Search

This study explored the consequences for long-term projections and impact analysis of the uncertainty in model parameters and initial conditions. Using the Global Forest Products Model, multiple replications of projections were carried out with parameters or initial condition data sampled randomly from their assumed distribution. The results showed that parameter uncertainty led to uncertainty of the projections increasing steadily with the time horizon, and more rapidly than the uncertainty stemming from initial conditions. Among the parameter uncertainties, those in the supply and demand elasticities tended to dominate the uncertainty in the other parameters describing forest growth, manufacturing activities, and trade inertia. In an application to impact analysis it was found that, due only to the uncertainty of the model parameters, and conditional on other assumptions, an assumed rise in global temperature of 2.8 °C over a century caused the forest stock in 2065 to be 2.4% to 4.0% higher in developed countries, and 2.5% to 3.9% lower in developing countries, with 68% probability, a conservative estimate of the true uncertainty given all the other factors involved in such a prediction.

This study explored the consequences for the forest sector of a trade agreement between Australia, Canada, Chile, Brunei Darussalam, Japan, Malaysia, Mexico, New Zealand, Peru, Singapore, the United States, and Vietnam (TPP12), and of an extension to include India, China and the Republic of Korea (TPP15). The forecasts used a partial equilibrium model to predict the effects of reduced trade costs due to tariff elimination and of forest product demand shifts resulting from the macroeconomic impacts of TPP12 and TPP15 predicted by Narayanan and Sharma (http://wtocentre.iift.ac.in/workingpaper/workingpaper14.pdf). The results showed the effects of the agreements on the consumption, production, and trade of roundwood, sawnwood and panels, and pulp and paper, for the countries directly involved in the agreements, the European Union, and the rest of the world (ROW). Under TPP12, the main total welfare gains of wood producers and forest product consumers were in the United States and Vietnam, and the greatest losses in China and Korea. Global welfare gains were considerably larger under TPP15, to the advantage of China, Japan and Korea, but to the detriment of the United States, the European Union and the ROW.

The importance of international trade for the welfare of actors in the forest sector was estimated by comparing the current state of the world with a world in pure autarky with zero imports and exports of roundwood and manufactured wood products. The analysis was done with a comparative statics application of the Global Forest Products Model. The model was first calibrated to replicate observations in the base year 2013, and then solved under autarky conditions. The results showed much variation in the effects of international trade on production, consumption, and prices across countries and sub sectors. Globally international trade did have a positive effect on the economic welfare of the sector. This was due mostly to the positive effect on the surplus of consumers, and to a lesser extent on the increase in value added in forest industries. But value added profited manufacturers in developed countries much more than in developing. Furthermore, while wood producers in developed countries increased their profits with trade, those in developing countries incurred heavy losses that negated any incentive to invest in forest conservation, management and new plantations.

Long-term projections for the global forest sector are particularly sensitive to the parameters of the demand equations for the end products. To get more precise estimators, and test statistics with more power than with pure time-series, the elasticities of demand are typically estimated from panel data, pooling time-series across countries, and thus assuming that the elasticities are the same in all countries. The objective of this study was to recognize potential differences between countries, while using the prior information obtained by pooling. The proposed method estimated with quadratic programming country-specific elasticities that minimized the sum of squares of the errors across countries and over time, while keeping the country elasticities within the confidence intervals of the pooled elasticities. The method was applied to international data for seven product groups from 1992 to 2016. Compared with pooling, country-specific elasticities reduced the root mean square error of in-sample predictions by 7% to 43% depending on the product. The country-specific elasticities had smaller standard errors than the pooled elasticities, and they tended to cluster near the bounds of the confidence intervals of the pooled elasticities. With country-specific elasticities in a global sector model the projected world prices in 2065 were 3% to 13% higher, depending on the product, than with pooled elasticities. World consumption in 2065 with country-specific elasticities was from 2% lower to 42% higher depending on the product, with large differences across countries and product groups.

