deforestation and forest transition: theory and evidence ...zhangy3/deforestation and forest...

WORLD FORESTS FROM DEFORESTATION TO TRANSITION?

41

1 Introduction

While the forests in the developed coun-tries have ceased to shrink in area and haveeven begun to expand � a reversal whichin this article is called forest transition �the developing countries are still in the de-forestation stage. The growing demands of

an increasing population for agriculturalland and for timber are widely recognizedas the most important causes of deforesta-tion. However, the scarcity of timber, andindustrialization are generally viewed asaccounting for forest transition (Rudel1998). Both conclusions seem to conflicteach other but are likely to be correct. The

FOREST TRANSITION

Deforestation and Forest Transition:Theory and Evidence in China

by Yaoqi Zhang

Abstract: A general theoretical framework on deforestation and forest transition is presented fol-lowed by empirical evidence from China. Relative scarcities of food, timber and environmentalgoods resulting from both population and economic growth are believed to be the most funda-mental causes of forest change. A relative scarcity of population � a factor of production as well �is considered to drive population change and re-allocation. The time required from deforestationto forest transition may be prolonged by the time lag in forest regeneration, and by the transactioncosts, i.e. the costs in transferring, defining and protecting property rights of land and forests. Theinstitutional issue is specially addressed throughout this article because the �exclusion cost�, i.e.the ex post cost of transaction in property rights of land and forests, is relatively large comparedwith other aspects of property protection. Consequently, active forest management is not econom-ically justified on a large part of land which otherwise should be under active management. Thisarticle concludes with a preliminary forecast of the future trend in China�s forests and policyimplications on future forest development.

Keywords: Deforestation; forest transition; reforestation; transaction costs; economic reform; landuse; population; China.

M. Palo and H. Vanhanen (eds.), World Forests from Deforestation to Transition?© 2000 Kluwer Academic Publishers, Dordrecht.

42

FOR

EST

TR

AN

SIT

ION problem is that we do not identify the ori-

gins of forests and analyze them separatelybased on the origins. Forests can be eithernatural endowment, or social products, butthey have different responses to socio-eco-nomic environments. The net change inforests is determined by the total effects.

Joining in the current debate and con-cerns regarding deforestation in developingcountries, an attempt is made to formulatean integrated theoretical framework of thedeforestation and forest transition and toprovide empirical evidence from China.The forecasts and policy suggestions to bepresented are argued to be relevant not onlyto China but also to other countries.

2 Framework of Deforesta-tion and Forest Transition

The theoretical framework of forest landchange, which is extended from the modelby von Thunen (1875), is illustrated by Fig-ure 1. Land is used for the option whichcreates the highest land rent in a competi-tive land market. Therefore, the land useoption and forest management method aredetermined by the land quality and socio-

economic environment. Land quality is abroad concept here, referring not only tosoil, steepness, incidence of rocks, sourcesof water and exposure to wind and sun, butalso to distance and types of roads to mar-ket and habitation, etc. The socio-economicenvironment refers to the output and inputprices, and institutions. The input and out-put prices can be measured on-site and in amarket. Prices may vary greatly when trans-portation costs are significant.

2.1 Land Use for Agriculture versusForestry

The land rent for forestry or any other landuse is determined by the output prices, theinput costs and the land quality. For sim-plicity, agriculture is considered as the onlyother land use option. If we use (p

a ,w

a) and

(pf ,w

f) to represent the set of output price

and input cost by agriculture and forestryrespectively, the land rent for agricultureand forestry with land quality q will bepa(pa,wa,q) and pf(pf ,wf ,q). If pa(pa,wa,q)>pf(pf ,w,q), the land will be allocated to ag-ricultural use. Otherwise, the land will beused for forestry. For instance, the land tothe left of A in Figure 1 should be allocat-

Figure 1: Land use for agriculture and forestry by land quality


43

ed to agricultural use, and the land to theright of A should be used for forestry. Onthe boundary, the rents for agriculture andforestry are equal.

Point A in Figure 1 shifts when on-siteoutput or input prices for agriculture andforestry and land quality change. For sim-plification, we just consider the changes ininput and output prices and do not consid-er land quality change that may derive fromroad expansion, soil degradation, etc.Graphically, if prices for agricultural goodsrise (or input price decreases), the curve ofthe agricultural rent will shift to the right,and some of the forestry land will be de-voted to agriculture. Alternatively, if theprices of forest products rise (or its inputprice decreases), some of agricultural landwill be devoted to forestry.

Mathematically, to understand the effectsof the change in forest product prices onthe shift of the forest-agriculture frontier,the total derivative of the equationpa(pa , wa , q*) � pf(pf , wf , q*) = 0 (q* refersto the boundary land where rent createdfrom forestry and agriculture is equal) withrespective to q* and pf is: (¶pa/¶q*) × dq*� (¶pf /¶q*) × dq* � (¶pf /¶pf) × dpf = 0.Therefore, (dq*/dpf) = (¶pf /¶pf) / [(¶pa/¶q*)� (¶pf /¶q*)]. As long as the higher qualityland is used for agriculture, [(¶pa/¶q*) �(¶pf /¶q*)] will be positive, so (dq*/dpf) isalso positive since (¶pf /¶pf) is positive. Thenew equilibrium point q* will thereforemove to better land quality q** when pric-es of forest products rise, thus some of thehigher quality land currently used for agri-culture will be devoted to forestry (q**> q*). By the same procedure, it is possi-ble to derive the land use shift betweenforestry and agriculture in response to thechanges of agricultural products and theinputs in forestry and agriculture as

Forestry land = Af (pa ,wa , pf ,wf ) [1] � + + �

It must be emphasized that land conver-sion is not costless and the cost is oftensunk. In reality, land use conversionbecomes much less flexible when the costassociated with conversion is considered,while land modification, referring tochange within one category, e.g. among dif-ferent agricultural crops, is quite flexible.Land use conversion from forestry to agri-culture is more likely to happen when treesare grown to (or quite close to) an econom-ically mature age. The capital in silvicul-ture is sunk and cannot be recovered (evenpartly) if the trees are too young. Land con-version from forestry to agriculture requiressignificant investment. It may take years tocut down the trees, remove stumps andstones, erect buildings and fences, todevelop large-scale irrigation and drainagesystem, and to expand the initial clearingsinto productive farms. Most of such invest-ment is also sunk. If the expected outcome(the prices and costs) does not differ great-ly from the reality, the land is likely to re-main in agricultural use.

Forest land conversion to agriculture andother uses might be significant during theearly stage of economic development, butit would not be critical with respect to thetotal land area available for forestry. Themore critical problems are: (1) the stagnanttimber price, poor credit markets, risk aver-sion, and poorly developed institutions donot justify forest investment for the logged-over land; (2) forestry may be unprofitableon a quite large area abandoned by otherland uses due to soil degradation and othersocio-economic reasons. That is why de-forestation is much faster than theexpansion of agricultural land in most de-veloping countries.

2.2 Management of Forest Land

FAO (1995a) defined forest land underactive management as �forest and other

44

FOR

EST

TR

AN

SIT

ION woody land that is managed according to a

professionally prepared plan or is otherwiseunder a recognized form of managementapplied regularly over a long period (fiveyears or more)�. Land will be under activeforest management if 0< pf(pf ,wf ,q) andpf(pf ,wf ,q) > pa(pa ,wa ,q). In Figure 1, onlythe land to the left of B is under active for-est management.

The amount of investment and the lengthof rotation, the decision variables in forestmanagement, are functions of the land qual-ity, output and input prices. To calculate theoptimal investment and rotation, the Faust-mann model can be applied

where, pf (q) is annual land rent created byforestry for land quality class q; p(Q, q) ison-site stumpage price, which is a functionof the status of forest (Q), e.g. biomass den-sity and tree species and land quality suchas steepness and distance from the marketunder given market price; w is on-site unitcost of silviculture, which is a function ofthe market costs (both materials and labor)and the distance from the labor supply, thelikely labor market and materials supply, ris the capital cost or simply interest rate; Tand E are length of rotation and on-site unitsof silviculture efforts; g is annual accumu-lative in situ value of forest at age t. yrepresents factors such as population, in-come and economic structure, which affectthe demand for in situ forest products (Parket al. 1998).

The optimal solution of rotation T andsilviculture efforts E can be obtained fromthe first order maximizing conditions. Anextreme case is that the optimal solution is

. This means that the forest manage-ment is undertaken for only environmental

pI��T��

7 (�

PD[ U{��

7

× J�W��(��Y��T��H�UW�GW��

S�4�T�4�7��(��S�Z�T�H�U7±�Z�T�(}��[��±�H�U7�]��

[2]

7��

7��

purposes: if forest does not exist, trees willbe planted; if the forest is already existing,logging will not be undertaken, e.g. natu-ral reserves and public parks. However, amore common solution could not be ,i.e. joint management of timber produc-tion and environmental services could bemore common in active forest manage-ment regimes. Clear cutting is graduallyregulated, while its alternative, selectivecutting, is more accepted due to apprecia-tion of environmental value. It means thenew rotation always start from a certainage of trees.

