1. THE GLOBAL FOREST PRODUCTS MODEL

1.1 Introduction

The Food and Agriculture Organization of the United Nations (FAO) has been producing long-term projections of future consumption, production and trade of forest products since 1982. The production of good projections has been a major concern to the FAO forestry economics and statistics staff and to all users, since they are likely to influence government planning and private decision making in member countries.

Earlier evaluation of the current FAO forest products projection system showed that when the system was applied to project consumption and production, it led to systematic errors for some products, and the projections towards the year 2010 also seemed to be too high. Furthermore, the projections of world consumption and production did not agree which, of course, they must in practice (Zhang and Buongiorno 1994).

To address these problems, a study was initiated in 1996 to improve the FAO global forest products outlook projection system. The approach used was a dynamic spatial equilibrium model to simulate the world forest sector. The model was applied to project forest product consumption, production and trade under provisional scenarios of economic growth and timber supply. A key feature of the model is that the projected demand for final products in the world cannot exceed the global wood supply, after allowing for trade in raw materials or finished products between countries. Therefore, the assumptions regarding economic growth and timber supply are critical in the projections. Three preliminary scenarios were considered in the Provisional Outlook (FAO, 1997). This report presents new projections based on revised economic growth rates, new information on timber supply, and revised econometric parameters in the GFPM model.

The report is organized as follows. Chapter 1 reviews the methodology of the Global Forest Products Model (GFPM), describes the countries/regions and the products included in the model, and discusses the demand equations and the timber supply assumptions. Chapter 2 presents projections for 1995 to 2010 of consumption, production and trade for the main regions of the world. Projections of forest products consumption, production and trade by country and product, from 1980 to 2010 are given in four appendices at the end of this report.

1.2 Background to the Global Forest Products Model

The GFPM model was developed with the Price Endogenous Linear Programming System (PELPS III) (Zhang et al. 1993). The technique simulates market equilibrium by mathematical programming. There have been several applications of the technique over the last two decades. The theory is that of spatial equilibrium in competitive markets. PELPS III solves the equilibrium by maximizing the value of the products, minus the cost of production, subject to material balance and capacity constraints in each country and each year. Because material flows throughout the system must balance, it insures data consistency, and coherence of projections.

In each projection year, for each country and commodity, supply (domestic production plus imports) is equal to demand (final consumption, plus input in other processes, plus exports). Final demand is price responsive; demand for wood or intermediate products derives from the demand for final products through input-output coefficients that describe technologies in each country. The supply of raw wood in each country is price responsive. The supply of recycled paper is constrained by the waste paper supply, which itself depends on the paper consumption and the recycling rate. Imports and exports, by country and destination are balanced in each country. Projected prices are such that they clear markets: at those prices demands are equal to supplies in each country.

From one year to the next, demand changes in each country due to changes in income. The wood supply shifts according to the chosen scenario. Technology also changes, especially the amount of recycled fibre used in making paper and paperboard. Capacity increases or decreases according to new investments that depend on past production and the profitability of production in different countries, as revealed by the shadow price of capacity. Then, a new equilibrium is computed subject to the new demand and supply conditions, new technology, new capacity, and inertia constraints that limit the change in exports and imports from year to year. The general principle of the PELPS system is, then, those markets optimize the allocation of resources in the short run (within one year). Long run resource allocation is partly governed by market forces, as in capacity expansion and trade, and also by political forces such as: the wood supply shifts determined by forest policy, the waste paper recovery rates by environmental policy, the trade inertia that depend in part on international agreements, and the techniques of production determined by exogenous progress.

PELPS has been used by the United States of America and Canadian Forest Services to develop the North American pulp and paper model (NAPAP), the North American solid wood model (NASAW), and by the International Tropical Timber Organization (ITTO) to develop the Asia-Pacific Tropical Timber Trade (Market) model. The Asia-Pacific Forest Products Model (Zhang et al. 1997) was also built with PELPS.

1.3 Model structure

The Global Forest Products Model has a hierarchical structure. It consists of one main model called WORLD which is solved first, and of four sub-models: AFRICA, AMERICA, ASIA, and EUROPE, each one solved with trade flows that are consistent with the WORLD model.