Most economic studies of forest decision making under risk assume a fixed interest rate. This paper investigated some implications of this stochastic nature of interest rates. Markov decision process (MDP) models, used previously to integrate stochastic stand growth and prices, can be extended to include variable interest rates as well. This method was applied to Douglas-fir/western hemlock forests in the Pacific Northwest of the United States. An MDP model was used to find the harvest decisions that maximized the forest value of a stand in a particular state, given the price level and interest rate. This optimal policy was compared with the policy that would hold in the same context but with an interest rate fixed at its historical average. The results showed that when the interest rate was lower than its historical average, the best decision differed for 52 of the 192 possible combinations of stand state and price level. Assuming a fixed interest rate underestimated the forest value by 4% to 16% depending on the initial condition. However, applying the harvest policy derived with a fixed interest rate led to a loss of no more than 7% depending on the initial condition. Taking explicit account of the variability of interest rate in setting harvest policies had some unexpected ecological benefits.

The Global Forest Products Model (GFPM) was modified to link the forest sector to two scenarios of the Intergovernmental Panel on Climate Change (IPCC), and to represent the utilization of fuelwood and industrial roundwood to produce biofuels. The scenarios examined were a subset of the “story lines” prepared by the IPCC. Each scenario has projections of population and gross domestic product. These projections were used as input in the GFPM simulations. The IPCC also makes projections of forest area, which were integrated in the timber supply sub-model of the GFPM. The IPCC scenarios also predict bioenergy production. These projections were used in the GFPM to determine forest area, forest stock, and the demand, supply, prices, and trade of forest products up to 2060. The main finding concerns the important impact of the high demand for biofuels implied in some of the IPCC scenarios. In particular, scenario A1B would induce a nearly 6-fold increase in the world demand for fuelwood by 2060. As a result, the real price of fuelwood would rise and converge towards the price of industrial roundwood by about 2025. At that point, industrial roundwood, which was used in the past to manufacture sawnwood, panels, and pulp, would begin to be used for energy production. The price of all wood would then continue to rise steadily up to 2060, and the price of manufactured product would increase in concert. The high fuelwood harvest would imply ecologically stressed forests in several countries, even under scenario A2 with a nearly 3-fold increase in fuelwood production by 2060.

ARTICLE I NFO We assessed the impact on the world forest sector of a progressive elimination of illegal logging. The analysis compared predictions from 2007 to 2020, with and without a gradual reduction of illegally logged industrial roundwood from 2007 to 2011. A large part of the curtailment of timber supply due to the stoppage of illegal logging would be compensated by increased legal production incited by higher prices. As a result, without illegal logging the world annual production of industrial roundwood would decrease by no more than 1%, even though it would decrease by up to 8% in developing countries. World prices would rise by 1.5 to 3.5% for industrial roundwood and by 0.5 to 2% for processed products, depending on the assumption on illegal logging rates. World consumer expenditures for wood products and producer revenues would rise by 1 to 2% without illegal logging. World value added in forest industries would remain the same. However, the changes in consumer expenditures would be more than double the changes in producer revenues in countries dependent on illegally logged timber of domestic or foreign origin such as Indonesia and China. Symmetrically, changes in producer revenues would be almost twice the changes in consumer expenditures in countries with little illegal logging and efficient industries, such as Canada, Germany and the United States. Value added in forest industries would decrease most in countries with heavy illegal logging (12% in Indonesia and up to 9% in Brazil), and it would increase most in Germany, Canada (4%), and the United States (2%). Without illegal logging, the world forest inventory would increase slightly, as the increase in developing countries would more than compensate the decrease in developed countries.

