Industrial supply chain network
The biochemical process at the source of forest products is photosynthesis. In a sense, forests are a source of renewable solar based material for the industry. To il-lustrate imagine a typical forest tree transformed into diﬀerent products by diﬀerent sub-sector of the industry. The lower part of the trunk is transformed in sawnwood. The upper part is transformed in pulp for paper or in wood panels. The remaining branches are used for wood energy purposes. Many diﬀerent variations of this sce-nario are possible, the whole trunk could be used for a single purpose and branches could be left in the forest to maintain litter. This description could be refined almost indefinitely by talking about the multiple economic uses of all available tree species. But the main transformation remains from a natural resource to primary indus-trial products to final consumer goods. It is our goal to understand how consumer behaviour changes and how these changes can impact the forest sector. When con-sumers decide to purchase furniture, books or houses, their choices and preferences impact forest products consumption. But the variety of products involved make it impossible to study final consumer behaviour at the macroeconomic level. This is why I analyse the aggregated consumption of primary industrial forest products.
Forest trees are growing by a certain volume each year called the annual increment (on the order of 1.3 billion m3 yearly in the EU1). From this volume, 700 million m³ stay in the forest and the rest is consumed by 3 industrial sectors: 300 million m³ are transformed by the wood products industry, 100 Mm³ transformed by the paper industry and 200 Mm³ are burned to generate energy. Wood products flows are set into motion by final products consumers. While each sector requires diﬀerent wood material qualities, there is considerable overlap and complementarity in the sourcing process of each industry (see also 1.1.2). Complementarity is evident in the case of sawmills which process a large part of the yellow flow representing the wood products industry. Round wood harvesting is a costly operation. It is generally less costly to re-use industrial residues as illustrated by the orange flow (Figure 1.1) going from the wood working industry to other industries. Reuse of co-products and recycling of used products greatly increase the material eﬃciency of the forest sector. Over time as recycling rates progress, more products can be created from the same amount of raw material.
Interesting as it may be from an engineering perspective, a picture of physical flows alone lacks a critical piece of information, namely prices. Indeed at each market stage, a producer and a consumer exchange a good only if they both agree on its price. In other words as long as a consumer is willing to pay more than what a producer is willing to accept. For each product, in each country, interactions between all consumers and producers can be represented by a classical market equilibrium illustrated Figure 1.8. I will come back to market equilibrium mechanisms in section 1.3.1.
Section 1.3 will later describe how forest management, the industrial supply chain and international trade are structured by several layers of market interfaces.
The impact of public policies on forest products supply and demand
Forests provide a habitat for animal species and a place for human leisure activities. Environmental services such as animal habitat and landscape beauty do not have a market valuation. But their values are far from negligible, indeed several studies estimated that the value of forest recreation (Zandersen and Tol, 2009) for example can be as high as the value of timber production. Environmental policies take into account the increased economic value generated by recreation amenities. As they potentially reduce the amount of wood available for harvest (Verkerk et al., 2008), forest protection policies are meant to interact with the industry. Forest policy is a balancing act between economic, environmental and social constraints. This section highlights some of the public policies that aﬀect the forest sector and the interactions between them.
Renewable energy and CO2 emissions policies
Even though the major part of the wood consumption volume globally is used for energy purposes, by far the major part of its value is generated from wood-based products. One of the purposes of this thesis is to analyse policies influencing ma-terial uses of wood. But because they are based on the same material, renewable energy policies have a strong influence on the forest sector. Biomass is the first source of renewable energy in Europe. Indeed biomass and renewable waste ac-count for two thirds of the primary renewable energy production in the EU-28 in 2013 (EUROSTAT, 2015), much higher than hydropower and wind combined. It should be noted that wood biomass represents only a part of total biomass con-sumption. Used mainly for heat production and to a lesser extend for electricity generation, biomass consumption has continued to increase in recent years, though at a slower pace than solar and wind. Increased consumption has been incentivised by EU renewable energy targets which encourage alternatives to fossil fuels (Euro-pean Parliament, 2009). Yet the increased use of biomass in the renewable energy mix has been criticized by the forest sector for over emphasising the use of fuel wood at the expense of material uses of wood (Mantau et al., 2010). On the other hand, biomass co-production is essential to the profitability of forest sector activities.
