand is unaffected by these regulations. Therefore, the significant decreases in production that occur in the southern US, British Columbia and Finland, for example, will be largely compensated by increases in production from other parts of the US, inland Canada and Russian Korelia, respectively, as well as with additional imports from developing countries. In each of these cases, the production shifts are largely from managed forests of regions in the third stage of forest development to the natural forest frontiers of regions in the second stage of forest development. Environmental standards are generally lower or non-existent in the regions of alternative production. Therefore, while the higher environmental standards improve environmental quality in the southern US, British Columbia and Finland, the same regulations effectively export environmental damage and deforestation to other parts of the world.

This effect is generally overlooked by those interested in improving the forest environment. The same interests that demand environmental improvement at home and desire to reduce deforestation globally are forcing a share of all production to shift away from sustainably managed operations at home to unsustainable harvesting at the frontier of other regions. While the trade-off between environmental improvement in one region and environmental decline elsewhere is unknown, the magnitude of additional silvicultural costs in the southern US, British Columbia and Finland suggests that trade-off must be substantial. Shifting more than 5 to 18 percent of the production in only these three regions to other, more environmentally risky regions could imply serious global environmental losses.

Certification: Certification refers to the assurance from an independent authority that forest products originate from sustainably managed forests. Although certification is not the official policy of any government at this time, it is widely encouraged, especially among western Europe’s environmentally conscious consumers, and it could become policy in some countries in the future (Hansen and Juslin, 1999). Until it becomes a formal government policy it can be anticipated that landowners will seek certification when they perceive that it brings them advantage in the markets for their products, either a price advantage, improved market share or an opportunity for penetration into new markets. This means that the anticipated benefits of certification, the higher prices or increased share of an environmentally aware market, must offset the costs of obtaining the certification.

We can anticipate that only the owners of managed forests from regions in the third stage of forest development should be able to provide legitimate evidence and make successful requests for certification. Northern Europe is largely within the third stage of forest development and many firms in its industries rely on wood supplies from managed forests. Certification does not impose a significant cost on them. In fact, more than 25 million hectares, or half the global total, have been certified in the Nordic countries (Bass et al., 2001).³ Consumers express greater concern for the certification of forest production from tropical and developing countries. A much greater share of the forest products from these countries tends to come from natural forests in regions within the first two stages of forest development. Indonesia’s example illustrates some of the problems for certification in these latter cases. Indonesia is the largest exporter of forest products in Asia and forest products are the third largest source of Indonesia’s export earnings. Indonesia’s large producers are rushing to obtain certification and thereby gain a share of the European market. They could become major exporters to Europe. The small and new Indonesian certification institute cannot keep up with the demands for its services.

Nevertheless, it is difficult to imagine successful certification for a significant share of Indonesia’s forest products. Less than 10 percent of Indonesia’s annual harvests come from managed forest plantations and most of its timber markets are characterized by the first two stages of forest development where sustainable management is not yet financially viable. Certification faces an additional problem where the unsustainably managed share of all forest products taken from the unmanaged natural frontier is so large. The products of managed and unmanaged forests and unsustainably managed forests are similar and one easily substitutes for the other at the mill. Therefore, we can anticipate that either the monitoring costs associated with certification will be great or certification itself will not be a reliable guarantee of production from genuinely managed forests.

In sum, the concept of certification is young and the procedures for its administration are still developing. It will face its most serious challenges in producing regions that are largely characterized by the first two stages of forest development and reliance on the natural forest, yet regions that are also significant suppliers for consumers who wish to have the assurance of certification. Indeed, the measure of certification’s success should be whether certification ever becomes the reason why any logging company operating at the forest frontier converts to a programme of long-term forest management.

ADJACENT SECTORAL POLICY AND INSTITUTIONAL SPILLOVER EFFECTS

As policy changes induce expansion or contraction in the sectors that compete with forestry for inputs, these policies also affect forestry. Similarly, as policy changes induce expansion or contraction in sectors that consume forest products, they too affect forestry. The important competition for inputs is the competition for land, most generally between agriculture and forestry. A variety of industries consume the outputs of the forest in their own production processes: rubber; producers of exotic fruits, nuts and natural herbs; outdoor recreations, etc. Of these, the wood products industries have the greatest effect on most forests. The effects of policy change in these two sectors, agriculture and wood products, will be the focus of our attention.

Extensions of the local infrastructure and modifications in the local and national institutions are both functions of public policy and they too can be important determinants of forest development. Therefore, changes in regional infrastructure and institutions will be added to the discussion. While neither infrastructure nor institutions are regularly considered as components of policy, policy can alter both and both are important determinants of forest development.

Agricultural policy spillovers: Agricultural policies affect agricultural land use, including those agricultural lands that are newly converted from forests. Therefore, agricultural policies may affect forests as well. However, the discussion of the pattern of forest development showed that the relationship between agricultural and forest land use is not consistent across the three stages of forest development. Furthermore, not all agricultural improvements use more land. The impacts of agricultural policy on the forest are complex for these reasons.

Many countries have had policies that encourage agricultural growth. For example, the US and Canada encouraged agricultural settlement on their frontiers in the nineteenth century, in the 1980s Indonesia financed the movement of new settlers to the forest frontier in a policy called “transmigration”, and in the early 1990s Finland compensated farmers for clearing new land for agricultural use. For a period in the 1990s, Brazil gave title to Amazon forest land to anyone who cleared it for agricultural or livestock use. In each case, agricultural (or population) policy shifted the agricultural value function V_a upward, at least at its far extremity, and encouraged agricultural expansion within a region in the first stage of forest development.

As we progress into the second and third stages of forest development, agricultural policies no longer have an effect on the forest frontier. By this time, the more common agricultural policies are subsidies for agricultural inputs and price supports for agricultural outputs. Input subsidies tend to encourage the use of fertilizers and water, in particular. They decrease capital costs and increase the marginal product of a dollar’s worth of capital input. They cause producers to shift input proportions, increasing the relative use of the subsidized capital and decreasing the relative use of labour and land inputs. In terms of our figures, these capital-consuming and land-saving technologies shift the agricultural value function upward along its vertical axis and increase its slope. Depending on the magnitude of the subsidy, these programmes can either expand or contract agriculture’s use of the degraded open access lands (between points B and D) in the second stage of forest development and can either increase or decrease agriculture’s competitive position at the intensive margin of forest management (B”) in the third stage.⁴

The effects of agricultural input subsidies are compounded by the effects of government agricultural research programmes. These are not among the programmes usually considered when we discuss government interventions in the market. However, within agriculture, they represent substantial government investments and some of these investments have produced phenomenal increases in agricultural productivity and decreases in agricultural costs (see Ruttan, 1982 for a survey). They tend to make capital inputs less expensive and more productive. Therefore, they also tend to be relatively capital-consuming and land-saving and their effects on the forest are comparable to those of agricultural input subsidies.