In the context of forest management, a fixed harvesting policy consists in trying to convert stands of trees to a chosen state at fixed intervals, regardless of the stand state and of the state of the market. In an adaptive policy, instead, the post-harvest state and the timing of the harvest depend on the stand and market states at the time of the decision. The objective of this study was to determine the practical gain from the theoretically superior adaptive policies. To this end, we compared optimal fixed and adaptive policies obtained with identical models and assumptions, and with data from the Douglas-fir/western-hemlock forests in the Pacific Northwest of the United States. In maximizing economic returns from harvests over an infinite time horizon, the net present value was 17 percent higher with an adaptive than with a fixed policy. It was 22 percent higher when the objective was to maximize annual harvest. The adaptive policy was even more superior with undiscounted, non-economic objectives, such as the area of spotted owl habitat (+37 percent gain), or the area of late-seral forest (+51 percent), but less so in maximizing the stock of high quality logs (+6 percent). The adaptive formulation also lent itself readily to multi-objective management.

The objective of this study was to determine the relationship, or lack thereof, between growth and diversity of tree species and size in conifer stands of western North America. Growth was measured by net basal area growth and its components: survivor growth, recruitment, and mortality. The analysis used inventory data from permanent plots in the Douglas-fir/western hemlock forest type in Oregon and Washington, and in the mixed-conifer forest type in California. The methods consisted of generalized least square regression with spatial autocorrelation, controlling for the effect of other stand characteristics. Other things being equal, in the two forest types under study there was a strong positive relationship between net basal area growth and tree-species diversity. This effect was associated with higher recruitment in stands of higher tree-species diversity. Neither mortality nor growth of survivors was related to tree-species diversity. The relationship between growth and tree-size diversity was less clear. For Douglas-fir/western hemlock, net basal area growth was negatively correlated with tree-size diversity, essentially because recruitment waslower onplots of high tree-size diversity.Formixed conifers, netbasal area growthtended alsoto be lower inplots of high tree-size diversity, but this was mostly because mortality was higher in plots of higher tree-size diversity. # 2007 Elsevier B.V. All rights reserved.

Background This paper explored the long-term, ceteris-paribus effects of potential CO 2 fertilization on the global forest sector. Based on the findings of Norby et al. (PNAS 2005, 102(50)) about forest response to elevated [CO 2 ]. Methods Forest productivity was increased in the Global Forest Products Model (GFPM) in proportion to the rising [CO 2 ] projected in the IPCC scenario A1B, A2, and B2. Projections of the forest area and forest stock and of the production, consumption, prices, and trade of products ranging from fuelwood to paper and paperboard were obtained with the GFPM for each scenario, with and without CO 2 fertilization beginning in 2011 and up to 2065. Results CO2 fertilization increased wood supply, leading to lower wood prices which in turn induced modest lower prices of end products and higher global consumption. However, production and value added in industries decreased in some regions due to the relative competitive advantages and to the varying regional effects of CO 2 fertilization. Conclusion The main effect of CO 2 fertilization was to raise the level of the world forest stock in 2065 by 9 to 10 % for scenarios A2 and B2 and by 20 % for scenario A1B. The rise in forest stock induced by fertilization was in part counteracted by its stimulation of the wood supply which resulted in lower wood prices and increased harvests.

This paper seeks ways of managing forests for landscape diversity, tree species and size diversities and timber production in a stochastic environment. A Markov chain model was developed to describe stand transition between pre-defined states, with high-frequency shocks and rare natural catastrophes. Stochastic optimization was then used with this model to study the trade-off between landscape diversity and other management objectives. This approach was applied to the mixed loblolly pine (Pinus taeda L.) and hardwood forests in the southern United States. The results showed that natural catastrophes enhanced the diversity of the landscape, but impaired the tree species and size diversities. Following current management would generate high landscape diversity, but low timber productivity. A highly diverse landscape could be maintained while keeping the tree species and size diversities near the achievable maximum. But managing to maximize tree size diversity or species diversity would much decrease landscape diversity. The opportunity cost of preserving a highly diverse landscape was high in terms of foregone timber production.