Non-Active Forest Management

The unexploitable forests, which are locat-ed to the right of D in Figure 1, means thatthe stumpage price is zero, p(Q, q) = 0. Ofcourse, land value for forest managementis also zero. Since the forests already exist,their value is not related to soil productiv-ity, but is closely related to the status of theforests (Q), depicting species, the growingstock of the forests, the distance and theaccessibility to the market, and harvestingand transport technology.

Except for the unexploitable forests, allforestry land is assumed to be under activemanagement based on traditional produc-tion economics which consider notransaction costs. Because forest regenera-tion can be achieved naturally, without anyor very minor on-site silviculture efforts formost logged-over land ¶Q(T, E; q)/¶T>0when E»0, the p(Q, q) shall become posi-tive when Q reaches a certain level, e.g.the original level after a certain years. Sinceit is theoretically possible for the forest torecover to its original biomass density, pos-itive rent is possible, and active manage-ment has value.

However, much land world-wide istoday not under active management. Thisis particularly so in developing countries.


45

The land not under active management isoften interpreted by economists to be theresult of open access. On the other hand,while the change from open access toclosed access can create Pareto efficiency,it is difficult to understand why open ac-cess is so common in developing countries.The problem lies in that traditional produc-tion economics does not consider transac-tion costs. The transaction costs associatedwith land ownership change includes exante and ex post elements (Mattews 1986).It is not costless to move from open accessto closed access.

Increasing timber prices or decreasingextraction costs � due to road constructionor improved logging or transportation tech-nology � will make more remote forestsincreasingly exploitable, shifting the for-est frontier (Point D in Figure 1) and proba-bly some secondary forests (Point C) to theright. Frontier forests on a large scale are lo-cated in remote and less populated regions.Since most of these forests are owned bythe state or are in open access, any changein the frontier will greatly depend on pub-lic policies in road expansion, populationsettlement and logging regulations.

Between these two extremes, there issubstantial forest or forestry land that isunder non-active management, the area

between B and D in Figure 1. Because theexclusion cost is the most importantreason that makes forest management un-justified, the shifts of Point B and C inFigure 1 are highly sensitive to general so-cio-economic, political and institutionalconditions. Figure 2 shows the differentbiomass growth under open and closed ac-cess.

The Cost of Exercising Ownership

Here the ex ante costs of transaction arenot discussed, as these are associated withdefining and transforming property rights.The ex post costs of transaction are moreimportant regarding trees and forested land,but these are often ignored by academicstudies. The common situation is that theland is given ownership, even legal rightseither by the state, the community, privatepersons or organization, but the owners failto exercise rights when the gains from im-plementing their exclusion rights aredeemed insufficient. Thus the land is leftin public domain, in open or quasi-openaccess.

The costs of ownership not only includethe cost of obtaining titled or recognizedrights, but also the cost of protecting exist-ing rights. If exercising existing rights is

Note: If E = 0, the early growth function of closed access is the same as open access before time t.

Figure 2: Forest biomass density growth under open access and closed access

46

FOR

EST

TR

AN

SIT

ION too costly, the ownership becomes nomi-

nal: the legal rights are retained but not theeconomic rights (for more about the differ-ence between legal rights and economicrights see Barzel 1997). Therefore, an an-swer to open access could be the high exante and ex post costs associated with trans-actions where the land is originally notowned, or with ex post �exclusion costs�associated with transaction where the landis already owned. In other words, the actu-al land value is the profit calculated by thetraditional production economics minus thetransaction costs of obtaining and protect-ing the land and forests.

When the exclusion cost is added to theproduction function, Equation [2] is trans-ferred as follows

where m is the annual protection costs forforests at age t, which is a function of thestatus of forests, the institutions (F ), landidentity, and perhaps other socio-econom-ic variables. Now it is much easier tounderstand why quite a large land area haszero land rent, and active forest manage-ment is not justified. In this case, theoptimal solution is E = 0, and T is the timewhen the biomass density rises to justifyextraction. The m is also a decision varia-ble. The owners must make a trade-offbetween a possibility of losing their eco-nomic rights and investment in protectingtheir rights. For simplicity, let us unrealis-tically keep it as cost of protecting the fulleconomic rights with no possibility of loss.

To emphasize the importance of exclu-sion costs, let us use China as an exampleto show how big the magnitude of the costs

can be in association with property rightprotection. According to officially docu-mented statistics from the Ministry of For-estry (1992), there were one million peo-ple employed in preventing illegal accessto forests. Of these more than one thirdwere full-time jobs in the early 1990s.There were another 200 000 officials whoseduties were associated with designing andimplementing regulations, and assisting dis-puted settlements concerning forest prop-erty rights. These figures demonstrate theconsiderable magnitude of the costs of pro-tection. Forest management has to consid-er these costs. Even so, illegal logging isstill rampant in China due to the character-istics of the forests such as easy access, noclear maturity, and traditional conventions,norms and laws as well as wide-spread pov-erty.

Illegal extraction in forests, particularlyin common forests (not to be confused withopen access forest), is often looked uponas being less criminal than in agriculturebecause, historically, trees are �wilder� thanagricultural crops. Trees are often viewedto be in the public domain, at least to someextent. On the other hand, if stealing a treeworth one-month�s wage is without muchrisk of being caught and with no seriouspenalty if caught, illegal logging will re-main difficult to prevent. It is not surprising,therefore that Poore et al. (1989) concludethat non-active forest management is socommon and sustainable forest manage-ment is so rare in the tropics.

The Cost of Exercising Exchange

As discussed above, the costs of exercis-ing ownership can be interpreted as part oftransaction costs. Now we turn to anotherclosely related transaction cost: the cost ofexercising exchange through the market.This was interpreted by Coase (1937) asa more fundamental reason for the emer-

pI��T�� 7 (�

PD[ U{��

7

× J�W��(� Y��T��H�UW�GW�

��S�4��T��4�7��(��S��Z��T�H�U7�±�Z�T��(��

±�

7

× m��W��(� F��T��H�UW�GW��}��[��±�H�U7�]��

[3]


47

gence of firms. He defined the transactioncost as �the cost of using price mecha-nisms�, �the cost of carrying out a transac-tion by means of an exchange on the openmarket� or simply �marketing costs�.Coase (1960) further explained transactioncosts as �to discover who it is that one wish-es to deal with, to inform people that onewishes to deal on what terms, to conductnegotiations leading up to a bargain, todraw up contract, to undertake the inspec-tion � and so on�.

The Faustmann model clearly does notinclude the marketing cost of the output �neither the cost of organizing the produc-tion, nor the �purchasing� cost of inputfactors, such as labor and capital. This maynot be important if such costs are relative-ly small compared with the value inexchange. However, these costs are notsmall due to the characteristics of forestproducts (or stumpage) and the status ofmarket instruments in the developing coun-tries. Even in an advanced economy, thecost of capital, the simplest product, is usu-ally more than 5%, the difference betweeninterest rates of deposit in and loan from abank. The long persistence of self-suffi-ciency economies in developing countriesis largely consequent upon the cost ofexchange through the market.

2.3 Causes for Shifts in Land Useand Management

Point A in Figure 1 shall shift when on-siteoutput or input prices for agriculture andforestry � or land quality � change. The on-site price change is often difficult to ob-serve. More generally, the land use changemay be considered to have resulted fromthe changes associated with population,infrastructure, technology, or institutions.These factors are not addressed separately,but discussed in the framework of popula-tion and institutional changes.

The Role of Population Change

In the short term, it is assumed that popu-lation settlement remains static. Point A inFigure 1 could be very sensitive to thechange of the relative market price of tim-ber and other goods that are largely basedon land. During the early stage of econom-ic development of a country with substan-tial natural forest growing stock, the elas-ticity of timber price to population growthis likely to be less than the elasticity of foodprice to population growth. Consequently,agricultural expansion is likely, and PointA in Figure 1 usually shifts to the right.

The above is based on static human set-tlement, but populations are always seekingfor cheap resources, including land andforest resources. Population growth chang-es the land-labor ratio. Consequently,changes in the relative prices of land andlabor lead to an increase in the value of in-ferior land. With respect to forest resources,the decline in the forest-labor ratio increas-es the relative value of forests and decreasesthe value of labor. Thus increasingly infe-rior forest becomes valuable and laborspreads out to seek its higher value. In oth-er words, when population migration is lesscostly than transporting the resources, pop-ulation will move to locations close to theresources. The population movement canoccur either as a slow spread of popula-tion, or as big new settlement projectswhich are rapidly established. This proc-ess will continue as long as any virgin landand frontier forests remain. Historically,population migration plays a much moreimportant role in forest change than thelocal growth of population.