The effects of price supports on agricultural outputs contrast with the effects of input subsidies and government research. Agricultural price supports shift the agricultural value function upward throughout its range and introduce agricultural management on some of the degraded open access lands for regions in the second stage of forest development. For regions in the third stage of forest development they improve agriculture’s ability to compete with the intensive margin of managed forestry, thereby converting some land away from forest management. The net effects on forests for regions in the second and third stages of forest development are further complicated as some developed countries have introduced incentives to remove land from agricultural production. Some of this land, no doubt, reverts to forest. In fact, the US Agricultural Conservation Reserve programme specifically subsidizes the reversion to forest land. Ireland has a similar programme. Therefore, in some countries programmes are observed that support the conversion of forest to cropland in some parts of the country while other programmes support the opposite, cropland reversion to forest, in other parts of the country.

The mixture of effects of agriculture policy is complicated yet further by the variability of programmes between agricultural products. For example, in the US wheat may be a greater beneficiary of output price supports while cotton and livestock may be greater beneficiaries of water and grazing subsidies. The total effect of agricultural programmes causes some products, like milk and cotton and peanuts in the US, to be produced in regions where they would never appear without government assistance. In all cases, the pattern of effects traces through the entire agriculture sector, with government assistance to higher-valued crops affecting the land use margins of those crops with lower-valued crops until eventually the shifts affecting some lower-valued agricultural product causes a shift in its competition for land use with either the degraded open access forest (stage II) or managed forestry (stage III).

The net global effect is only speculation, but many would speculate that the very large scale of the North American and European Union agricultural support programmes is the determining factor. The US and Canada spend roughly US$40 billion annually on agricultural programmes and the EU spends an even larger amount. These immense expenditures fund the excessive use of all inputs, including large shares of North American and European land that would otherwise be under tree cover. They create excessive agricultural production, some of which is exported to developing countries where its low prices drive out competition from local agriculture products. Some of the disemployed local commercial farmers and agricultural workers revert to subsistence farming, which uses land more extensively than commercial agriculture, and still more land is converted from forest to subsistence agricultural use because of the North American and EU agricultural policies.

Spillovers from policies designed for their impacts on the wood products industry: Policies directed towards wood products industries affect the demands of those industries for their raw materials and through these demands they also affect forestry itself. The common objectives of policies designed for these industries are the protection of the domestic industry from international competition and the promotion of its development, although controls on environmental pollution can also be important, especially for the pulp and paper industry. Log export bans are the common means of protecting the processing industries. Their effects are much like the effects of restrictions on timber shipments discussed previously. A policy to restrict exports removes international competition from the domestic market. With only domestic processors competing for timber, demand declines and domestic prices shift downward. This is a benefit for the domestic wood processing industry but it is a disincentive for forestry itself. The net forest value function V_f shifts downward throughout its range. Both the area in managed forestry and the level of managed forest production decline for regions in the third stage of forest development. Harvests at the natural forest frontier also decline - for all three stages of forest development.

Several Asian countries have log export bans and both the US and Canada have banned the export of certain classes of logs from the Pacific Northwest. Indonesia’s experience provides a good example. Indonesia’s log export ban was intended to promote forest-based industrialization and employment. It did help accomplish this objective, especially in the plywood industry. Indonesia was the world’s largest exporter of tropical hardwood logs in 1980, exporting 60 percent of its harvests. A gradual ban on log exports, introduced in the early 1980s, became total in 1985. The domestic wood processing industry expanded immediately. The export of both plywood and sawnwood increased from a very low level in the mid-1970s until exports were more than half of Indonesia’s total production of processed wood by the mid-1980s. Indonesia became the world’s largest exporter of hardwood and plywood and wood products became the third largest source of Indonesia’s export earnings prior to the East Asian financial crisis in 1997. However, growth in the plywood industry came at a cost. Domestic log prices fell below the world price through the 1980s and 1990s. Fitzgerald (1986) estimated that for every dollar gained in the value of plywood exports, four dollars were lost in foregone log exports. The employment gains were small, less than one person for every 50 hectares harvested (Gillis, 1988). The reduction in demand and lower prices coincided with a reduction in harvests by as much as 50 percent. Most of the reduction came at the natural frontier as most of Indonesia was in the second stage of forest development and only about 3 percent of Indonesia’s harvests at that time originated from plantations.

Government-sponsored research in wood processing, as in agriculture, has been another source of industrial development - and a source of reduction in the demands for wood as a raw material. US government-sponsored research for the southern pine plywood industry produced rates of return in the neighbourhood of 300 percent per annum throughout the 1960s and 1970s. One research improvement alone, the powered back-up roller, improved wood utilization and reduced the plywood industry’s demand for logs by 17 percent. Overall research-induced cost reductions in southern pine plywood were a critical component of the rapid expansion of the industry throughout the 1970s and 1980s. Expansion in this industry increased the demand for southern pine logs. However, southern pine plywood substitutes for western pine plywood and also for lumber in some of their common final uses as construction material. Southern pine plywood typically utilizes more of the log than either of the latter products. Furthermore, the southern pine region is in the third stage of forest development with a significant share of its timber originating from managed forests, while western US forests are more generally characterized by the second stage of forest development and a much larger share of western US harvests originate from mature natural forests. Therefore, it can be safely concluded that government-sponsored southern pine research decreased the total US demand for logs as well as total US harvests at the natural forest frontier.

Infrastructure: The term infrastructure includes public utilities and public services like education, hospitals and communication networks. Antle (1983) showed that the full collection of items identified with infrastructure has a significant and positive effect on rural economic development in general. Improved roads in particular have important effects on the general condition of natural forests. They have a lesser effect on managed forests because the latter occur in regions that are already developed. In the earliest stage of forest development, the community and the forest are in the closest proximity, or even intermingled. All roads affecting the community also affect agricultural development and the natural forest. New roads improve the conditions for local development. They shift the agriculture and forest value functions to the right, converting some degraded forest into permanent agricultural land but also degrading and deforesting additional lands - as in Figure 3.4.