We analyze the willingness to pay for non-timber values (NTV) by non-industrial private forest (NIPF) landowners of uneven-age mixed species forest stands in the south central United States. The preferences of NIPF owners are revealed by the way they manage their timber. Many NIPF owners maintain a more diverse and natural stand structure than that of a more profitable even-aged industrial plantation. In so doing, the average NIPF owner was willing to forego 60% of the timber profit for the NTV of their more natural stands compared to a less diverse industrially managed even-aged plantation.

Transformation requires, in part, the setting of goals for the stand structure that should be ultimately achieved. This structure can be described by the number of trees per unit area of different sizes, and their species, before and after the harvest. The structure should be sustainable under the prescribed intensity of harvest and cutting cycle. In addition, the harvest and residual growing stock should be such as to meet management goals in the best possible way. These goals usually involve economic and ecological dimensions that can be quantified with financial criteria such as net present value, and with ecological criteria such as diversity indices. The core of quantification is a growth model suitable for the specific forest type under consideration. The growth model allows the determination of feasible, sustainable regimes, which are rarely unique. The choice of a specific regime can be guided by optimization methods, for example, to maximize economic returns, subject to ecological constraints. Once the desired state is decided upon, the conversion to this state can be achieved by harvesting only the trees in excess of the desired distribution. Alternatively, the conversion can also be optimized, for example, by calculating the regime that brings the stand as close as possible to the desired state, within a specific amount of time.

The objective of this study was to project the effects on the world forest sector of eliminating quickly all import tariffs. The projections were done for two scenarios: (1) progressive tariff reduction according to the current schedule of the General Agreement on Tariffs and Trade; (2) complete elimination of all tariffs within the Asia-Pacific Economic Cooperation (APEC) countries. The projections were obtained with the global forest products model (GFPM), for the years 1998–2010. The model gave market equilibrium projections of quantities produced, consumed, imported and exported for each of 180 countries, and for 14 commodity groups, covering roundwood, sawnwood, wood-based panels, pulp and recycled fibers and paper and paperboard. The model also projected world equilibrium prices. The results showed that, with or without ATL agreement, the world consumption and trade of all forest products would continue to grow along the historical trends, and the world prices of forest products would increase moderately. With the elimination of tariffs in APEC countries, the projected world production and consumption of all products would change little (less than 0.5%). The tariff reduction effects would be larger for trade, and the commodity composition of world trade would shift from raw materials to more processed products. The timber harvest would change in a number of countries, but the net effect at the world scale would be small. While the US consumption and production of forest products would change little, the US would reduce its export of logs and increase exports of most processed products. The US timber harvest was expected to be indistinguishable compared to what it would be in the absence of the Accelerated Tariff Liberalization agreement.

Data from a large-scale (over 8000) discrete choice contingent valuation study with a follow-up are used to investigate the effects of the hedonic equation for willingness to pay for access to forest recreation in Ireland. Selected forest descriptors for recreational quality show meaningful impacts and signs through a number of probit models, representing a gradient of restrictions gradually moving from the bivariate probit with heteroskedasticity to the interval data assumptions. Significance and magnitude of parameter estimates are found to be stable across the investigated models. This result supports the hypothesis that value estimates from contingent valuation methods exhibit theoretically consistent relationships with important forest attributes.

A partial equilibrium model was applied to the global forest sector in order to assess regional and global impacts of changes in economic growth, timber supply potentials, and technical trends. The model uses recursive price-endogenous linear programming and deals with eight geographical regions and 16 products. The base line projections of the model gave an average annual increase in global supply of industrial roundwood of 1.2% until the year 2010. The real price of sawlogs and sawnwood was found to remain approximately constant, whereas the prices of pulpwood and particles increased significantly during the first years, and then declined after the year 2000. The real prices of pulp and paper increased less than those of pulpwood and particles. The assumed variations in GDP growth rates had limited influence on quantities supplied and traded due to restricted timber supply potentials, but affected the real prices, especially of pulpwood and particles. Changes in the assumed timber supply potentials and technical change affected the real prices of pulpwood and particles significantly. Introduction of a price responsive timber supply also dampened the price peaks of pulpwood. Possible improvements of the methodology include empirical estimation of timber supply and of key parameters that determine capacity expansion, trade inertia, and technical changes.