The general path on forest change in thecontext of population change is, therefore,as follows: First, as the population growsand the economy develops, the speed withwhich A in Figure 1 shifts to the right foragricultural expansion, and the speed with

48

FOR

EST

TR

AN

SIT

ION which D shifts to the right for timber ex-

traction from frontier forest is faster thanthe movement of B to the right. The areabetween B and D is largely bareland, openforest and shrubs, and non-actively man-aged forest land. Whenever a large forestarea or other land resources are not inhab-ited and used, population migrations intothese areas are likely to occur. Therefore,deforestation could occur for a quite longperiod of time. When Point D approachesE, pushing D to the right is difficult becauseof increasing marginal costs. The speedwith which B moves to the right increasesas price increases. Finally, forest expansionbecomes dominant. At this time, large pop-ulation movements are mostly character-ized by urbanization. Forest cover wouldcease to decrease, remain stable, or perhapseven begin to increase.

The Role of Institutions

The impacts of institutions on economicperformance have been acknowledgedsince John Stuart Mill, Adam Smith andKarl Marx. Recently, institutions havereceived great attention by institutionaleconomics that include economics of trans-action costs and the economics of propertyrights (see e.g. Coase 1937, 1960; Lewis

1955; Kuznets 1966, 1973; Stigler 1966;Alchian and Demstz 1972; North and Tho-mas 1973; Dahlman 1979; Cheung 1983;North 1981, 1990; Williamson 1975, 1985;Barzel 1997). Using similar arguments thatapply to more general economic perform-ance, it is hardly controversial that changinginstitutions may have great influence onland use. It is still argued as to what kindof causality relationship exists betweeneconomic performance, institutions andtechnology. It seems plausible that they areinteractive.

In the long-term, institutional evolutionmay be greatly affected by the process ofsocio-economic development. Along withthe increase of scarcity, the gains fromtransformation from open access to closedaccess will gradually exceed the costs andhence have an effect on institutional ar-rangements. Open access may change intoclosed access. In addition, the scarcitymight justify transfers of ownership, forexample state forests might be transferredto community ownership. These again maybe transferred into private ownership. Thatis why we may find much less communityowned land in the developed than in thedeveloping countries. However, increasingscarcity of environmental goods mightagain justify some transactions from pri-vate forest land into public domain. Thegeneral pattern of land property regimes isillustrated in Figure 3.

Institutional changes, either in propertyright institutions or market (non-market)institutions, may not always evolve effi-ciently and instantly. Institutional rigidityand inertia may persist for a considerableperiod of time. Inertia may prevent institu-tional adaptations to the environmentalchanges as well as causing institutions tobe inefficient. Information and knowledgeconcerning alternative institutions are al-ways limited, the judgment and decisionmaking in choosing therefore vary from

Figure 3: The rent gradient and propertyregime (Bromley 1991)


49

society to society. In addition, institutionalarrangements, particularly in the short term,are to a large extent the product of publicchoice under the balance of power amongdifferent interest groups. Significant differ-ences are observable in the institutions ofcountries with similar environments.

Following the insights of transaction costeconomics, the characteristics (as immobil-ity, long-waited maturity, difficulty indelineating property rights) of relativelylarge magnitude of the transaction costscompared to production costs in forestmanagement, make forestry more sensitiveto institutional arrangements. In develop-ing countries agriculture, which is moretechnology oriented than silviculture, hasreached levels which are not too far behindthe level in developed countries. In forest-ry, there is much more poor managementin developing than developed countries.Even silvicultural technology is in generalstill primitive in developing countries com-pared with developed ones. The differencein the performance of forest managementin developing countries compared with de-veloped ones is largely due to institutionaldifferences.

3 History ofDeforestation Process

China has a good record of informationregarding forest and land use changes. Verydetailed records have been kept on vegeta-tion, flora, fauna, climate and agriculturedevelopment for more than 3000 years.However, until the 1950s, there was littlematerial in the English language, except forsome reports contributed by professionalplant hunters from Europe and America inthe late 19th and early 20th centuries, andby a few Chinese scholars who were edu-cated abroad. According to Murphey(1983), Jean Baptiste du Halde�s descrip-

tion of the Empire of China, published inParis in 1735 and in English in London in1738, could be the earliest work for west-erners. Perkins (1969) explored agriculturalland development from 1368 to 1968. Shaw(1914) also could also be among the earli-est commentators on China�s forests andtimber supply. Murphey (1983) focused ondeforestation in the 19th and the first halfof the 20th century. Menzies (1994) de-scribed the historical pattern of land useand deforestation in China by referencingancient literature.

If the current territory of China is divid-ed by a diagonal line from the northeast tothe southwest, the original vegetation cov-er can be classified into three categories:Forest land dominated the area with anextensive and dense forest cover of over70�90% to the right of the diagonal; grass-land and forest with 20�30% forest coverdominated in the transition zone, along thediagonal line; and as the third categorydesert and bareland dominated with lessthan 5% scattered forest to the left of thediagonal. The types of forests varied fromboreal forest in the northeast, temperatedeciduous broadleaf in the north, and mixeddeciduous and evergreen broadleaf forestin the south, with tropical forests in thesouthernmost region of China. Thus, abouthalf of the land was originally covered byforest. Liu and Wang (1989) estimated thatin 2700 BC there were six provinces withmore than 90% forest cover and 14 prov-inces with more than 50%.

3.1 Loess Plateau andGuangzhong Region

Significant deforestation began at least asearly as 3000 BC in the Loess Plateau Area,the place of the origin of the Chinese civi-lization. A temporal process of deforesta-tion started from the heartland of theGuangzhong region, currently Gansu and

50

FOR

EST

TR

AN

SIT

ION Shaanxi Provinces, at the junction of the

Wei and Yellow Huang Rivers. Accordingto archeological records, the populationwas dense in this area around 3000 BC. Theoriginal status of the vegetation was and isa subject of much debate between advo-cates of the forested vegetation view (Teng1947; Wen and He 1979; Wang and Sheng1994) and the non-woody vegetation view(Handel-Mazetti 1931). Based on informa-tion contained in the Book of Odes (ShiJing) and the Classic of Mountains andRivers (Shan Hai Jing), among other sourc-es, it is more widely accepted in China thatthe forest was dominant before human in-tervention. Because this site is located inan ecological transition zone from moist tosemi-arid and arid climate, it would morelikely vary between forest-dominated andgrass-dominated vegetation, rather than ariddeserts. Menzies (1994) considers that theimage of the pristine landscape of north-ern China was neither of dense forest norof barren steppe, but rather of a mosaic ofvegetation: riparian forest followed thecourses of rivers, with rather sparse decid-uous woodland and brush on the mountainsand denser deciduous forest on the foothills.

It is widely recognized that the naturalenvironment of the Loess Plateau Area wasgreatly damaged around 1000�500 BC.There are interesting records which con-cern the Wei and Yellow Rivers in thisrespect. According to Wei Feng in Shi Jin,the rivers nearly 3000 years ago were veryclear, indicating no soil erosion. The riversmay still have been occasionally clear 2500years ago. Thereafter the rivers were alwaysyellow. The Yellow River, probably namedduring the period 320�311 BC, indicatesthe acute soil erosion and high silt loads atthat time. It should be noted that climatechange has been suggested as the cause ofthis deforestation. However, the fragileenvironment and climate change only madethe recovery of vegetation more difficult.

It is logical to assume that the causes ofdeforestation and degradation are to befound in agricultural expansions that be-gan around 3000 BC and reached theirclimax around 1000 BC, following loggingas the population grew and spread.

3.2 North China

As the environment deteriorated, the LoessPlateau region shrank and the populationwith its economic and cultural centersmoved eastward along the Yellow River tothe current Shanxi, Henan and Shandongprovinces. Rich forests covered the TaiMountains region, consisting of severalhigh mountains: Tai, Lu, Yi, Men, Daze andKunlun among others. According to ancientliterature, the vegetation was dominated bynatural forest and grassland 3200�3000years ago (Wang and Sheng 1994). Thisview is supported by many notations on theactivities of large wild animals, includingtiger, elephant, rhinoceros and others re-corded in the Meng Zi-Teng Wen Gong, ShiJi-Zhou Ben Ji, Lu Shi Chun Qiu and otherclassical literature.