In the second and third stages of development, roads that extend to the forest continue to have a direct effect on the forest. The improved access makes the region’s land more valuable in all uses. Once more, the agriculture and forest value functions shift right, moving the important land value margins (B and D in stage II; B’, B” and D in stage III) with them. This extends the claims of permanent agriculture and shifts the entire degraded open access area further into the geographic interior in stage II. In stage III, it extends agriculture into the area of previously managed forest and extends managed forestry into the area of previously degraded forest. The natural forest frontier shifts further into the interior as the deforested area expands in both the second and third stages of forest development. Thailand provides a very clear recent example. Thailand built roads into its more sparsely inhabited northeast in the 1960s. The policy objective was security - military access and encouragement for human settlement to secure the region against encroachment from Laos and Cambodia during the Vietnam War. Of course, timber harvests followed.

In fact, rights to adjacent lands and timber are often part of a government’s payment to private roadbuilding contractors. The US government transferred alternate square-mile patches of public domain for 10 miles either side of some railroads as payment for their construction in the nineteenth century. The government of Laos recently made a similar transfer of timber rights for building a highway through its northern forests to provide access to the rapidly developing markets of southern China.

Institutions: Like infrastructure, the institutional arrangements for local exchange, and particularly the arrangements for local property rights, have a forest development effect. In developed countries, we tend to think of property rights in terms of formal titles to land and other resources. However, Feder et al. (1988) used an example from Thailand to demonstrate that formal titles are insufficient protection unless the titles can be transferred easily as higher-valued uses of the resource present themselves. On the other hand, Migot-Adholla et al. (1991) used examples from sub-Saharan Africa to show that formal titles are unnecessary where customary rights are sufficiently enforced by the local community.

Figure 3.5 - Transferring property rights from a central agency to a local institution (the example is for a region in Stage III)

Both points are important in forestry today, as many countries are finding that formal rights to forests, which tend to be the official responsibility of the forest ministry, are not easily enforced. Yet some local communities protect their forests even when they have few formal rights to these resources. In fact, these considerations provide the justification for the transfers of public forest to some form of local private or local community-based forest management. The familiar terms; “community forestry”, “joint forest management”, “public participation’ and “devolution” all refer to transfers to local users of some or all of the rights to forests that were previously the unambiguous responsibility of central forest ministries. The economic argument for these transfers is that local users of the land and forest resources know those resources and the demands on them better that the officials of the forest ministry. Furthermore, the local users live closer to the forest and their daily activities put them in closer contact with it. Therefore, local users can manage the resources more efficiently than the ministry and they will have greater success than the ministry in enforcing their management intentions. In terms of our figures, the shift to local management will reduce the transactions cost function C_r, as in Figure 3.5, more land will be managed sustainably (points B or B” shift to the right) and the area of degraded open access forest will decline.

These arguments and the shift to local management work well where the forest values at stake are local, where they are shared by most of the community and where they are relatively large enough to justify local management of the resource. Therefore, local management can improve long-term land management for agriculture, timber and other extractive products of the forest and also for local non-market values like some erosion control and recreational forest use. The list of successful examples of community management is almost endless and it comes from all corners of the world. However, local property rights and local management do have their limits.

Transfers of rights to local communities are less successful where the local values for forest products are very low relative to the community’s values for other uses of their land and their own time. This is the general case in the first stage of forest development. The forest is plentiful and members of the local community have little interest in using their own scarce resources to protect it. It is also the case for the lands that remain in an open access condition under the best institutional arrangements (the lands between B or B” and D) in the second and third stages of development.

Local management will be less successful in several other cases within the second and third stages of development. It will be less successful where important local groups have competing demands on the forest. For example, community management has not been successful for the eroding hillsides around Basantapur in Nepal. The higher income agricultural households in the community prefer to exclude all activity from the forests on these hillsides in order to prevent erosion and downstream damage to their agricultural lands. The poorer landless households rely on these forests for fuel. The higher income households formed a forest protection committee and hired forest guards, but the guards have been unable to prevent fuelwood collection and further degradation of the forest by their trespassing poorer neighbours (Dangi and Hyde, 2000).

In addition, local management may be no more successful than management by the forest ministry where either the local property rights are incomplete or the policy environment is uncertain. Incomplete property rights are common in forest transfers as the ministries often desire to maintain some degree of oversight - but often succeed only in removing the incentives for local management. The Philippines provides an example of the former, incomplete rights. The Philippine Bureau of Forest Development (BFD) established a policy of land transfers to communities in the early 1990s supported by a US$40 million loan from the Asian Development Bank. However, various NGOs were concerned about the potential distribution of benefits from the transfers and BFD was concerned that the communities would fail to follow its own perception of good management practices. Therefore, BFD required each community to hire a forester, to report management plans to regional and central advisory committees for approval and to return 44 percent of gross revenues from the community forests to a central budget used to pay down the loan. These requirements were too severe and not one community in the entire country applied for the transfer of community forests until these requirements were withdrawn in 1997 (Hyde et al., 1997).

Yin and Newman (1998) illustrate the latter problem, an uncertain policy environment, by contrasting experience from two regions in China. Following the introduction of agricultural reforms in 1978, the Chinese authorities gradually extended to farmers the rights to manage and harvest trees on lands of China’s former farm and forest collectives. In the northern plains, the authorities paid little attention to forestry, the farmers initially harvested trees as they obtained the new rights, but they also planted and, over time, both the standing forest and the level of harvests increased sharply. In a second region just south of the first, the authorities gave rights to farmers and then subsequently took rights back from some farmers. In fact, they had changed farmers’ rights another three times in the 20 years preceding reforms. Farmers in this region responded to their uncertainty about the permanence of their rights by harvesting without replacement. Both the level of the standing forest and the level of successive years’ harvests declined despite the fact that the rights to the land and trees were similar for both regions by 1985.

Finally, local management is never a successful substitute for regional or national management, or even global assistance, when the values at stake are public values that are shared by the broader regional, national or global community. Carbon sequestration, biodiversity and some kinds of tourism are all in the global interest. In these cases, broader institutional management is necessary to ensure that broader interest, but local involvement is also necessary in order to ensure local cooperation - instead of local trespass in search of local extractive products. Numerous national parks in both developed and developing countries have successfully included local participation in the benefits of the park (e.g. providing various tourist services and also employment in park management itself). Even in these cases some amount of trespass is inevitable. It may be anticipated that creating a local interest in sequestering carbon or protecting biodiversity will be even more difficult (Box 3.3).

MODERN POLICY EXAMPLES

This section of the chapter discusses five key forest products and environmental services: i) timber and non-timber forest products, ii) carbon sequestration to protect against global climate change, iii) erosion control and general watershed protection, iv) biodiversity and critical habitat and v) tourism.