Data from a large-scale contingent valuation study are used to investigate the effects of forest attributes on willingness to pay for forest recreation in Ireland. In particular, the presence of a nature reserve in the forest is found to significantly increase the visitors’ willingness to pay. A random utility model is used to estimate the welfare change associated with the creation of nature reserves in all the Irish forests currently without one. The yearly impact on visitors’ economic welfare of new nature reserves approaches half a million pounds per annum, exclusive of non-recreational values.

The objective of this study was to update and extend the geographic range of a forest growth model for northern hardwoods, developed previously with data from the fourth Wisconsin inventory (Lin et al., 1996. Ecol. Model., 91: 193-211.). To this end, Lin's model was recalibrated with data from the recent fifth inventory of Wisconsin and Michigan, and with the addition of a site variable to reflect variations in land productivity. After the introduction of site effects, there were still statistically significant differences between the equations of ingrowth, upgrowth. and mortality for Wisconsin and Michigan. Thus, two models were maintained. one for each state. Each model predicted well the growth of trees on post-sample plots, and simulated adequately the tree distribution in old stands in its own state. Applied to stands of the same initial distribution of tree species and size, the equations predicted faster early growth and higher basal area in the steady state in Michigan than in Wisconsin, with more marked differences on good than poor sites.

Uneven-aged silviculture in loblolly pine ( Pinus taeda L.) stands has many economic and ecological benefits. Here, the consequences of various uneven-aged management regimes are predicted with the SouthPro simulator. Results indicate that target distributions for pines with residual merchantable basal areas of ≈12.5 m 2 ha −1 , maximum diameters of ca. 40 cm, and q -ratios of 1.2–1.25 for 2.5 cm DBH classes are likely to provide high economic returns on good sites when combined with hardwood control. Increasing this maximum diameter would enhance tree-size diversity, but reduce sawtimber production and profits. Retaining a hardwood component with 1.15–2.3 m 2 ha −1 of basal area could enhance tree-species diversity, but this too would result in moderate reductions in income. Insisting on maximizing tree-size diversity or tree-species diversity among softwoods, soft hardwood, and hard hardwoods would be quite costly in terms of lost income and production. Results also illustrate how short-term economic incentives can lead to high-grading practices, despite substantial reductions in stand productivity and net returns in the long term.

A matrix growth model with constant parameters has been modified, by making survivor growth and mortality of trees functions of stand basal area. The functions, estimated from re-measured plots in Wisconsin maple-birch forests, showed that basal area affected mostly the growth of small trees. However, even when the effects were statistically significant, they were small. Fifteen-year forecasts with the variable-parameter model were almost identical to those of the fixed-parameter model for a set of 37 random plots. Long-term, steady-state predictions gave similar diameter distributions for both models, but a 20% lower basal area for the variable-parameter model. Steady-state management regimes that maximized soil rent were identical for both models in terms of intensity of harvest and length of cutting cycle. The basal area, harvests, and net present values predicted by both models, in the long run and during conversion, were also nearly the same. For managed maple-birch forests, the matrix model with constant parameters has the advantage of simplicity of form, estimation, and application, especially in optimization mode.