For the purpose of maximising emissions reduction, diﬀerent levels of wood resource use should be distinguished. A meta analysis of twenty studies shows that on average 1 ton of carbon used in wood products creates an emissions reduction of 2 tons of carbon (Sathre and O’Connor, 2010). Wood products are a valuable means to reduce CO2 emissions when used in a building or furniture. In fact, when looking at avoided emissions, wood processing is more eﬃcient than alternative materials such as plastic, aluminium, steel or concrete. Some have concluded that wood energy should be used as a last resort, at the end of life of its valuable material use.
Climate change mitigation option within the AFOLU IPCC 5th assessment 3rd report on mitigation mention changes in consumption behaviour as a mitigation option: “Demand-side options (e. g., by lifestyle changes, reducing losses and wastes of food, changes in human diet, changes in wood consumption), though known to be diﬃcult to implement, may also play a role (Section 11.4).”
Biodiversity conservation and forest recreation policies
Biodiversity consideration might seem remote from the production imperatives of the forest sector, they are nonetheless central to the integrity of forest ecosystems. In this section I would like to briefly describe the trade-oﬀ between biodiversity conservation and intensification of forest management practices. Although the link with forest products consumption may seem tenuous, biodiversity protection issues could participate in the environmental consciousness of forest products consumers.
Drivers of long term change
The trade oﬀ between production intensification and biodiversity protection is de-cided upon by public policies. Policy answers to this trade-oﬀ can lead to several outcomes on an axis with intensified harvesting on the one side and complete pro-tection on the other side. Plantation and short rotation coppices being the most intensive types of management. Plant and animal diversity are high on naturally re-generated forests with mixed species and they are low on mono-species plantations. In general, the low environmental impact of forest management practices mean that commercially managed forests harbour biodiversity. There are several degrees of biodiversity importance, and the methodologies for assessing economic impacts on biodiversity are still in development. It is clear that private and public forests are important elements in the connectivity of protected national parks. Together, all forests – with various degrees of harvesting activities within them – harbour the ma-jor part of continental biodiversity (Myers et al., 2000). Another important source of policy interest linked with animal presence in western European forests is the browsing by animals such as deer. Over browsing tends to reduce tree species di-versity, although the eﬀect on tree species dispersion are not fully understood Gill and Beardall (2001).
Forest structure also contribute to the quality of the landscape for recreation pur-poses such has hiking and cycling. Consumer preferences for specific landscapes has indirect measurable market impacts on tourism and housing prices for exam-ple. Surveys evaluate how some consumers have a preference for mixed forest stand in comparison to mono-specific plantations (Abildtrup et al., 2013; Nielsen et al., 2007).
Countries vary greatly in the way they have built forest regulations to deal with biodiversity related trade-oﬀs. Some have set aside part of the forest area for com-plete conservation and allowed intense management in the remaining areas. This is the approach taken in countries such as the USA and New Zealand. European countries on the other hand have taken an integrative approach, where forest man-agement integrates biodiversity conservation principles (Bollmann and Braunisch, 2013). Such management principles are called multifunctional: they should achieve the joint purposes of wood production, biodiversity conservation and recreation in the same forest. Additionally, increased forest protection in developed countries can lead to potential leakage and forest degradation in other countries.
Overall, forests role as biodiversity habitat contributes to the image of forest prod-ucts as environmentally friendly products. Consumers certainly do not have a direct influence on forest management, but they could have an indirect influence by shift-ing to certified products. Indeed the Forest Stewardship Council (FSC) certification for example requires harvesting operations to set aside some trees for biodiversity purposes.
Information Technology and Paper products scenarios
Since the mid 1990ies, a decline in newsprint consumption was observed in the United States and in other countries. Hetemäki (1999) emitted the hypothesis that this decline was due to the rise of information technology. But there was little data at the time to support this hypothesis. Similar to above developments, there is yet too little data to analyse composite wood products and multi storey construction impacts on a macroeconomic level. However additional data for the paper market which has accumulated since the work of Hetemäki (1999) 20 years ago shows a newsprint consumption decline in a large number of countries and a similar decline in the demand for printing and writing paper. Chapter 4 provides a detailed ac-count of studies that have attempted to add new explanatory variables – related to Information Technology – that could explain the structural change. I contribute a new approach by using information technology as a threshold variable to explain the non linearity in paper demand.