Timber and non-timber forest products: The focus of the policy discussion in forestry has shifted over the last 40 years. Timber supply to support a domestic industry and economic growth, and its European counterpart, the creation of a strategic timber reserve, were central to the discussion in Europe and North America for at least 100 years. The colonial powers in these continents exported their concerns to government forestry agencies and forestry schools in the developing world until timber supply became the dominant forest policy issue in almost all countries and the world now has long experience with policies addressing the “timber supply problem”.

The conclusion of this experience is that timber is a market valued product and the market provides for it, either from natural stands at the frontier or from managed stands, in just the manner described in the first section of this chapter. In fact, the market has responded to increasing demands for timber, but policies designed to improve timber supply have not had overwhelming success (see Section 3.3). Forest incentives and technical assistance have probably caused smaller expansions of timber supply when they were directed at regions in the third stage of forest development, regions that already contained some managed forests. These policy instruments have not been successful in increasing timber supply in regions where all timber supply still originates from the natural forest frontier because those frontier harvests are still less expensive than timber potentially coming from managed stands.

Numerous regulations have been tried as a means of ensuring improved timber supplies or restricting declining supplies. Reforestation requirements and harvest and shipment restrictions are the most common instruments. Invariably they have removed the incentives for future timber management and long-term timber supply declines as a result.

The experience with fuelwood has been similar, just as it has been for other marketed non-timber forest products as well. Production behaviour with respect to these forest products also follows the market pattern of our Figures 3.1-3.3 and these products are generally market-valued themselves - despite some assertions to the contrary. Local markets exist for most of these products in most regions where they are consumed. The markets may be thin and most households may collect for their own domestic consumption in preference to purchasing in the market. However, local markets generally do exist and whether local collectors choose to buy or sell instead of collecting for their own consumption is an economic decision. In fact, a few local residents do buy and a few do sell and the numbers who participate in the market change as the local prices for these products change.

Furthermore, household decisions regarding these non-timber forest products (NTFP) are market-related even for households that neither buy nor sell them. These households use their own labour to collect these resources and that labour is exchangeable for market-valued goods - sometimes as hired labour that earns a wage which can be used to purchase goods, sometimes as labour used in the domestic production of marketable agricultural products. Therefore, all households have incentives to apply their labour for collecting NTFP until the marginal value products of that labour approximates the marginal value product of their labour in those other, marketable activities. Therefore, it is not surprising that the instruments for improving the supply of NTFP, usually free or discounted seedlings or government-sponsored plantations, have a poor history of success. Where these products are valuable enough for local farmers and communities to plant and manage, the evidence is that farmers and communities do just that. In fact, two recent examinations of the global fuelwood situation have arrived at exactly this conclusion (Arnold et al., 2002; Hyde and Kohlin, 2000).

Carbon sequestration and the mitigation of global climate change: The discussion on forestry and global climate change has two components: a) the effect of climate change on forests - which has to do with changes in the extent of forest cover in response to anticipated global climate change (Sedjo and Solomon, 1990) - and b) the mitigating effects of the forest on climate change. The latter is probably of greater interest. It can help preclude the former.

Increases in atmospheric levels of CO₂ cause global warming. On this most are in agreement. The level and importance of global warming, to date or in the foreseeable future, remain debatable issues (Mendelsohn and Neumann, 1998; Nordhaus and Boyer, 1999). The importance of the forest in mitigating global warming also remains a debatable issue.⁵ However, it is clear that fossil fuel combustion and land use change (mostly agricultural conversion of the forest) release CO₂ into the atmosphere. The acts of harvesting timber and processing wood products also release CO₂, although not nearly at the rates of fossil fuel combustion and land use change. Indeed, wood processing releases only a portion of the carbon originally contained in a tree because so much is stored in the soil (roots, soil carbon, detritus on the forest floor), in the product (lumber) or in landfills (paper).

Reducing the rate of forest conversion to agriculture, particularly where conversion involves burning the forest cover, would decrease the rate of atmospheric build-up of CO₂. Growing trees in new plantations, harvesting and storing the final biomass would retain additional CO₂. However, growing trees and storing them on the stump would only yield a short-term increment to stored carbon because little additional storage occurs once the trees achieve maturity and the rate of tree growth declines. Nevertheless, additional tree planting and tree growth could be a means of buying time until world markets substitute other energy sources for fossil fuels.

The problem confronting any attempt to use trees to control global climate change is the mismatch between the non-exclusive global public impact and the more identifiable but still general location of the forest activities to control it. Protection from climate change is a public good in the broadest sense of the concept. Literally everyone everywhere benefits and the exclusion of anyone from the benefit is impossible. However, the depletion of mature natural forests and the loss of carbon stored in forests are most rapid at the forest frontier of the developing tropical countries - or at the margin of point D in Figures 3.1-3.3.

Various import restrictions, taxes and subsidies have been suggested for addressing the problem. Import restrictions such as certification are in favour today, but enforcing the forest management requirement for certification must be difficult while less expensive harvests from the natural forest remain available to local loggers. The natural solution for economists would be to tax the negative externalities, land conversion and fossil fuel consumption, while using tax revenues to subsidize the positive externality, forest management. Taxes on land conversion would be difficult because they largely occur in developing countries as a response to population growth, insecure land tenure or domestic development policy. Taxes on fossil fuel consumption have attracted more attention. Weimar (1990) showed that even a small tax on fossil fuel emissions comparable to 0.5 percent on the price of a barrel of petroleum would extract US$9 billion annually from the developed countries. However, we can anticipate that even this large sum would be rapidly dissipated in a) the transfers necessary to induce each of the many marginal users of the tropical forest to halt their extractive activities, b) the additional transfers necessary to encourage a significant expansion in tropical forest plantations, plus c) the costs of monitoring and enforcing (M&E) the promised behaviour by so many individuals spread over a vast forest landscape, as well as d) the administrative costs of the entire transfer.

For regions in all three stages of development, the costs of establishing and maintaining property rights C_r are greater than the market value of the resource at the natural forest frontier (point D). In the two less developed stages (Figures 3.1 and 3.2), these costs also exceed the value of marginal agricultural land use (from point B to point C). This means that M&E, for carbon sequestration or any other purpose, will be more costly than the market value of the protected resource. Few developing country resource management agencies will have the budgetary means to accomplish the M&E task under these conditions, while the local population will have competing private incentives to convert land and extract forest resources.