A density-dependent matrix model of stand growth is presented. The model was calibrated with re-measured plots in the northern hardwood forests in Wisconsin. Trees were divided into shade-tolerant, intermediate, and shade-intolerant species in twelve diameter classes. Model predictions were tested against post-sample observations for short-term (15-year) and long-term (steady-state) accuracy. This growth model was then applied to study stand growth dynamics with and without management. Starting with the current average stand state, the predicted stand basal area of an unmanaged stand rose steadily to a maximum, dropped gradually, and finally reached a steady state. The time paths of basal area and tree size diversity were highly and positively correlated. Instead, tree species diversity decreased when size diversity and basal area increased, and vice-versa. When harvesting was introduced, it was found that more intensive harvest increased the diversity of tree species, while decreasing the diversity of tree size. A 20-year cutting cycle could produce higher economic returns than a 10-year cutting cycle, with similar size and species diversity. Harvesting most of the saw timbers every 20 years could be a good compromise between economics and tree diversity, at the stand level.

The aim of this study was to predict the long-term effects of alternative management regimes for uneven-aged mixed-species forests in the Italian Dolomites, among the most beautiful in the Alps range. For this purpose, a non-linear matrix model of stand growth was estimated from re-measured stands, and the predictions compared with actual data. Then, long-term simulations were used to determine the effects of (i) extracting some mortality only; (ii) applying Susmel's (1980) management guides; (iii) cutting only trees above a diameter limit, (iv) saving the beech, and (v) continuing the current harvest rate. The ecological criteria to evaluate the different management regimes were basal area and stand composition by tree species and size. The economic criterion was the net present value of the harvests. The results suggest that, for a 10-year cutting cycle, applying Susmel's guide or a diameter-limit cut of 70 cm for conifers and 50 for beech would keep diversity of species and size equal to that of an unmanaged stand, though lowering basal area. Of the two, Susmel's guide would give 25% higher income. However, income could be doubled by lowering the diameter-limit cut to 60 cm for conifers and 45 for beech. This would give the same basal area and species diversity, but decrease size diversity by about 6%. The policy that did not allow to cut beech resulted in stands that had less beech than those obtained by Susmel's guide or diameter-limit cuts. Perpetuating the current cutting rate gave the highest tree size diversity, but lower species diversity and the lowest income. Of cutting cycles of 5, 10, and 15 years, the longest gave, other things being equal, the highest basal area and size diversity, slightly lower species diversity, but substantially lower income.

Quantitative analysis and forecasting of forest product markets began in the 1950s, based almost exclusively on pure time-series analysis. Since then, considerable improvements have been made in the theoretical basis of the models, the statistical methods of estimation, and the coverage of the data. Especially noteworthy is the exploitation of panel data in the analysis of demand for final products. To model the industry supply of intermediate and final products, however, activity analysis seems more promising than econometrics. It allows for a detailed description of the techniques of production, and, therefore, a better modeling of technical change than the econometric approach. This paper presents a class of models, currently used for policy analysis and forecasting, that present a blend of econometric and mathematical programming, with a dash of system dynamics. Econometrically estimated demand and supply functions are combined with activity analysis of production and transportation, into optimizers to calculate spatial equilibria in multi-product markets. The shadow prices of the optimization are then the key inputs in predicting capacity change by region and process, in system dynamics fashion, and based on Tobin's q theory of investments.

A growth model is presented, which recognizes differences in tree species, quality and size. The model was calibrated with data from permanent sample plots in hardwood forests in Wisconsin, classifying trees in twelve sizes, each divided into high-value, low-value and non-commercial trees. The model predicted well the short- and long-term (steady-states) growth of stands, in terms of number, size and value of trees. This growth model was incorporated into a linear program to compare six cutting guides in terms of the soil rent and ecological diversity they would obtain in a steady state. A simple guide, cutting all trees of at least 41 cm every 15 years, led to a diversity that was nearly 95% of that of a natural stand, and a soil rent that was about 70% of the maximum achievable. High-grading the stand led to a negative soil rent and the lowest diversity among all six alternatives. These long-term, steady-state differences between cutting guides would be missed by analyses that relied on the forest value criterion, because forest value is determined largely by the initial stand state.