According to the Meng Zi-Gao Zi, landreclamation had begun at the time of theZhou Dynasty (1100�256 BC) duringwhich significant areas of natural forestwere logged and transported to nearby are-as, which were the most civilized and pop-ulated in China at that time. The populationdensity was still low, however, and peoplemainly resided in the lowlands. By 500 BC,the forest had been logged extensively, butmuch forest still remained in the middle ofmountains due to difficult accessibility. Theforests, in general, were seriously damagedduring the period 500�0 BC. In the follow-ing years, logging and agricultural expan-sion steadily progressed to the mountainousrange and spread southward and northward.The agents of deforestation were mainlylogging, followed by pastoral husbandry


51

and steadily increasing reclamation towardsthe marginal area.

The population continued to move north-ward and the capital was relocated inBeijing during the Yuan Dynasty (1271�1368) in 1297. Rich forest grew on the YanMountains. This site was located in a so-cio-economic and political transition zone,varying from developed, permanent agri-cultural area in the south (North ChinaPlains) inhabited by the Han Chinese, tothe pastoral agricultural area in the north,inhabited by minorities. This location,according to some research, was still dom-inated by forests and grass land and mainlyinhabited by nomads since only marginalreclamation was occasionally carried outby Han prior to the Yuan Dynasty. In thebeginning of the Yuan Dynasty, this areawas listed as a hunting preserve of the em-perors and was well protected. The growingpopulation, together with an increasinglysignificant economy, promoted the demandfor food, fuelwood and timber. The wholeprocess of deforestation was fulfilled dur-ing the Yuan Dynasty, Ming Dynasty(1368�1644) and Qing Dynasty (1616�1911). The driving forces were intensivehunting, repeated logging for timber andfuelwood, followed by pastoral husbandryand cultivation. Fuelwood collection isassumed to be a critical factor in deforest-ation, since fuelwood collection, regardlessof the size of the trees, may be repeatedtoo frequently and intensively. In the be-ginning of this century, no large area offorests remained in this region.

3.3 South and Southeast China

During the late Song Dynasty (960�1279),southern China was experiencing a rapidpopulation growth and migration from thenorth when northern China was occupiedby the Mongolians. The capital was trans-ferred to Hangzhou in the current Zhejiang

Province in 1127. A significant area of for-est was converted into agricultural landduring the following decades. Because ofits mountainous landscape and favorableclimate, southern China still had an abun-dance of virgin and secondary forests du-ring the Song and the Ming Dynasties andthe early part of the Qing Dynasty. The mostsignificant loss of forest occurred duringthe late Qing Dynasty, particularly duringthe war of Taiping Rebellion (1851�64).

The population was not limited to themainland, a large number of the peopleimmigrated across the Qiongzhou strait toHainan Island during the Song and YuanDynasties. From the Han Dynasty (206 BC�220 AD) to the Tan Dynasty (618�907), thisisland was populated along the southeastcoast mostly by the minority Li. Shiftingcultivation was in fact firstly widely car-ried out along the southeastern seashorewhere the Li first settled. They then retreat-ed steadily to the central mountains afterthe Song Dynasty. At that time, the shiftingcultivation was gradually replaced by semi-permanent cultivation, followed by perma-nent cultivation in the northwest andsoutheast. With the introduction of newtools, the areas along the seashore settledby the Han, who had come there from themainland, steadily came under permanentcultivation. Before the Song and Ming Dy-nasties, large areas of virgin forests stillexisted since the forest was mainly used asthe source of non-wood forest products:berry collecting, hunting and limited har-vesting for household use. Around the 16thcentury, timber extraction for industry anddirect supply to be market began. At thebeginning of the 20th century Hainan wasthreatened by a lack of timber (Pen 1922).

Due to the difficult accessibility to thecenter of Hainan Island, there was still largeareas of tropical forest in the middle of themountains at the early part of the 20th cen-tury. The most significant timber extraction

52

FOR

EST

TR

AN

SIT

ION from the rain forest took place during the

20th century, particularly from the 1950sto the 1980s. Most of the tropical crops,particularly the rubber tree were introducedaround 1900, and have promoted signifi-cant agricultural expansion, especiallyduring the last 50 years.

3.4 Southwest and Northeast China

The southwest parts of Sichun and Yunnanprovinces still had a lot of natural forestsremaining at the beginning of the 20th cen-tury. The most significant deforestationoccurred during the last 50 years. Current-ly, only Tibet and some parts of Yunnan,for instance Xishunbanan, still have anyremains of virgin forests.

Northeast China had large areas of vir-gin forest at the latter part of the 19th cen-tury. The Changbai Mountains region wasalmost pristine land before the Qing Dy-nasty, and it was still well preserved by theemporers as a traditional Manchurian baseand imperial hunting ground in the earlyQing Dynasty. Han (Chinese) settlementsbegan to spread northeastward beyond theGreat Wall on large scale only after 1870when the Qing authority weakened and thepopulation pressure in North China in-creased. The earliest Han immigrants, how-ever, first arrived on the Liaodong Plains,south of Liaoning Province, which is thesouthwest to the Changbai Mountains. Thepopulation was still small relative to the

large territory. Only marginal natural for-est was logged by the end of the 19th cen-tury. Russian railway construction in thelate 19th century and the Japanese takeo-ver in northeast China in the early 20th cen-tury promoted significant deforestation inthe Changbai Mountains. It is estimated thatabout 100 mill. m³ of timber had been trans-ported out and 6 mill. ha of forests waslogged, accounting for approximately halfof the total area (Liu and Sun 1985).

The remaining forests in the ChangbaiMountains region were further logged af-ter 1949 when the People�s Republic ofChina was founded. Over 30 forest bureauswere set up after 1949 with the intention tolog the forests. About 10 mill. m3 of indus-trial logs flowed out of the region and hugeamounts of timber were annually consumedlocally (Ministry of Forestry 1987). Fromthe 1950s to the 1980s, the logged areaamounted to 1.5 mill. ha. It is estimated thatby the end of the 20th century there will beno accessible mature natural forest left.Large areas of forest land were convertedinto agricultural use to serve the growingpopulation, which has more than doubledduring the last 50 years. Soil erosion is be-coming very serious. For example, in theseven counties located east of LiaoningProvince, more than 0.2 mill. ha of culti-vated land have had to be abandoned dueto soil erosion.

Heilongjiang Province, located to thenortheast of the Changbai mountains, has

Table 1: Forest resources decline in Heilongjiang Province

(Heilongjiang General Bureau of Forestry 1987)

<HDU� )RUHVWHG�DUHD�

�PLOO��KD��

*URZLQJ�VWRFN�

�ELOOLRQ�P��)RUHVW�FRYHU�UDWH�

��

�

1RWHV�

�� 4LQJ�'\QDVW\�

�� -DSDQHVH�LQYDVLRQ�RI�0DQFKXULD�

�� (QG�RI�-DSDQHVH�RFFXSDWLRQ�

��

��

�


53

a shorter history of deforestation (see Ta-ble 1). One of the most important forestregions is the Xiaoxinganling range that isbordered by the Heilongjiang River in thenorth, by the Songhuajiang River in thesouth, the Daxinganling range in the north-west and a large plain in the southwest.Logging was started during the Japaneseoccupation, but only along the railway fromHarbin to Jiamushi. From 1950 to the mid-1980s, a total of 24 state-owned loggingenterprises were set up, and about 0.3 bill.m3 of logs were produced, approximately1/7 of China�s total industrial timber pro-duction during that period.

Most of the state-owned logging enter-prises are now short of mature forestresources in Heilongjiang Province. From1950 to 1978 in the Yichun prefecturealone, in the heart of Xiaoxinanliang, thegrowing stock decreased from 4146 mill.m3 to 2414 mill. m3, the share of conifer-ous forest in total inventory decreased from68.5 % to 51%, and the proportion of Ko-

rean pine in the total from 13% to 6.4%(Ministry of Forestry 1987). In addition,agricultural clearances claimed much of theforest land. For example, in Heihe alone,the area under agriculture increased by35 000 ha from 1962 to 1980.