However, this same description also suggests an institutional solution: improved property rights. Any policy improvement that reduces the costs of establishing and maintaining property rights (reducing the height of function C_r) will extend the area of sustainable management (shifting B or B” to the right in the figures). This will decrease the level of forest conversion for regions in stage I, reduce the area of open access forest degradation for regions in stage II and expand managed forests for regions in stage III - thereby protecting threatened forests and stored carbon for regions in stages I and II and adding new forests and sequestering more carbon for regions in stage III. Two changes that could satisfy this description are the institutional adjustments that would permit local farmers to register their land use claims more easily and the transfers of these lands from public agencies to private land management. Successful community forestry is an example of the latter. General improvements in the overall macroeconomic policy environment are a third example. These are the kinds of improvements that increase the confidence of local land managers in their future prospects in general and cause them to extend their planning horizons and invest in conservation improvements in their land (see Section 3.4).

The second set of potential solutions is related to unplanned spillovers from agricultural activity. Decreases in the general class of incentives that induce agricultural expansion (e.g. crop subsidies) would decrease agriculture’s infringement on the natural forest for regions in stage I (shifting point C to the left), decrease agricultural infringement on the degraded open access forest for regions in stage II (also shifting C to the left) and permit forest management to out-compete agriculture and expand into some former agricultural land for regions in stage III (shifting point B’ to the left). Of course, altering these agricultural incentives will not be an easy task as strong political interests support them in almost all developed countries.

The issue of agricultural incentives becomes even more complex when it is recognized that some incentives, such as those that induce shifts to higher yield varieties, induce agricultural intensification. These would have the same beneficial effects on the forest and on carbon sequestration as those achieved by ridding the world of the incentives for agricultural expansion! Furthermore, agricultural intensification generally increases the demands for labour, raises rural wages and attracts workers away from the forest. This too saves on forest exploitation and controls the destruction of carbon already stored in mature trees. Therefore, any proposal to modify agricultural incentives for the purpose of improving carbon sequestration must be selective regarding the incentives in question. Some agricultural incentives reduce agriculture’s competition with forests, extend forest area and improve carbon sequestration, and some decrease carbon sequestration (Box 3.4).

Erosion control and general watershed protection: Erosion control and watershed protection include all the wind, water and soil movement services of trees and watersheds; for example, water supply, storm protection - especially in coastal areas - and control of the upstream loss of soil nutrients and downstream effects of sediment deposition. Like carbon sequestration, watershed protection may be divided into two broad classes of activity - those requiring new conservation investments such as tree planting in shelterbelts to deter wind erosion and along streams and gulleys to deter water erosion and those maintaining the services of existing forested watersheds and controlling their deterioration.

The difference in economic effect is that watershed values are local, while carbon sequestration is a global public value. Depending on the watershed management activity, it can benefit an individual landowner who makes the conservation investment and improves the productivity of his or her own land, or it can benefit a range of downstream or other off-site land managers in the same watershed. In the latter case, when watershed management is a public good, it involves a smaller and more local public than the global society affected by carbon sequestration, the benefits of the activity are clearer and more immediate for the local community and obtaining support and compliance with local public management is somewhat easier than for carbon sequestration.

Many of the first class of watershed management activity, those requiring new conservation investments, are responses to human development. They are a means of improving the productivity of existing (often agricultural) land uses. These investments typically occur on private lands - to the left of the intersection of the agricultural and forestry value functions with the rising property rights cost function in the three figures (points B or B”). Therefore, the increases in the manager’s private long-term productivity are generally sufficient to induce private conservation investment. In fact, this observation conforms to the established analytical economic evidence for developed countries where land use rights are generally more secure (e.g. Crosson, 1985; Crosson and Stout, 1984). More recently, Yin (2000) and Alemu (1999) demonstrated its reliability in China and Ethiopia, respectively, once farmers in those countries received new longer-term land use rights. That is, economic reforms in the 1970s and 1980s in both countries gave farmers more secure and longer-term land use rights. The farmers responded by increasing various inputs, including trees to protect against wind and water erosion. Farmers in both countries obtained statistically significant increases in the levels of their private agricultural productivity due to the newly planted trees and related conservation investments.

The second class of watershed management activity protects the upland watershed or the coastal wetlands for the benefit of off-site residents of the same local area. Grazing livestock are a common source of upland erosion and downstream sedimentation [e.g. the Lake Victoria watershed in central Africa (personal communication with ICRAF) and the Philippines (Cruz et al., 1988)]. Upland collection of fuelwood and fodder [e.g. Nepal’s hills (Dangi and Hyde, 2001)] is a second source in some severely degraded forest areas. Even less intrusive activities like picnicking can also affect local water supplies to the extent that the city of Portland (Oregon, USA), for example, restricts all human activity on one critical section of its adjacent Mt. Hood watershed. Finally, logging and land conversion in many coastal areas have damaged local fisheries and decreased local protection against the effects of tropical storms. These examples are all characteristic of either the open access degraded forest or the boundary of the mature natural forest in Figures 3.2 and 3.3 (between points C and D). The costs of protecting these watersheds exceed the open access private (grazing, fuelwood collection, recreation, timber harvesting or land conversion) values on these lands. That is, in terms of the figures, the function C_r explaining the costs of property rights is above the agricultural and commercial value functions. Therefore, private management will be unsuccessful and only public monitoring and enforcement can ensure common watershed benefits for the local community.⁶

Where most or all members of the local community share in common watershed benefits, then the local residents also share a common incentive to protect the watershed and monitoring and enforcement may be a relatively simple matter. In other cases, where the local incentives are dissimilar, M&E will consume more resources and protecting the public benefits will be a more difficult task (Box 3.5). For example, some poorer households in the village of Basantapur in Nepal rely on the watershed for fuelwood, but their fuelwood collection degrades the watershed and increases the off-site flow of soil to the detriment of the agricultural productivity of the better-off households at the base of the watershed. This community’s two interests in the watershed are in conflict and the community’s forest guards have been unable to fully restrict access. The degradation of the watershed continues (Dangi and Hyde, 2001).

Biodiversity and critical habitat: The protection of biodiversity is based on the idea that currently undeveloped, and even unknown, species possess characteristics of potential future value. Preserving their habitats helps preserve the species and the option to obtain future advantage from them once we learn more about them. Numerous examples demonstrate the great social and economic gain from preserving select species for which we had no prior scientific knowledge - although some doubt remains regarding the significance of the very probabilistic value to be gained from preserving unknown species in general (Simpson et al., 1996).