The WORLD model has four demand and supply regions, Africa, America, Asia, and Europe (including the former USSR). Each of the four regions has: (1) demand (consumption) of fuelwood, other industrial roundwood, sawnwood, veneer & plywood, particle board, fibreboard, newsprint, printing & writing paper and other paper & paperboard; (2) production (supply) of fuelwood, industrial roundwood, other fibre pulp, waste paper, sawnwood, veneer & plywood, particle board, fibreboard, mechanical pulp, chemical pulp, newsprint, printing & writing paper and other paper & paperboard. Also, each region imports from and exports to the "world market" (see Figure 1.1).

Figure 1. 1 Structure of the WORLD model.

A similar structure applies to the regional sub-models. For example, in the AMERICA sub-model, there are 35 countries, each with a demand and supply of the same products as in the WORLD model. Each country has imports from and exports to the "world market". However, the total imports and exports of all countries in, say, the AMERICA are constrained to be equal to the total imports and exports of the AMERICA estimated in the WORLD model.

1.3.1 Countries

The countries included in each sub-model are in Table 1.1. The AFRICA sub-model has 50 countries in Africa. The AMERICA sub-model includes 35 countries in North, Central and South America. The ASIA sub-model is composed of 50 countries in Asia and Oceania. The EUROPE sub-model has 45 countries in western and eastern Europe and the former Soviet Union.

Table 1.1 Countries in the Global Forest Products Model¹.

Code In Model	Country	Code In Model	Country	Code In Model	Country	Code In Model	Country
	AFRICA		AMERICA		ASIA		EUROPE
1	Algeria	1	Bahamas	1	Afghanistan	1	Albania
2	Angola	2	Barbados	2	Bahrain	2	Austria
3	Benin	3	Belize	3	Bangladesh	3	Belgium-Luxembourg
4	Botswana	4	Canada	4	Bhutan	4	Bosnia-Herzegovina
5	Burkina Faso	5	Cayman Islands	5	Brunei Darussalam	5	Bulgaria
6	Burundi	6	Costa Rica	6	Cambodia	6	Croatia
7	Cameroon	7	Cuba	7	China, People's Rep. of	7	Czech Republic
8	Cape Verde	8	Dominica	8	Cyprus	8	Denmark
9	Central African Republic	9	Dominican Republic	9	Hong Kong SAR, China	9	Finland
10	Chad	10	El Salvador	10	India	10	France
11	Congo, Rep of	11	Guatemala	11	Indonesia	11	Germany
12	Cote d'Ivoire	12	Haiti	12	Iran, Islamic Republic	12	Greece
13	Djibouti	13	Honduras	13	Iraq	13	Hungary
14	Egypt	14	Jamaica	14	Israel	14	Iceland
15	Equatorial Guinea	15	Martinique	15	Japan	15	Ireland
16	Ethiopia	16	Mexico	16	Jordan	16	Italy
17	Gabon	17	Netherlands Antilles	17	Korea, DPR	17	Macedonia, Fmr Yug RP of
18	Gambia	18	Nicaragua	18	Korea, REP	18	Malta
19	Ghana	19	Panama	19	Kuwait	19	Netherlands
20	Guinea	20	Saint Vincent	20	Laos	20	Norway
21	Guinea-Bissau	21	Trinidad and Tobago	21	Lebanon	21	Poland
22	Kenya	22	United States of America	22	Macau	22	Portugal
23	Lesotho	23	Argentina	23	Malaysia	23	Romania
24	Liberia	24	Bolivia	24	Mongolia	24	Slovakia
25	Libyan Arab Jamahiriya	25	Brazil	25	Myanmar	25	Slovenia
26	Madagascar	26	Chile	26	Nepal	26	Spain
27	Malawi	27	Colombia	27	Oman	27	Sweden
28	Mali	28	Ecuador	28	Pakistan	28	Switzerland
29	Mauritania	29	French Guiana	29	Philippines	29	United Kingdom
30	Mauritius	30	Guyana	30	Qatar	30	Yugoslav Fed. Rep.
31	Morocco	31	Paraguay	31	Saudi Arabia	31	Armenia
32	Mozambique	32	Peru	32	Singapore	32	Azerbaijan
33	Niger	33	Suriname	33	Sri Lanka	33	Belarus
34	Nigeria	34	Uruguay	34	Syrian Arab Republic	34	Estonia
35	Reunion	35	Venezuela	35	Thailand	35	Georgia
36	Rwanda			36	Turkey	36	Kazakhstan
37	Sao Tome and Principe			37	United Arab Emirates	37	Kyrgyzstan
38	Senegal			38	Viet Nam	38	Latvia
39	Sierra Leone			39	Yemen	39	Lithuania
40	Somalia			40	Australia	40	Moldova, Rep.
41	South Africa			41	Cook Island	41	Russian Federation
42	Sudan			42	Fiji	42	Tajikistan
43	Swaziland			43	French Polynesia	43	Turkmenistan
44	Tanzania, United Republic			44	New Caledonia	44	Ukraine
45	Togo			45	New Zealand	45	Uzbekistan
46	Tunisia			46	Papua New Guinea
47	Uganda			47	Samoa
48	Congo, Dem Rep. of			48	Solomon Islands
49	Zambia			49	Tonga
50	Zimbabwe			50	Vanuatu