3.5 General Pattern of Deforestation

To summarize, China�s deforestation start-ed in the Loess Plateau Area, moving east-ward along the Yellow River to the currentShanxi, Henan and Shandong Provinces,then spread northward to the current He-bei and Beijing regions, and finally movedbeyond the Great Wall to northeast China.Meanwhile, the population moved south-ward to southern China, even across the seato Hainan Island. In general, the plains andlocations along major rivers were usuallyoccupied and cultivated first, then pro-gressed to low hilly regions and finally tomountainous ranges. Table 2, using six re-gions as an example illustrates the popula-

Table 2: Deforestation in six selected sites

� /RHVV�3ODWHDX�DUHD�

7DL��PRXQWDLQV��

<DQ��PRXQWDLQV�

+DLQDQ��,VODQG��

&KDQJEDL�PRXQWDLQV�

;LDR[LQJDQOLQJ��DUHD�

� � � � � � �$UHD��NPô��

��

� � � � � � �/HQJWK�DQG�SHULRG��RI�GHIRUHVWDWLRQ��

��\HDUV��%&²��%&��

��\HDUV��%&²��

��\HDUV��²��

��\HDUV��²��

��\HDUV��²��

��\HDUV��²��

� � � � � � �7KH�ORZHVW�IRUHVW��FRYHU�UDWH�

��

� � � � � � �&XUUHQW�SRSXODWLRQ�GHQVLW\��SHUVRQV�NPô��

��

� � � � � � �0DLQ�DJHQWV�RI�GHIRUHVWDWLRQ�

ORJJLQJ��VXEVLVWHQFH�DJULFXOWXUH��FOLPDWH�FKDQJH�

ORJJLQJ��VXEVLVWHQFH�DJULFXOWXUH��

ORJJLQJ��VXEVLVWHQFH�DJULFXOWXUH�

ORJJLQJ��VXEVLVWHQFH�DQG�WURSLFDO�FURSV�FXOWLYDWLRQ�

ORJJLQJ��FXOWLYDWLRQ�

ORJJLQJ��FXOWLYDWLRQ��

� � � � � � �8QGHUO\LQJ��FDXVHV�

SRSXODWLRQ��JURZWK�

SRSXODWLRQ�JURZWK�

SRSXODWLRQ�JURZWK�

SRSXODWLRQ�JURZWK��WLPEHU�GHPDQG�RXWVLGH�WKH�UHJLRQ�

WLPEHU�GHPDQG�RXWVLGH�WKH�UHJLRQ��SRSXODWLRQ�JURZWK�

WLPEHU�GHPDQG�RXWVLGH�WKH�UHJLRQ��SRSXODWLRQ�JURZWK�

54

FOR

EST

TR

AN

SIT

ION tion movement, deforestation, and agents

and underlying causes.Over these thousands of years, there has

been a cycle of deforestation, mild recov-ery, and more severe deforestation (Liao1987). In general, intensive logging in thefrontier forests and better management ofthe secondary forests coexisted at the be-ginning of each dynasty because relativelygreater population growth and the strongereconomy demanded more agricultural landand timber. Better management was alsosecured by a more stable political regimeand a strong government. Rampant forestdestruction, due to wars, conflicts and weakgovernment and institutions, was frequentlyfound at the end of each dynasty.

The demand for agricultural land andtimber following human settlement areclearly the primary causes of forest loss.However, the rate of agricultural land ex-pansion is much lower than that of thepopulation, and deforestation is much fasterthan the agricultural land expansion (seeFigure 4). As early as the Tang and Yao era(about 2500 BC), shifting cultivation wasextensively carried out and continued forone thousand years during the Xia and Shan

Dynasties (2100 BC�1100 BC). During theearly Han Dynasty (206 BC�220 AD), theChinese began a systematic land reclama-tion and irrigation schemes.

During the 20th century, the area of ag-ricultural land has not significantly increa-sed, but the loss of forests continues. A largepart of the loss is not because the land isconstantly occupied by other uses, but rath-er because it is abandoned and not con-verted to active forest management. Thisapplies to the logged-over forest land.

4 From Deforestation toForest Transition

Data concerning China�s forest resourceswas published in the China Yearbook in1943 and on at least ten other occasionsbefore the first national forest inventory(1973�1976). All the data, basically rely-ing on the data of 1943, stated that the for-est cover rate ranged from 5% to 8%; the8% is more often cited. These figures, how-ever, are misleading. The mistakes result-ed from both incorrect estimates andvarying definitions, such as definition offorest, the territory of China, etc. Unfortu-nately, many public officials and academ-ics, both domestic and international, haveused this data as a point of reference (Chu1988). Thus, this 1943 forest area becamethe most widely recognized benchmarkwhen the People�s Republic of China wasfounded in 1949.

When the first national inventory waspublished, the forest cover rate was givenas 12.7%. It seemed that forest resourceshad significantly increased over the 8% fig-ure given for 1949. In fact, it is reasonableto believe that the 1949 forest area propor-tion was about 15% assuming the samedefinition of forests. Be that as it may, theprocess of deforestation continued duringthe first 30 years of the People�s Republic

Figure 4: Changes in population and cultivatedland area in China, 1400-1995 (Perkins 1969; Wang

1992; Ministry of Agriculture 1950�1995)


55

of China. Logging in northeast and south-west China may have had social and eco-nomic benefits, but the radical deforestationduring the Great-Leap-Forward in 1958 andthe Cultural Revolution from 1966�1976 wasa complete abuse of natural resources.

However, since the 1970s, China seemsto have embarked on a road to forest tran-sition. The analysis here is mainly basedon the four national forest inventories con-ducted during 1973�1976, 1977�1981,1984�1988 and 1989�1993. FAO (1995b)assumes that the reliability class is 1 forthe state estimates and 2 for the change as-sessment (range from best = 1 to worst =3). The present study divides China�s terri-tory into five regions: Northwest, North,Northeast, South-Southeast, and Southwest(see Map 1). The land use in these five re-gions is illustrated in Figure 5. From thefigures, some general conclusions can bedrawn: the turn from contracting to expand-ing forest area in the Northwest of Chinaoccurred during the late 1970s; in the Northand South-Southeast the turn occurred dur-ing the early 1980s; in the Northeast andSouthwest the transition started during thelate 1980s and early 1990s.

Table 3 illustrates in more detail the on-site land conversion among different cate-gories. The estimates are based on 200 000repeated on-site observations. In general,the shifts of land use from one category toanother have become less significant. Theincrease in forested area was more likelyderiving from open forest, shrubs and bare-land. Therefore the increase of the forestland area does not necessary mean a con-version of other productive land use cate-

Table 3: Land use conversion between forestry and other land uses in China, 1980�1986 and1986�1992

Map 1: China�s regions and provinces

Note1: Other land uses might not necessarily mean non-forested land. In China�s statistical system, the definition offorestry land and forested land refer to the potential use and the sectoral ministry who manages the land. For instance,some forested land owned by agricultural farms is likely to be listed as agricultural land. Therefore, the real conversionbetween forested land and other land uses is much smaller that the figures show.

(Ministry of Forestry 1989; 1995)

� ,QWR�IRUHVW�ODQG��PLOO��KD��

)URP�IRUHVW�ODQG��PLOO��KD��

1HW�IRUHVWHG�DUHD�FKDQJH��PLOO��KD�� ²��

� � � � � � �

� ��²�� ²�� ²�� ²�� ²�� ²��

2SHQ�IRUHVW� �� ²��

%XVK�DQG�VKUXE� ��

1HZ�SODQWHG�IRUHVW� ��

%DUHODQG� ��

2WKHU�ODQG�XVH��

7RWDO� ��

56

FOR

EST

TR

AN

SIT

ION gories into forest land. The conflicts be-

tween land use are often overestimated.

Scarcity of timber and EnvironmentalGoods

As discussed in the theoretical section, ascarcity of timber makes more remote for-

ests exploitable, while the scarcity of landand forest resources compared with popu-lation causes populations to move. How-ever, as was indicated earlier, the relativescarcity of timber also justifies active for-est management. Hyde et al. (1993) andHyde and Seve (1996) had an intensive dis-cussion on the response to timber scarcity

Figure 5: Major land uses in China by regionsfrom 1970s to 1990s (Ministry of Forestry 1978,

1982, 1988, 1995; China�s Statistical Bureau 1975�1996)

Northwest China

North China

Northeast China

South-Southeast China

Southwest China

Note: The provinces for each region refer to Map 1. Thebareland only includes the technically forestable land,so the sum of land is possibly not equal to total land area;the build-up area includes urban and industrial areas,roads, etc; some discrepancies are due to statisticalproblems, particularly in Southwest, where the boundaryof Tibet was not quite clear.


57

in terms of combating deforestation andargued that deforestation has an economiclimit that precedes its physical limit. Fig-ure 6 shows the scarcity of forest resourcesas measured by the deflated timber price.Prior to the late 1970s, timber price wastightly controlled by government. Pricemight not be good as a measure of scarcityof timber, as the price did not show a largeincrease. However, timber was listed as oneof the top scarce products. Since then, tim-ber prices have risen throughout the 1980s.Since the mid-1990s, timber prices havesomewhat decreased because of a growingtimber supply from the plantation forests,timber imports and some changes in theeconomic structure of the country.

Figure 7 indicates that the timber pro-duction in regions with poor forests, suchas in Northwest and North China, increasedearlier than in the well forested areas, suchas in Northeast China. The transition mayin fact come out as an increase in timberproduction through a change from a regimeof non-actively managed forests to a regimeof actively managed forests.