If the importance of preserving species is accepted, then we know that the remaining natural forest, the forest beyond the frontier at point D in our three figures, contains most of the critical habitat. That is, natural forests (rather than forest plantations, other managed forests and trees and the degraded open access lands) must be the focus of most of our preservation effort because only these forests contain undisturbed habitat. Furthermore, among natural forests, the tropical moist forest contains more than a half of all species of flora and fauna, including many that are still unknown. This particular forest is all the more important because its marginal habitats are more threatened than those of either the tropical dry forests or the temperate and boreal forests. Therefore, the remaining mature tropical moist forest is an appropriate focus for a large share of our interest in protecting biodiversity.

The economic problem, once more, is the mismatch between the locations of the most concerned public and critical habitats. It contains elements of similarity with the carbon sequestration problem just discussed. That is, the greater wealth of the developed countries puts them in a better position to act on the potential future value of biodiversity and to pay the price necessary to protect endangered habitats, while the poorer local farmers and extractive users of tropical forests bear most of the costs of foregone development and habitat protection.

However, these elements of similarity diminish in importance when we focus on the physical sources of carbon and biodiversity. Carbon sequestration improves with the protection and expansion of forests and trees on all marginal forest lands. Most lands and trees are of similar value for carbon sequestration, but all lands and trees are definitely not similar for biodiversity. Biodiversity requires the protection of selective “islands” of specialized habitat. In terms of our three figures, these islands are often in the neighbourhood of point D or beyond it and the foregone future development opportunities associated with them are often smaller than even the current opportunities foregone when protecting marginally commercial forest lands for carbon sequestration at points B’, B” and D. This means that many critical habitats remain unthreatened today simply because they are beyond the limits of access for either agricultural land conversion or extractive forest activities. Their inaccessibility protects them - for now. In other cases, some habitats can be protected in the normal course of management for other forest products and forest-based environmental services. Hyde (1991) showed that this is the case, for example, for many red-cockaded woodpecker habitats in the southern pine region of the US.

This leaves two fundamental problems for policy resolution, protecting that currently threatened habitat that is in competition with other, consumptive uses of the forest (or monitoring to ensure that its management and use is not competitive with the endangered habitat) and identifying additional specialized habitat that is inaccessible now and protecting it before it becomes threatened in the future. The currently endangered habitat will mostly often appear in the neighbourhood of point D in our figures. Protecting it requires the establishment of boundaries around the endangered habitat and the permanent exclusion of incompatible land uses. Excluding land conversion and agricultural use of the forest is a problem only for habitats at new frontiers in regions described by Figure 3.1. Therefore, reducing the incentives for agricultural expansion will only be important for the preservation of habitat in newly settled regions like some parts of the Amazon or the interior of Papua New Guinea.

Otherwise, for regions characterized by the more developed forestry described by Figures 3.2 and 3.3, the threats to diverse forest habitats originate from either the development of new infrastructure or extractive activities in the remaining natural forest. Changes in the infrastructure can affect all non-market forest values. Extractive activities in these natural forests are typically undertaken in a transitory manner by loggers or other operators who are here today and gone tomorrow. They have all the private economic incentives to harvest resources, often timber, and controlling their activities is a difficult task. It requires continuous monitoring of the boundaries of the protected habitats as long as the habitat is to be protected. Therein lies another problem. The resource management agencies of most developing countries do not have either the human or the financial resources to adequately ensure the protection of many critical habitats. Furthermore, most donor-assisted projects and most debt-for-nature swaps fail to include long-term continuing maintenance costs. In fact, long-term commitment to any single project is unusual in the practice of donor agencies. Therefore, the long-term protection of currently threatened habitat remains a largely unsolved policy problem and, in the presence of very limited resources, we have the additional problem of deciding which of many threatened habitats is most critical and most worthy of the long-term demands on these limited resources.

The second fundamental policy problem has to do with those critical habitats that are currently inaccessible and unthreatened. If the broad regions within which they are contained do develop, then some of these habitats may not be threatened in the future either. That is, local forests tend to recover after regions develop beyond some level. However, many critical habitats could be lost forever before the regions around them attain this level of development. Can we identify those threatened habitats and preserve them before they become commercially accessible and are lost forever?

Identification is a task for field research and it is one reason for the current interest in developing ‘indicators’ of forest quality. As current opportunities for biotechnology expand, the values of some endangered habitats may become clearer to private investors and those private interests may undertake the effort to save some of them. Otherwise, this too is a public problem and one that the donors and public research institutions of the developed countries may be particularly well suited to address. Once the public researchers do identify critical habitat, however, policy-makers are still left with the previously discussed and largely unsolved task of finding a publicly-funded means for excluding extractive uses of the forest from the critical habitat.

Finally, ex situ preservation can be a partial solution for long-term preservation. Ex situ preservation is the off-site storage of the endangered biological material. Once the threatened flora or fauna is identified, its germplasm can be catalogued and stored, preferably in a public institution that will maintain the biological material for the benefit of all the world’s people in this as well as future generations. Zoological parks are one example. In addition, some institutions in the CGIAR system of international agricultural research stations maintain cold storage banks of germplasm. Gene banks, however, are static institutions. They fail to capture and preserve dynamic genetic evolution. They can be important contributors to the preservation of existing biodiversity while we search for ways to identify and permanently protect critical forest habitat (Box 3.6).

Tourism: A broad range of people obtain the benefits of tourism and a broad range of forested sites provide for it. Those who benefit range from wealthier global tourists to local picnickers and others just looking for a pleasant moment in a natural setting. Site characteristics range from unique global resources like Yellowstone, the Serengeti or Sagarmantha to pleasant local forested groves and even village parks. The unique sites can provide a focus for substantial demands for tourist support services like restaurants, motels, guide services and outdoor equipment shops; and these can be important sources of hard currency for the local economy.

Once more, the economic problem is either one of protecting “islands” of specialized forested sites within lands that are valuable for other, commercial uses (out to point D in our figures), or identifying commercially inaccessible forest lands (beyond D) that are uniquely attractive for nature tourism and protecting them before they become accessible for those commercial uses. For the most unusual sites, fees may be charged at points of limited access and the revenue from them used to establish boundaries and to monitor and enforce the exclusion of undesirable uses of the park, as well as the provision of tourist services. For example, Kenya charges an increased visa fee for national entry for global tourists, Nepal charges for trekking permits which may be obtained at the offices of guide services and many national parks with natural boundaries charge a gate fee.

There are two problems, however, even in the case of these unique resources. First, the fee is general. It makes no distinction for particular resources or specialized services within the region covered by the fee. For example, fees collected at national borders do not distinguish between different parks within a country or between particular resources within a park. Therefore, there is no market signal that assists managers in the allocation of revenues among numerous resources and services. Sometimes this problem can be addressed by placing the general management of all resources under one integrated operation - a national park service, for example - and competitively allocating concessions for specialized services like hotels and guides with each park boundary.