1.3.2 Products

The products considered in the model are shown in Table 1.2. There are econometric demand equations for: fuelwood & charcoal; other industrial roundwood; sawnwood; veneer sheets & plywood; particle board; fibreboard; newsprint; printing & writing paper; and other paper & paperboard. Figure 1.2 shows the wood and noon-wood material flows assumed in the model.

Table 1.2 Products included in the Global Forest Products Model.

Code

Products

Units

Demand

Supply

Fuelwood and Charcoal

Sawlogs and Pulpwood

Other Industrial Roundwood

Veneer/Plywood

Particle Board

Fibreboard

Mechanical Wood Pulp

Chemical/Semi-Chemical Wood Pulp

Other Fibre Pulp

Waste Paper

Newsprint

Other Paper and Paperboard

1000 CUM

1000 MT

Note: E indicates the relation is represented with an econometric equation; I indicates the relation is represented with input-output coefficients.

1.4 Demand

The demand equations of the Global Forest Products Model assume that national demand for forest products is determined by economic activity (GDP) and real prices. In the short-run, demand may adjust only partially to GDP and prices. Let q_it (of price p_it) be the amount of a forest product consumed or imported by a country (or region) i during year t; with i_it an index of the price of all other goods and services used in combination with this product to produce an output designated by the variable y_it, which in this study was measured by the real gross domestic product. Then, assuming cost minimization under a Cobb-Douglas technology with constant returns to scale, coupled with a partial adjustment of actual consumption (or imports) towards desired imports, leads to the following derived demand (Chou and Buongiorno 1984, Baudin and Lundberg 1987):

The long-run elasticities, _ ^` and _^', used to make projections, were calculated from the short ones by the equations:

The demand equation for each product was estimated for three income groups: HIGH (developed market economies), LOW (developing market economies), and WORLD (combination of developed and developing economies). The HIGH elasticities were used for developed countries or regions. The generally larger LOW elasticities were used for Africa, and the developing countries in Asia.

Figure 1.2 Wood/non-wood material flows.

Table 1.3 Country groupings.

Group	Country
HIGH income	Australia, Austria, Belgium, Canada, Denmark, Finland, France, French Greenland, Germany, Iceland, Ireland, Israel, Italy, Japan, Kuwait, Luxembourg, Netherlands, Norway, New Zealand, South Africa, Spain, Sweden, Switzerland, United Kingdom, United States of America
LOW income	The rest of the world

Source: FAO (1994, p.ix)

Commodity prices were weighted averages of unit values of imports and exports, by country. The data on production, imports and exports were obtained from FAOSTAT·PC (FAO 1997). GDPs were measured in constant US dollar at 1987 prices, based on GDP in domestic currency, exchange rates, and the US GDP deflator. Alternatively, they were converted in constant US dollars with the domestic deflator, and at the 1987 exchange rate. The two methods gave similar estimates of elasticities for high-income countries, but quite different elasticities for low income countries. The elasticities used here pooled both types of GDP estimates. The data on GDPs, exchange rates, and the US GDP deflator came from the World Tables of the World Bank (1995).

The parameters of demand equations were estimated by analysis of covariance (OLS with one dummy variable for each country). The results in Table 1.4 show that the equations explained consumption well across countries and over time. GDP elasticities were all positive, except for fuelwood and other industrial roundwood in high-income countries, as expected. Price elasticities were low, though statistically and economically significant in many cases.

In the GFPM model, prices are endogenous (the equilibrium between demand and supply leads to the prices). So, no assumption has to be made regarding future prices. Only national GDPs shift demand and are strictly exogenous.

The assumptions about GDP growth in each country, and the implications for the regions are summarized in Table1.5. Real GDP was assumed to grow at about 4% per year from 1994 to 2010 in Africa, at 2.7% per year in America, 4.6% per year in Asia, and 2.4% per year in Europe.

1 Region names are those used in the model.