It should particularly be emphasized thatscarcity includes environmental goods thatgenerally do not yet have a market price. Ifwe view the economy as being open, tim-ber can be imported from other regions orother countries, so scarcity � if we use priceas the measurement � can hardly go beyonda certain level. But most environmentalgoods are supplied on-site and cannot beimported. Local supply is therefore oftennecessary. An increasing share of the af-forestation in China is environmentoriented as it is in other countries and re-gions. As an example forest managementin Singapore and Hong Kong or other high-ly populated regions is clearly not fortimber production, but for the supply of en-vironmental goods.

In China, the timber scarcity since the1990s has been eased to some extent due

to massive plantations of fast-growing treesand imports. However, the demand for en-vironmental goods is also growing fast dueto population growth and economic devel-opment. In China more than 16 mill. ha ofprotection forests, accounting for 14% ofthe total forested area, is aimed at environ-mental services. The protection forests aremainly located in the three northern areas(North, Northwest, Northeast; see Table 4).Ten of the eleven recently implemented

Figure 6: Timber procurement price index inChina, 1951�1995 (China�s Statistical Bureau 1989and 1996)

Figure 7: Industrial timber production in Chinaby regions, 1974�1996

Note: The price is deflated by the retail price index.

58

FOR

EST

TR

AN

SIT

ION

massive afforestation programs were initi-ated specially to combat environmentalproblems (Zhang et al. 1999). In addition,about 9 mill. ha of forestry land are pre-served as natural reserves, for eitherenvironmental or ecotourism purposes.

Rudel (1998) has argued that a scarcityof timber might not induce reforestation be-cause of potential substitutes for wood andthe time lag of reforestation. Substitutes forwood really ease wood scarcity, may pre-vent wood price increase and can delay thereforestation. However, looking for woodsubstitutes is a process actually induced bythe scarcity timber. The time lag of regen-eration is really considerable, and withsustainable forest management a consider-able time between two rotations is required.Forest change is thus sometimes misunder-stood, partly because of the short-termactivities overshadow the long-term per-spectives.

Institutional Transition

Palo (1994) among others has argued thatthe forest transition is often not realizedbecause of market failures and inappropri-ate property rights, political unrest andmissing information in developing coun-tries. This argument only explains that scar-city is not a sufficient reason, but does not

deny it as necessary reason for the foresttransition. It must be realized that scarcityis the motivation for ownership and mar-ket exchange. China experienced shortag-es of timber supply in the 19th century(Shaw 1914; Pen 1922; Murphey 1983).Without considering either the transactioncosts or risks, China�s forest transition � atleast at a regional level � should have oc-curred in the late 19th century: in fact, for-est management in China has over 1000years of history. Unfortunately, between thelate 19th and the mid-20th centuries, Chi-na suffered from continuous wars. Con-fronted with such unrest the environmentfor long-term investment, such as forestmanagement, was hostile.

Within the People�s Republic of China,great efforts have been made concerningafforestation. China was widely recognizedas having the world�s largest forest planta-tion area (Westoby 1975; FAO 1978). Thecontinued political struggles did mean thatafforestation measures, more specificallyinstitutions governing the measures, provedto be ineffective. Logging on the naturalforests still outweighed the efforts in affor-estation. The radical deforestation duringthe Great-Leap-Forward in 1958 and theCultural Revolution from 1966�1976 wascompletely a result of institutional prob-lems.

Table 4: Main environment oriented forests in China

(Ministry of Forestry 1989; 1995)

5HJLRQ� 7RWDO�ODQG�DUHD��PLOO��KD��

3HUFHQWDJH�RI�ODQG�DUHD�RI�&KLQD�

7RWDO�IRUHVWHG�DUHD��PLOO��KD��

3URWHFWLRQ�IRUHVW�DUHD��PLOO��KD��

� � � � �7KUHH�1RUWK�5HJLRQ� �� 8SSHU�DQG�PLGGOH�UHDFK�RI�<DQJ]H�5LYHU�

��

� � � � �&RDVWDO�UHJLRQ� �� 7KH�SODLQV�UHJLRQ� �� 7RWDO��

�


59

The inappropriate assessment of a 8%of forest cover in 1949 and 12.7% in 1976,as well as rampant false reports on the af-forestation achievement, overshadowed thedeteriorating status of China�s forests. Onlywhen the data from second national forestinventory (1977�1981) showed a declineof the forest resources, was the serious stateof China�s forests given attention by thegeneral public. A growing demand forwood, resulting from both populationgrowth and economic development, had torely on huge imports of forest productsduring the 1980s. The state-owned loggingenterprises also faced a serious economiccrisis. Deforestation resulted in an increasein environmental damage and the frequen-cy of natural disasters.

These signals and concepts of scarcitycalled for a revision of the general forestpolicies. To increase both the forest coverand the growing stock became a top prior-ity. The strategy was changed from �min-ing� the forests � moving Point D and C toright in Figure 1, to a silviculture-basedforestry � moving Point B to the right. Largescale funding and efforts have been allo-cated to silvicultural investments. Nume-rous massive programs for industrial tim-ber supply, soil and water conservation,coastal natural disaster prevention, biodi-versity preservation, etc. have been imple-mented (Zhang et al. 1999).

Institution Building in Forestry

Institutions have clearly played an impor-tant role in both deforestation and the for-est transition. China�s land tenure haddeveloped already at a much earlier timeto the level of today�s capitalistic countriesin the beginning of the past century. Almostall agricultural land and a large part of theforest land had been owned by landlordsand some rich peasants for centuries. Someinstitutions and land market mechanisms

for such private ownership had also beendeveloped. Those institutions have been thebasis for socio-economic development.Unfortunately, the land reform from 1949to 1952 and the follow-up socialist trans-formation, completely destroyed these in-stitutions. In the short term, the land reformdid have some positive effects on rural de-velopment and equal income distribution,but the long-term negative impact is sig-nificant and far-reaching. Building institu-tions is much more difficult than destroyingthem. North (1991) pointed out that insti-tutional change is a complicated process.Although formal rules may change over-night as the results of political or judicialdecisions, informal constrains embodied incustoms, traditions, and codes of conductare much more impervious to deliberatepolicies.

Recent reforms are trying to rebuild theinstitutions and management organizationswith respect to (1) clarification of forestryland property rights through forestry landowners� re-titlement. While there is still norecognition of private land ownershiprights, at least private land-use rights andforest property rights are recognized; (2)the role of law in forest management isbeing addressed, and significant forces arebeing provided to protect economic rightsin forests, as well as helping to providesolutions where disputes occur. The ForestLaw was firstly enacted in 1985 and amend-ed in 1998; (3) a free timber trade isgradually being permitted which recogniz-es the economic rights to forest products;(4) the de-collectivization and reorganiza-tion of collective forestry land is progres-sing through the Household Responsibili-ty System and the Share-holding Systemand other types of joint or co-operativemanagement; also (5) de-centralization ofthe state-owned forestry land is occurringthrough dissolution of the managementauthority and budget regime.

60

FOR

EST

TR

AN

SIT

ION Institution building has proved difficult

and time consuming, not only in term ofthe institutions themselves, but also interms of the credibility of the government.Even after two decades of considerable ef-forts by various governments promisingthat forests owned either privately or col-lectively will be legally protected by laws,the farmers are still in doubt whether thegovernment will always maintain such apolicy. This again demonstrates how diffi-cult it is to rebuild institutions once theyhave been destroyed. However, these re-forms have already helped the foresttransition to some degree (Ruiz-Peres et al.1996; Yin and Newman 1997; Song et al.1998), and will play a more important rolein the future.

5 Prospects for the Future

Forecasts of China�s forest and land usetrends have been made by the Chinese So-ciety of Forestry (1987) and Gu (1988)among others. Errors in the forecasts arebecoming obvious. The problems lie in thefact that they are based either on past trends,government planning or political opinions.Even though some of the models are verysophisticated and seem to consider com-prehensive variables, the most importantmechanisms driving the forest transitionand the response to scarcity are still nottaken into account. Alternatively, let usexamine a very simple way to predict thefuture of forest land change.

Forestable versus Unforestable Land

Before the conversion between differentland uses is analyzed, technically unforesta-ble land must be excluded. Unforestableland refers to the biological limits of treegrowth because of factors such as temper-ature, soil, altitude and water. Currently,about one third of the land in China, most-

ly located in the northwest and southwestChina, is to some extent technically un-forestable. Of course, �unforestable� is arelative concept in the long-term. Some ofthe most favorable �unforestable� terrainhas been and will be forested, while someforested land and potentially forestableland has degraded into unforestable due tothe loss of reversibility resulting from heavyand repeated damage, for instance over-harvesting and over-grazing. Significantshifts in land use do not seem likely, in spiteof the great efforts such as silviculturaltechnology developments, which have beenmade to overcome some of the biologicallimits in the northwest region. Meanwhile,some marginal land continues to degradeto become unforestable. A minor net posi-tive change depends on the economicstatus, technological development, policyand public awareness of the environmentin the future.