The second problem exists even when this first problem can be solved. While tourists from long distances can be excluded unless they pay the costs of park management, the exclusion of local users is more difficult. Therefore, monitoring and enforcing restrictions against competitive local uses of the unique resource is more difficult. Poaching within East Africa’s game parks and timber harvesting within the boundaries of southeast Asian natural reserves are examples. A partial solution to this problem can often be obtained by establishing an interest in the park’s tourist services within the local population. When some of the local population gain, as from employment in park concessions or by providing guide services, then local citizens have their own incentives to assist in the M&E of park boundaries and management and their own incentives to discourage their neighbours’ violations of the park boundaries. Even this solution, however, does not prevent all local trespass.

For those forest resources that are not unique and do not attract global tourists the most common non-consumptive users are members of the local community. Local institutions are generally better suited to manage these resources and many successfully protected village parks and forest sanctuaries have been observed around the world. The employees of local institutions have a better understanding of both the patterns of demand on these resources and the arrangements of their most appealing physical characteristics. Therefore, the management costs of a local institution will be lower than for a comparable level of management by a national authority. Nevertheless, these costs will still be positive and, since the exclusion of local users is difficult, the local community must bear these costs either as part of its community budget or as part of a commonly respected decision not to exploit the extractive resources within the park (Box 3.7).

CONCLUSIONS

It should be clear from this review that various selective policies can have favourable consequences for the protection of particular forest resources. For example, reducing forest taxation and increasing forest incentives improve the opportunity for managed forests - but only in those regions like the southern US where the products of managed forests are financially competitive with extracted products from the natural forest. It should also be clear that some policy instruments have not had the desired effects - particularly on timber supply. For example, reducing forest taxation and increasing forest incentives has no effect whatsoever on regions like Alaska, Siberia or many tropical countries where managed forests are not yet financially competitive with extraction from the natural forest. And regulations that control harvest and shipment may even decrease the natural market incentives for both the development of managed forests and the protection of natural forests.

The most important lesson, however, is that it is crucial to consider the policy objective first, then to think about the best way to achieve it. Often the best means of achieving a forest policy objective is by modifying activity in an adjacent sector like agriculture, or by improving the local institutions providing for property rights of the land that provides the forest resource in question. Very often modifications in road design improve the likelihood of achieving a particular forest policy objective as well.

In sum, it should be clear that policy spillovers from other sectors of the economy often have even greater impacts on forest resources than forest policies themselves. Furthermore, perhaps the general problem of rural poverty is the greatest source of deforestation and loss of the various forest resources. Therefore, general economic development and associated improvements in personal economic welfare may have greater positive effects on the protection of these resources than any set of selective and specific policy or institutional improvements. The focal objectives of modern forest policy are sustainability and the control of forest degradation and deforestation. Sustainability is achievable if we define it as “sustainable options” or “maintaining, in perpetuity, options for all different uses of forest resources, market and non-market, consumptive and non-consumptive, known and unknown.” This definition would permit some shifts of forest land to agriculture, from agriculture back to forestry, or from natural forest to managed forest, so long as the land’s productive base and the genetic base of the forest are maintained. Values will change with time and preferred arrangements of land use will change with them, but we can ensure that changes in land use do not destroy opportunities for new and different land or resource uses in the future.

This perspective of sustainability is consistent with minimizing the area of degraded open access forest while locally regulating specific erosive watersheds, critical habitats and important aesthetic resources, both within and outside the degraded area. Minimizing the area of degraded forest is the objective because its elimination is impossible as long as secure property rights impose a cost and the public agencies responsible for managing the degraded open access area have limited budgets. The primary means for minimizing the degraded area are to reduce the cost of property rights and attract some human activity away from the forest. The first requires finding the best bundle of property rights and the institution that can provide this bundle at least cost. The second implies finding improved employment opportunities outside the forest for the lowest wage users of the forest. Once more, policies that improve general rural economic welfare may have the most beneficial impacts on the forest and, as a corollary, policies that are detrimental to the rural poor may have the greatest negative influences on sustainability and the control of deforestation.

REFERENCES

Alemu, T. (1999). Land tenure and soil conservation: evidence from Ethiopia. Unpubl. Ph.D thesis, Economics Department, Goteborg University, Sweden.

Amacher, G., Hyde, W. & Kanel, K. (1999). Nepali fuelwood consumption and production: regional and household distinctions, substitution and successful intervention. Journal of Development Studies, 35(4):138-63. Also in P. Parks and I. Hardie, eds. Economics of Land Use. Dordrecht. Kluwer (forthcoming).

Antle, J. (1983). Infrastructure and aggregate agricultural productivity: international averages. Economic Development and Cultural Change, 31(3):609-19.

Bass, S., Thornber, K., Markopoulos, M., Roberts, S. & Grieg-Gran, M. (2001). Certification’s impacts on forests, stakeholders and supply chains. London. IIED.

Bluffstone, R. (1995). The effect of labor markets on deforestation in developing countries under open access: an example from rural Nepal. Journal of Environmental Economics and Management, 29(1):42-63.

Boyd, R., & Hyde, W. (1989). Forestry sector intervention: the impacts of public regulation on social welfare. Ames. Iowa State University Press.

Cooke, P. (1998). The effect of environmental good scarcity on own-farm labor allocation: the case of agricultural households in rural Nepal. Environment and Development Economics, 3(4):443-469.

Crosson, P. & Stout, T. (1983). Productivity effects of cropland erosion in the United States. Unpublished mss. Resources for the Future.

Cruz, W., Francisco, H. & Conroy, Z. (1988). The onsite and downstream costs of soil erosion in the Magat and Pantabangan watersheds. Journal of Philippine Development, 15(1):48-85.

Dangi, R. & Hyde, W. (2001). When does community forestry improve forest management? Nepal Journal of Forestry, 12(1):1-19.

Feder, G., Just, R. & Zilberman, D. (1985). Adoption of agricultural innovations in developing countries: a survey. Economic Development and Cultural Change, 33:255-297.

Feder, G., Onchan, T., Chalamwong, Y. & Hongladarom, C. (1988). Land ownership security, farm productivity and land policies in Thailand. Baltimore. Johns Hopkins University Press.

Foster, A., Rosenzweig, M. & Behrman, J. (1997). Population and deforestation: management of village common land in India. Unpubl. draft mss. Department of Economics, University of Pennsylvania.

FAO (2001). Global forest resources assessment 2000. Rome. FAO Forestry Paper 140.

