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 environmental goods 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 deforestation to forest transition may be prolonged by the time lag in forest regeneration, and by the transaction costs, i.e. the costs in transferring, defining and protecting property rights of land and forests. The institutional 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 compared with 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. This article concludes with a preliminary forecast of the future trend in Chinas forests and policy implications on future forest development. Keywords: Deforestation; forest transition; reforestation; transaction costs; economic reform; land use; population; China. 1 Introduction an increasing population for agricultural land and for timber are widely recognized While the forests in the developed coun- as the most important causes of deforesta- tries have ceased to shrink in area and have tion. However, the scarcity of timber, and even begun to expand a reversal which industrialization are generally viewed as 41 in this article is called forest transition accounting for forest transition (Rudel the developing countries are still in the de- 1998). Both conclusions seem to conflict forestation stage. The growing demands of each other but are likely to be correct. The WORLD FORESTS FROM DEFORESTATION TO TRANSITION? M. Palo and H. Vanhanen (eds.), World Forests from Deforestation to Transition? © 2000 Kluwer Academic Publishers, Dordrecht. problem is that we do not identify the ori- economic environment. Land quality is a gins of forests and analyze them separately broad concept here, referring not only to based on the origins. Forests can be either soil, steepness, incidence of rocks, sources natural endowment, or social products, but of water and exposure to wind and sun, but they have different responses to socio-eco- also to distance and types of roads to mar- nomic environments. The net change in ket and habitation, etc. The socio-economic forests is determined by the total effects. environment refers to the output and input Joining in the current debate and con- prices, and institutions. The input and out- FOREST TRANSITION cerns regarding deforestation in developing put prices can be measured on-site and in a countries, an attempt is made to formulate market. Prices may vary greatly when trans- an integrated theoretical framework of the portation costs are significant. deforestation and forest transition and to provide empirical evidence from China. 2.1 Land Use for Agriculture versus The forecasts and policy suggestions to be presented are argued to be relevant not only Forestry to China but also to other countries. The land rent for forestry or any other land use is determined by the output prices, the input costs and the land quality. For sim- 2 Framework of Deforesta- plicity, agriculture is considered as the only tion and Forest Transition other land use option. If we use (pa ,wa) and (pf ,wf) to represent the set of output price The theoretical framework of forest land and input cost by agriculture and forestry change, which is extended from the model respectively, the land rent for agriculture by von Thunen (1875), is illustrated by Fig- and forestry with land quality q will be ure 1. Land is used for the option which pa(pa,wa,q) and pf(pf ,wf ,q). If pa(pa,wa,q) creates the highest land rent in a competi- >pf(pf ,w,q), the land will be allocated to ag- tive land market. Therefore, the land use ricultural use. Otherwise, the land will be option and forest management method are used for forestry. For instance, the land to determined by the land quality and socio- the left of A in Figure 1 should be allocat- 42 Figure 1: Land use for agriculture and forestry by land quality ed to agricultural use, and the land to the It must be emphasized that land conver- right of A should be used for forestry. On sion is not costless and the cost is often the boundary, the rents for agriculture and sunk. In reality, land use conversion forestry are equal. becomes much less flexible when the cost Point A in Figure 1 shifts when on-site associated with conversion is considered, output or input prices for agriculture and while land modification, referring to forestry and land quality change. For sim- change within one category, e.g. among dif- plification, we just consider the changes in ferent agricultural crops, is quite flexible. input and output prices and do not consid- Land use conversion from forestry to agri- er land quality change that may derive from culture is more likely to happen when trees road expansion, soil degradation, etc. are grown to (or quite close to) an econom- Graphically, if prices for agricultural goods ically mature age. The capital in silvicul- rise (or input price decreases), the curve of ture is sunk and cannot be recovered (even the agricultural rent will shift to the right, partly) if the trees are too young. Land con- and some of the forestry land will be de- version from forestry to agriculture requires voted to agriculture. Alternatively, if the significant investment. It may take years to prices of forest products rise (or its input cut down the trees, remove stumps and price decreases), some of agricultural land stones, erect buildings and fences, to will be devoted to forestry. develop large-scale irrigation and drainage Mathematically, to understand the effects system, and to expand the initial clearings of the change in forest product prices on into productive farms. Most of such invest- the shift of the forest-agriculture frontier, ment is also sunk. If the expected outcome the total derivative of the equation (the prices and costs) does not differ great- pa(pa , wa , q*) pf(pf , wf , q*) = 0 (q* refers ly from the reality, the land is likely to re- to the boundary land where rent created main in agricultural use. from forestry and agriculture is equal) with Forest land conversion to agriculture and respective to q* and pf is: (¶pa/¶q*) × dq* other uses might be significant during the (¶pf /¶q*) × dq* (¶pf /¶pf) ×dpf = 0. early stage of economic development, but Therefore, (dq*/dpf) = (¶pf /¶pf) / [(¶pa/¶q*) it would not be critical with respect to the (¶pf /¶q*)]. As long as the higher quality total land area available for forestry. The land is used for agriculture, [(¶pa/¶q*) more critical problems are: (1) the stagnant (¶pf /¶q*)] will be positive, so (dq*/dpf) is timber price, poor credit markets, risk aver- also positive since (¶pf /¶pf) is positive. The sion, and poorly developed institutions do new equilibrium point q* will therefore not justify forest investment for the logged- move to better land quality q** when pric- over land; (2) forestry may be unprofitable es of forest products rise, thus some of the on a quite large area abandoned by other higher quality land currently used for agri- land uses due to soil degradation and other culture will be devoted to forestry (q** socio-economic reasons. That is why de- > q*). By the same procedure, it is possi- forestation is much faster than the ble to derive the land use shift between expansion of agricultural land in most de- forestry and agriculture in response to the veloping countries. changes of agricultural products and the inputs in forestry and agriculture as 2.2 Management of Forest Land 43 Forestry land = Af (pa ,wa , pf ,wf ) [1] FAO (1995a) defined forest land under + + active management as forest and other WORLD FORESTS FROM DEFORESTATION TO TRANSITION? woody land that is managed according to a purposes: if forest does not exist, trees will professionally prepared plan or is otherwise be planted; if the forest is already existing, under a recognized form of management logging will not be undertaken, e.g. natu- applied regularly over a long period (five ral reserves and public parks. However, a years or more). Land will be under active more common solution could not be 7 , forest management if 0< pf(pf ,wf ,q) and i.e. joint management of timber produc- pf(pf ,wf ,q) > pa(pa ,wa ,q). In Figure 1, only tion and environmental services could be the land to the left of B is under active for- more common in active forest manage- FOREST TRANSITION est management. ment regimes. Clear cutting is gradually The amount of investment and the length regulated, while its alternative, selective of rotation, the decision variables in forest cutting, is more accepted due to apprecia- management, are functions of the land qual- tion of environmental value. It means the ity, output and input prices. To calculate the new rota- tion always start from a certain optimal investment and rotation, the Faust- age of trees. mann model can be applied 7 Non-Active Forest Management PD[ × UW Ã pI T 7( U{ J W(YT H GW [2] The unexploitable forests, which are locat- U7 U7 S 4T 4 7(SZT H ±Z T (}[ ±H ] ed to the right of D in Figure 1, means that the stumpage price is zero, p(Q, q) = 0.