Forestry versus Competing Land Uses

The boundary of forestry land and non-for-estry land is Point A in Figure 1. The causesof the boundary change has been discussedabove. Economically, agricultural use isbecoming less competitive: agriculturalproducts in general are more expensive thanimports. China�s entrance to the WorldTrade Organization will further lower thegeneral price level of agricultural goods.The long history of agriculture has left lit-tle potentially arable land in existingforested land, so the possible shift from for-ested land to agricultural land could becomparatively minor.

On the general issue of food security,China will invest in its agricultural infra-structure, which is intended to increase pro-ductivity and improve some of the degradedagricultural land. In the short term, importsof agricultural products are expected to in-crease. The policy for food self-sufficien-cy, at least at the regional level, is gradually


61

losing support due to its inefficiency. There-fore, increased government subsides foragriculture can be neither justified nor ex-pected. Since the social price of the out-puts from forest management is muchhigher than the timber market price, it islikely that subsides to forestry, governmentdirect investments and loans, will be in-creased. Agroforestry and urban forest de-velopment are other alternatives forovercoming land limitations and combin-ing forestry and agriculture.

Point A in the Figure 1 is not very ap-propriate because it only indicates the shiftbetween forestry and agriculture. Theoret-ically, a growing population and industria-lization will occupy more agricultural land,putting more pressure on agricultural landand pushing agricultural frontier A to theright, utilizing directly forestry land. How-ever, current population movements andindustrialization are characterized by roadimprovements, expansion and urbanization.In other words, new large settlement sites,even the establishment of small villages,as in the 1950s and 1960s, is less possibleand not encouraged by the government.Forest management may therefore benefitmore than agriculture in reduced costs re-sulting from road expansion and improve-ments.

To sum up, the total net conversion fromforestry land to the competing land uses,including agricultural, residential and in-dustrial land, may no longer be significant(see Table 3 and Figure 5). As a nation, Chi-na has �matured�. As far as land use is con-cerned, there will never again be radicaland rapid changes in major land uses suchas occurred during history. The changewithin each category is likely to be greaterthan the change between the categories.Changes in major uses of land will tend tobe localized since no large population mi-gration is possible except from rural areasto urban areas. The conversions, at least the

net change from forestry land to other cat-egories, are gradually becoming insignifi-cant and in the future may possibly be fromother land uses to forestry.

Forests in the Future?

The changes within forestry land refers tothe area located to the right of A in Figure1. Future forest changes will mostly bewithin this category. Firstly, consider PointD or C in Figure 1. Given the current pop-ulation density and distribution, Point D hasalmost approached E. That means Chinahas few frontier forests and little virginland. The remaining virgin forests, togeth-er with some secondary natural forests,have been protected to fulfill environmen-tal functions. The extensive floods of thelate 1990s across China led to a re-appraisalof the natural forests, resulting in the fullimplementation of the bans on logging innatural forests. Marginal timber extractionfrom natural forests might be unavoidablebut only selective cutting is allowed. In thissense, these forests are neither frontier for-ests nor non-actively managed forests, butmanaged forests and their value is reflect-ed in their environmental functions.Therefore, neither a small rise in timberprice nor a decrease in extraction costshould be a threat to them. In fact, the in-crease in environmental value caused bythe growing demand and decreasing sup-ply of environmental services will help theirprotection, management and expansion.Increasing large-scale protection forestsand nature reserves are good examples ofthis process.

Secondly, consider Point B in Figure 1.Forest land under active management �mostly plantation forests � are likely toexpand, assuming that timber prices willcontinue to rise. This assumption may bevalid since less timber will come from nat-ural forests, and more subsidies or tax cuts

62

FOR

EST

TR

AN

SIT

ION can be expected. More importantly, improv-

ing the socio-economic conditions andinstitutions, including laws and propertyrights security, will reduce the protectioncosts and the risks that investors must bear.The development of institutions has alreadyplayed a role in the forest transition andwill continue to be important in the sus-tainable expansion of forests.

Finally, consider the area from B to C inFigure 1, that is to say, forestry land underthe non-active management. This area maybe the most important section for futurechange, as about half of forestry land, or15% of the total territory of China, is inthis category (see the categories of bare-land, open forest and shrubs in Figure 5).As already discussed, it is the transactioncosts rather than the traditional productioncosts which make active forest managementuneconomic. The role of institutions, whichis important in reducing transaction costs,will be even more important for its transi-tion into active management.

To forecast the future changes in the for-ests of China, attention should again begiven to the land between B and C in Fig-ure 1, denoting bareland, open forest andshrubs. The area of this land category ac-counts for 15% of the total land area ofChina. Focusing on this land and consider-ing the future development of the economy,institutions and markets, some simple pre-dictions can be made as follows:

· In the short term (5 years, up to 2005),on the assumption that the current dis-torted timber price with heavy tax andfees continues, institutions regardingland tenure and macroeconomic policywill improve steadily but will not changesignificantly. It is expected that 17% ofthe land area of China, equal to the cur-rent forest cover of 15% plus 2% throughafforesting 5�15% of the land betweenB and C, is likely to be covered by for-est in 2005.

· In the medium term (20 years, up to2020), on the assumption that institu-tions will experience some significantprogress, the distorted timber supplyprice will be adjusted and more non-market values of the forest will beconsidered. It is appropriate to use theinternational prices of agricultural goodsand wood products as shadow prices.Under this assumption, around 22% ofthe land, which is equal to the currentforest cover 15% plus 7% through af-foresting 50% of the area between B andC, could be covered by forest.

· In the long term (50 years, up to 2050),it can be assumed that institutions willreach the same level as in the currentdeveloped world, and the non-marketvalue of water and soil conservation,carbon sequestration, biodiversity pres-ervation and other environmentalfunctions will be considered in govern-ment decision-making. It can thereforebe expected that about 28% of the land,over 200 mill. ha, which is equal to thecurrent 15% plus 13% through afforest-ing 90% of the area between B and C,will probably be under active manage-ment.

6 Concluding PolicyImplicationsDeforestation is unavoidable in any coun-try with rich virgin forests. The growingdemand for agricultural land and timber byeither a local population, or populationgrowth through resettlement, are importantreasons for deforestation. However, for acountry like the current China, where fewfrontier forests are left, timber will becomemore scarce and worth managing. There-fore, a forest transition will be likely tooccur. However, there is insufficient justi-fication for active forest management to be


63

derived from traditional production eco-nomics. The transaction costs of landproperty rights consist of the costs of exante and ex post of transactions. All po-tential land rent will dissipate whenever theland is in open access. However, the po-tential gains from open access to closedaccess will also dissipate if the transactionis too costly. The transaction costs are im-portant factors for the transition fromnon-active to active forest management.Therefore, institutions are critical, sincetransaction costs greatly depends on them.

Based on the previous discussion, somepolicy implications regarding China�s fu-ture forest development can be outlined:

· The first priority is to emphasize the de-velopment of institutions to significantlyreduce property right protection costs forforest management and to provide moresecurity and confidence for farmers con-cerning the continuity of the propertyrights. This will accelerate the transitioninto active management in the area be-tween Points B and C in Figure 1, whichtoday accounts for 50% of total forestryland.

· The second priority is to reduce taxeson timber production and adjust the tax-ation in order to increase the currentforest management intensity by usingprice incentives. This would effect thearea between A and B and would encour-age the shift of Point B to the right.Lump sum tax could be a good alterna-tive.

· The third priority is to allocate govern-ment funds to protect those highlyfragile and environmentally significantforests, located to the right of Point C inFigure 1.

Currently, it seems that government effortsare still focusing on direct investmentthrough launching massive forestation pro-

grams. This may not be the most effectiveapproach. Without the institutions that cre-ate cheap exclusion costs, closed access isunfeasible. Without closed access, any in-vestment in silviculture is wasted money.If forest management is profitable, privateinvestment will be induced. The priority ofgovernment should be to invest in institu-tion building that would promote thedefinition, transference and protection ofproperty rights.