The objectives of this paper were to define, measure and explain differences in the comparative advantage of countries which trade forest products. The theory of Heckscher, Ohlin and Vanek was used, which led to a direct relationship between a country's net trade, its endownment in resources, and its domestic income. From these equations an index of revealed comparative advantage was derived and used to rank countries in the early 1980s for major groups of forest products. This ranking led to country groupings which suggested a definite relationship between the comparative advantage of a country, its endowment in forest and other resources, and domestic demand. This relationship was tested formally with an empirical model based on the Heckscher-Ohlin-Vanek equations. This simple model explained a large part of the variation in the net trade of total forest product, roundwood, sawnwood, panels, pulp, and paper, in the early 1960s and 1980s. In general, there was a strong positive relationship between net trade and the level of domestic demand, reflected by income. For some commodities, especially logs, this relationship had become stronger between the early 1960s and 1980s.

A theoretical model was developed to determine the best cutting policy (cutting cycle and residual stock) in a regulated selection forest. This model was obtained by a simple modification of the objective function in a linear-programming model of a non-regulated stand. The parallel between the determination of the best rotation in an even-aged regulated forest and that of the cutting cycle in an uneven-aged regulated forest was apparent. This conceptual model could be applied to a wide variety of situations, including the selective management of natural tropical forests. In this paper, the model was solved with data for northern hardwood forests in the United States. The results showed that regulating an uneven-aged forest to get a yearly production had a cost, compared to cutting the entire forest at intervals equal to the cutting cycle. The cost of regulation increased with the fixed cost (cost per ha of preparing and administering a harvest). The results also showed that the choice of a cutting cycle in uneven-aged forests was critical only when fixed costs were low. Particular attention was given to the loss that would occur by computing the best cutting cycle and residual stock of a regulated forest as if it were not regulated. The results showed that this loss was small, except for extreme interest rates and fixed costs. For all practical purposes, the best cutting cycle and residual stock could be computed for a regulated forest while ignoring the cost of regulation.

We investigate the reliability of transferring benefit estimates of forest recreation obtained from discrete choice CV data and conditional on forest-specific attributes. The transfer reliability is checked against the forest-specific estimates of mean and median willingness to pay. We report and discuss the outcomes of formal tests of the null hypothesis of no difference for 26 recreational forests in Ireland, when the value transfer is based on single and double-bounded data collected at the remaining 25 forests. Contrary to the unconditional value transfers of Downing and Ozuna [6] we find that value transfers conditional on site-specific recreational attributes are mostly transferable.

Four time series of monthly prices of oak and softwood lumber and flooring are analysed using the autoregressive integrated moving average technique. The principles of the method are exposed and one model is estimated for each time series. The steps in model formulation, estimation and diagnostic checking are described in detail for the series of oak lumber prices. All series are stationary after a first order differencing. The behaviour of monthly prices of softwood products is well described by pure autoregressive models while the oak product prices require full autoregressive integrated moving average models. The models obtained generate one-month forecasts which are significantly more accurate than those obtained from a no-change extrapolarive model, except in the case of softwood flooring prices for which the behaviour is close to a random walk. Six-month forecasts are computed for each series with corresponding confidence intervals. An adaptive forecasting procedure is used to revise multiple-month forecasts after the errors on the first month forecast have been observed.

A gravity model was used (1) to synthesise the system of economic—and, to some extent, political—influences which determined the multilateral flow of tropical logs during the period 1967 to 1973 and (2) to evaluate the changes in the relative importance of these factors over time. Shipping distance was found to be the most important determinant of trade intensity, with lesser but significant roles being played by size of the importing country, relative price in the importing and exporting country, level of development and wood availability of the exporter and total volume of trade between countries. Trade preferences corresponding to old colonial ties played a minor, although statistically significant, role in determining the volume of log trade. The influence of most variables influencing trade was found to be stable over the period of observation, except for a systematic shift in the role of per capita income, reflecting a rapid movement of exporting countries towards domestic processing of logs.