FAO/UN RWEDP (Rural Wood Energy Development Programme) (2000). Wood Energy News, 15(1). Bangkok.

Fitzgerald, B. (1986). An analysis of Indonesian trade policies. World Bank CPD discussion paper 1988-22.

Gillis, M. (1988). Indonesia: public policies, resource management and the tropical forest. In R. Repetto and M. Gillis, eds. Public policies and the misuse of forest resources, pp. 43-114. Cambridge. Cambridge University Press.

Godoy, R. (1992). Determinants of smallholder commercial tree cultivation. World Development, 20(5):713-25.

Government of Indonesia, Ministry of Forestry and Estate Crops (1998). 1997-1998 forest utilization statistical yearbook. Jakarta. DG of Forest Utilization.

Gunatilake, H. (2002). Technical efficiency in sawmilling: implications for rainforest conservation in Sri Lanka. Unpubl. mss. Department of Agricultural Economics, University of Peradeniya.

Hansen, E. & Juslin, H. (1999). The status of forest certification in the ECE region. Geneva timber and forest discussion papers ECE/TIM/DP/14.

Heydir, L. (1999). Population-environment dynamics in Lahat: deforestation in a regency of South Sumatra province, Indonesia. In: B. Baudot and W. Moomaw, eds. People and their planet, pp. 91-107. New York. St. Martin’s Press.

Hyde, W. (1991). The marginal costs of endangered species management: the case of the red-cockaded woodpecker. Journal of Agricultural Economics Research, 41(2):12-19.

Hyde, W., Dalmacio, M., Guiang, E. & Harker, B. (1997). Forest charges and trusts: shared benefits with a clear definition of responsibilities. Journal of Philippine Development, XXIV(2):223-56

Hyde, W. & Kohlin, G. (2000). Social forestry reconsidered. Silva Fennica, 34(3):285-315. Also chapter 12 in W. Hyde, G. Amacher and colleagues (2000) The Economics of Forestry and Rural Development, p 243-287. Ann Arbor. University of Michigan Press.

Hyde, W. & Seve, J. (1993). The economic role of wood products in tropical deforestation: the severe experience of Malawi. Forest Ecology and Management, 57(2):283-300.

Jodha, N. (2000). Common property resources and the dynamics of rural poverty: field evidence from the dry regions of India. In: W. Hyde, G. Amacher and colleagues. Economics of Forestry and Rural Development, pp. 203-222. Ann Arbor. University of Michigan Press.

Lopez, R. (1998). The tragedy of the commons in Côte d’Ivoire agriculture: empirical evidence and implications for evaluating trade policies. World Bank Economic Review, 12(1):105-132.

Mendelsohn, R. & Neumann, J. (eds.) (1998). The impacts of climate change on the American economy. Cambridge. Cambridge University Press.

Meyer, L. & Shobayashi, M. (2001). Multifunctionality: toward an analytical framework. Unpubl. working paper of the working party on agricultural policies and markets of the committee for agriculture, joint working party of the committee for agriculture and the trade committee, OECD.

Migot-Adholla, S., Hazell, P., Barel, B. & Place, F. (1991). Indigenous land rights systems in sub-Saharan Africa: a constraint on productivity? World Bank Economic Review 5(1):155-75.

Nordhaus, W. & Boyer, J. (1999). Warming the world: economics models of global warming. Cambridge. MIT Press.

Pattanayak, S., Mercer, D.E., Sills, E., Yang, J. & Cassingham, K. (2002). Taking stock of agroforestry adoption studies. Unpubl. mss. USDA Forest Service, Research Triangle Park, N.C.

Rice, R., Gullison, R. & Reid, J. (1997). Can sustainable management save tropical forests? Scientific American, (4):44-49.

Richter, C. (1966). The trees. Athens. Ohio University Press.

Russakoff, D. (1985). Timber industry is rooted in tax breaks. Washington Post (24 March 1985), p. A2 ff.

Ruttan, V. (1982). Agricultural Research Policy. Minneapolis. University of Minnesota Press.

Scherr, S. (1995). Economic factors in farmer adoption of agroforestry: patterns observed in western Kenya. World Development, 23(5):787-804.

Sedjo, R. (1994). The potential of high-yield plantation forestry for meeting timber needs: recent performance and future potentials. Resources for the Future discussion paper 95-08.

Sedjo, R. (1999). Land use change and innovation in US forestry. In: R. Simpson, ed. Productivity in Natural Resource Industries, pp. 141-174. Washington, DC. Resources for the Future.

Sedjo, R. & Solomon, A. (1990). Climate and forests. In: N. Rosenburg, W. Easterling, P. Crosson and J. Darmstadter, eds. Greenhouse warming: abatement and adaptation. Washington, DC. Resources for the Future.

Simpson, R., Sedjo, R. & Reid, J. (1996). Valuing biodiversity for pharmaceutical research. Journal of Political Economy, (1).

Smith, W. (2002). The global problem of illegal logging. Tropical Forest Update, 12(1):3-5.

Stein, T. (2001). Savings grow on trees along Front Range. Denver Post (April 18), pp. 1A, 9A.

Weimar, D. (1990). An earmarked fossil fuels tax to save the rainforest. Journal of Policy Analysis and Management, 9(2):254-259.

Yin, R. & Hyde, W. (2000). The impact of agroforestry on agricultural productivity: the case of northern China. Agroforestry Systems, 50:179-94.

Yin, R. & Newman, D. (1998). Impacts of rural reforms: the case of the Chinese forestry sector. Environment and Development Economics, 2(3):289-304.

¹ The 47 percent is an estimate from personal communication with D. Garrity in May 2000. The official estimate of 24 percent is from the Government of Indonesia, Ministry of Forestry and Estate Crops (1998). 1997-1998 Forest Utilization Statistical Yearbook. Jakarta.
² The agroforestry evidence follows the agricultural experience, although the analytical evidence is not as extensive. See Pattanayak et al. (2002) for a review.
³ The experience is similar in the Republic of South Africa where all the commercial wood comes from plantations that were originally established in the 1890s by a government that recognized Europe’s demand for a reliable timber supply. Certification imposes no significant cost on these plantations and almost all of them are certified.
⁴ Where agriculture expands into the open access frontier (stage II) or converts land away from managed forestry (stage III), the policy has induced absolutely greater use of both capital and labour, but it remains relatively capital-consuming and labour-saving.
⁵ For example, see the discussion by F. Bohlin on the RESECON listserv June 25, 2000.
⁶ Nevertheless, Landell-Mills and Portas (2002) identify over 180 cases of markets for watershed services.