Acknowledgments

Thanks are extended to Jari Kuuluvainenof the University of Helsinki, Jussi Uusi-vuori of the Finnish Forest ResearchInstitute, William Hyde of the VirginiaPolytechnic Institute and State University(USA), Daowei Zhang of the Auburn Uni-versity (USA) and Chunjiang Liu of theUniversity of Helsinki for their comments.The Department of Forest Economics of theUniversity of Helsinki, the United NationsUniversity/World Institute of DevelopmentEconomics Research (UNU/WIDER) andthe International Development ResearchCentre/Economy and Environment Programfor Southeast Asia (IDRC/EEPSEA) provid-ed considerable support. All possible errorsare the responsibility of the author.

References

Alchian, A.A. and Demstz, H. 1972. Produc-tion, Information Costs, and EconomicOrganization. American Economic Review62: 777�795.

Barzel, Y. 1997. Economic Analysis of Proper-ty Rights. 2nd Edition. Cambridge UniversityPress, New York.

Bromley, D.W. 1991. Environment and Econo-my: Property Rights and Public Policy.Blackwell Publishers, Cambridge, USA.

Cheung, S.N.S. 1983. The Contractual Natureof the Firm. Journal of Law and Economics2(1): 1�44.

64

FOR

EST

TR

AN

SIT

ION China�s Statistical Bureau 1975�1996. China�s

Statistical Yearbooks. China�s Statistics Pub-lishing Houses, Beijing.

Chinese Society of Forestry 1987. China�s For-est Resource Trend in Year 2000. ScientiaSilvae Sinicae 1987(1): 51�59.

Chu, C. 1988. History on calculation of the for-est cover rate. Forest ResourcesManagement 1988 (3): 5�10.

Coase, R.H. 1937. The nature of the firm. Eco-nomica 16: 386�406.

Coase, R.H. 1960. The problem of social cost.Journal of Law and Economics 3: 1�40.

Dahlman, C.J. 1979. The problem of externali-ty. Journal of Law and Economics 22(1):141�162.

FAO 1978. China: forestry support for agri-culture. FAO Forestry Paper series 12. Rome.

FAO 1995a. Forest resources assessment 1990- Global synthesis. FAO Forestry Paper 124.Rome.

FAO 1995b. Forest resources assessment 1990- Non-tropical Developing Countries. Tech-nical Report FO: GCP/INT/474/FRA. FAO,Rome.

Gu, K. 1988. Structure and Simulating of theNational Forest Resources Prediction Mod-el. Journal of Beijing Forestry University1988(3): 57�65.

Handel-Mazetti, H. 1931. Die pflanzengeogra-phische Gliederung und Stellung Chinas.Botanische Jahrbücher 64: 309�23.

Heilongjiang General Bureau of Forestry 1987.Forestry Zoning in Heilongjiang. Harbin,China.

Hyde, W., Amacher, G.S. and Magrath, W. 1996.Deforestation, Scarce Forest Resource, andForest Land Use: Theory, Empirical Evi-dence, and Policy Implications. World BankResearch Observer 11: 223�248.

Hyde, W.F. and Seve, J.E. 1993. The economicrole of wood productions in tropical defor-estation: the severe example of Malawi.Forest Ecology and Management 57: 283�300.

Kuznets, S. 1966. Modern economic growth:rate, structure, and spread. Yale UniversityPress, New Haven.

Kuznets, S. 1973. Modern economic growth:findings and reflections, American EconomicReview 63: 247�258.

Lewis, W.A. 1955. Theory of Economic Growth.Harper & Row, London.

Liao, S. 1987. Introduction to Forestry Econom-ics. China�s Forestry Publishing House,Beijing.

Liu, C. and Wang, Z. 1989. China�s ForestChange Process, Structure and Patterns. Be-ijing Forestry Science and Technology1989(1): 41�48.

Liu, Z. and Sun, B. 1985. Forest resource ex-ploitation and management in ChangbaiMountain Area. China�s Forestry PublishingHouse, Beijing.

Mattews, R.C.O. 1986. The Economics of In-stitutions and Sources of Growth. EconomicJournal 96: 903�910.

Maxwell, N. (ed.) 1975. The Chinese Road toDevelopment. Pergamon Books Ltd.

Menzies, N.K. 1994. Forest and Land Manage-ment in Imperial China. St. Martin�s Press,New York.

Ministry of Agriculture 1950�1995. China�sAgricultural Statistical Yearbooks. China�sAgriculture Publishing Houses, Beijing.

Ministry of Forestry 1978, 1982, 1988, 1995.China�s Forest Resources Statistical Mate-rials. China�s Forestry Publishing House,Beijing.

Ministry of Forestry 1987�1997. China�s For-estry Statistical Materials. China�s ForestryPublishing House, Beijing.

Ministry of Forestry 1987. China�s Forest Zon-ing. China�s Forestry Publishing House,Beijing.

Ministry of Forestry 1992. Forest resourcesadministration and monitoring. China�s For-estry Publishing House, Beijing.

Murphey, R. 1983. Deforestation in ModernChina. In: P.T. Richard and J.F. Richard(eds.), Global Deforestation and the Nine-teenth Century World Economy. Duke PressPolicy Studies. New Haven, USA. Pp. 111�128.

North, D.C. 1981. Structure and Change inEconomic History. Norton, New York.

North, D.C. 1991. Institutions, InstitutionalChange and Economic Performance. Cam-bridge University Press, New York.

North, D.C. and Thomas, R.P. 1973. The Riseof the Western World. Cambridge Universi-ty Press, London.


65

Palo, M. 1994. From Deforestation into Sus-tainable Forestry � A transition based onmarkets and policies. Scandinavian ForestEconomics 35.

Palo, M. and Uusivuori, J. (eds.) 1999. WorldForests, Society and Environment. KluwerAcademic Publishers, The Netherlands.

Park, P.J., Barbier, E.B. and Burgess, J.C. 1998.The Economics of Forest Land Use in Tem-perate and Tropical Area. Environmental andResource Economics 11(3-4): 473�487.

Pen, C. 1922. Hainan Forests. Journal of Ge-ography 12: 8�15.

Perkins, D.H. 1969. Agricultural Developmentin China: 1368�1968. Aldine PublishingCompany, Chicago.

Poore, D., Burgess, P., Palmer, J., Rietbergen,S. and Synnott, T. 1989. No Timber WithoutTrees: Sustainability in the Tropical Forest.Earthscan Publications, London.

Richard, P.T and Richard, J.F. (eds.) 1983. Glo-bal Deforestation and the NineteenthCentury World Economy. Duke Press PolicyStudies. New Haven, USA.

Rudel, K.T. 1998. Is There a Forest Transition?Deforestation, Reforestation, and Develop-ment. Rural Sociology 63(4): 533�552.

Ruiz-Peres, M., Fu, M., Xie, J., Belcher, B.,Zhong, M. and Xie, C. 1996. Policy changein China: The effects on the bamboo sectorin Anji County. Journal of Forest Econom-ics 2(2) 149�176.

Shaw, N. 1914. Chinese Forest Trees and Tim-ber Supply. T. Fisher Unwin, London.

Song, Y., Burch, W., Geballe, G. and Geng, L.1997. New organizational strategy for man-aging the forests of southeast China � Theshare-holding integrated forestry tenure(SHIFT) system. Forest Ecology and Man-

agement 91:183�194.Stigler, G. 1966. The Theory of Price. Macmil-

lan Co, New York.von Thunen, J.H. 1875. The Isolated State in

Relation to Land Use and National Econo-my. Schmaucher Zarchlin, Berlin.

Wang, T. 1992. Quantitative economic analy-sis on the problem of decreasing cultivatedland in China. Economic Sciences Publish-ing House, Beijing.

Wang, Z. and Sheng, L. 1994. China�s ecolog-ical environment changes and populationpressure. China�s Environmental SciencesPublishing House, Beijing.

Wen, H. and He, Y. 1979. China�s Forest Re-source Distribution in History. Naturalresources 1979 (2): 72�84.

Westoby, J. 1975. Making Green the Mother-land-Forestry in China. In: N. Maxwell (ed.),The Chinese Road to Development. Perga-mon Books Ltd.

Williamson, O.E. 1975. Markets and Hierar-chies: Analysis and Antitrust Implications.The Free Press, New York.

Williamson, O.E. 1985. The Economic Institu-tions of Capitalism. The Free Press, NewYork.

Yin, R. and Newman, D. 1997. Impacts of ru-ral reforms: the case of the Chinese forestsector. Environment and Development Eco-nomics 2: 289�303.

Zhang, Y., Dai, G., Huang, H., Kong, F., Tian,Z., Wang, X. and Zhang, L. 1999. The For-est Sector in China: Towards a MarketEconomy. In: M. Palo and J. Uusivuori (eds),World Forests, Society and Environment.Kluwer Academic Publishers, The Nether-lands. Pp. 371�393.

deforestation and forest transition: theory and evidence ...zhangy3/deforestation and forest...

Documents