A model has been developed to evaluate long-term development strategies for the forestry sector of Indonesia. The model uses data for 26 regions of Indonesia, plus Japan and Singapore. The model consists of a mixed-integer program with goal programming variables. The objective function measures total costs which include harvesting, transportation, manufacturing, capacity expansion and cost of not meeting regional demand and export targets. The model variables represent the different movements of forest products at various stages of processing within and between regions. Goal variables measure deviations from domestic demand targets in each region and from export targets. Boolean variables are used to determine the economic location of one or two ports capable of handling Indonesia's exports to Japan. An application of the model is given to show the effect, keeping other variables equal, of altering the composition of Indonesian exports from logs only to 50% processed products and 50% logs.

The latest of three models designed to analyze long-term development plans of the forestry sector of Indonesia is presented. The planning problem consists in determining the geographic pattern of forest exploitation, industrial processing, and transportation which minimizes total costs. Costs include direct costs of various operations, and indirect costs of not satisfying domestic demand and of not fully using potential exports. The model has a goal programming structure. Goal variables measure deviations from demand and export targets, or industrial and port capacity expansion at various points. Other variables relate to the transport of logs, sawnwood and plywood, and fuelwood. Cost of port capacity expansion is a non-linear function of capacity, modeled by separable programming.

The purpose of this paper was to evaluate various techniques to obtain estimates of coefficients in a model of a selection forest developed previously. The data originated from 161 plots re-measured in 1961, 1964 and 1969. The detailed individual tree data available ib this particular case permitted a very simple and unbiased estimation of the growth matrix parameters. It was shown that, in situations where less information would be available, ordinary least squares, or seemingly unrelated (Zellner) estimation would lead to very serious bias and in fact meaningless results. A constrained estimation procedure was proposed to alleviate this bias. The method used the particular structure of the model and extraneous estimates of mortality rates.

A theoretical model of the United States demand for hardwood plywood imports was developed. Demand derives from activity in the construction industry where costs are minimised subject to a generalised Cobb-Douglas technology. This static model was then modified to reflect dynamic adjustments of imports to prices and housing starts. The resulting model was estimated with monthly data for the period January, 1974 to December, 1979. The choice of either c.i.f., f.a.s. or customs data did not significantly affect the statistical results. No significant substitution was apparent between imported hardwood plywood and domestic hardwood or softwood plywood. Import data appeared to be explained by expectations on housing starts and relative prices. Importers' expectations were represented by Cagan's adaptive expectations model and by Almon's polynomial lag equation. The latter model, relating imports to (i) housing starts and (ii) import prices, gave significantly better results. The maximum lag was of 9 months, with a mean lag of 4·5 months. The long-term elasticity of imports with respect to housing starts was estimated at 1·10 (±0·12). The long-term elasticity with respect to real price was −1·98 (±0·24), using the price of all commodities as the deflator.

This paper deals with the problem of forecasting the timber harvest in a small region, given predicted changes for the economy of the country to which the region belongs. It suggests that time-series methods be used to model the feedback between local timber harvest and non-local (macro-economic) variables. These measures of feedback could then be supplemented with multiplier analysis to predict the long-term effect of each macro-economic variable of interest on local timber harvest. This method was applied to five counties of Southwestern Oregon, an area of some 3 m ha, 75% forested, to measure the effects of housing starts and lumber prices in the entire United States on private and public timber harvest in each county. Local public harvest was found to have short-term elasticities with respect to national housing starts of 0·48 (±0·18) to 0·78 (±0·19). The corresponding long-term elasticities ranged from 0·78 (±0·18) to 0·99 (±0·21). National lumber prices influenced local public harvest in the short-term only, with an elasticity of 1·34 (±0·36). Local private harvest was not influenced significantly by national housing starts, nor by prices.