Forecasting of wood-products markets
To make informed decisions, policy makers and investors are interested in demand and supply forecasts. A simplistic forecasting method would consider for example that linear trends of the past continue in the future. But trends tend not to con-tinue for ever and when they change, numerous alternative future scenarios appear. Then to generate realistic scenarios, it becomes important to understand patterns of relationship between consumption and other relevant macroeconomic variables. For example a shock on final products demand will impact market prices and pro-duction. Other shocks on raw material supply will impact international trade flows. To understand these relationships, forest economists have used Samuelson’s theories (1952) on the equilibrium between demand, supply and international trade. Such models provide policy relevant simulations when changes are expected in the market.
The following section will describe similarities and diﬀerences between a few global and national forest sector models. Then I will describe the structure of a dynamic-recursive partial equilibrium model to prepare some context for chapter 2 and especially describe how assumptions on demand are an integral part of forest sector models. Finally I will describe some of the estimation issues related to demand models to provide elements of context for chapters 3 and 4.
Table of contents :
1.1. Forest resources supply
1.1.1. Sustainable Forest Management
1.1.2. Industrial supply chain network
1.1.3. International Trade
1.1.4. Forest products consumption
184.108.40.206. Fibre-based products
220.127.116.11. Substitute products
1.2. Drivers of long term change
1.2.1. The impact of public policies on forest products supply and demand
18.104.22.168. Renewable energy and CO2 emissions policies
22.214.171.124. Biodiversity conservation and forest recreation policies
126.96.36.199. Forest certification
1.2.2. The impact of structural changes
188.8.131.52. Composite material and wood fibre
184.108.40.206. Wood construction scenarios
220.127.116.11. Information Technology and Paper products scenarios
1.3. Forecasting of wood-products markets
1.3.1. Approaches to forest sector modelling
1.3.2. Applications of forest sector models
1.3.3. Details of a partial equilibrium model
18.104.22.168. Static market equilibrium
22.214.171.124. Dynamic market shifts
1.3.4. Econometric modelling of forest products demand
126.96.36.199. Theoretical derived demand model
188.8.131.52. Relevance of considering a demand function isolated from the rest of the market
184.108.40.206. Spurious regression issues
2. Potential impact of a transatlantic trade and investment partnership on the global forest sector
2.2.2. Global Forest Products Model
2.2.3. Effects of the TTIP
2.2.4. Macroeconomic scenarios
2.3.1. Price effects
2.3.2. Effects on industrial roundwood
2.3.3. Effects on sawnwood
2.3.4. Effects on wood-based Panels
2.3.5. Effects on wood pulp
2.3.6. Effects on paper and paperboard
2.3.7. Effects on value added
2.3.8. Welfare effects
2.3.9. Sensitivity analysis
2.4. Summary and conclusion
3. Reassessing forest products demand functions in Europe using a panel co-integration approach
3.3. Model and data
3.4.1. Panel non stationarity tests
3.4.2. Cointegration tests and estimation method
3.5.1. Panel unit root tests
3.5.2. Cointegration tests
3.5.3. Estimated demand elasticities
4. Information technology, substitute or complement to paper products demand?
4.3. Theoretical model
4.4. Estimation method and data
A. Résumé détaillé en français
A.1. Contexte et méthodes d’analyses de la consommation de produits bois
A.2. Impact Potentiel d’un Accord de Partenariat Transatlantique sur le Secteur Forestier Mondial
A.3. Réévaluer la demande de produits forestiers en Europe à l’aide d’une approche par cointégration en panel
A.4. Les technologies de l’information, complément ou substitut de la demande de papier?
B. Appendix to chapter 3 additional samples
B.1. Unit root tests
B.1.1. PANIC (2004)
B.1.2. Carrion-i-Silvestre (2005)
B.1.3. Bai Carrion (2009)
B.2. Cointegration tests
B.2.1. Westerlund (2007)
B.2.2. Westerlund and Edgerton (2007)
B.2.3. Banerjee Carrion (2015)
B.3. Demand elasticities
B.3.1. Estimation by DOLS and PMG
B.3.2. Comparison plot
B.4. GFPM demand scenarios
B.4.1. Comparison of estimated elasticities with the literature
B.4.2. GFPM demand scenarios
C. Appendix to chapter 4
C.1. Panel cointegration tests
C.2. Thresholds results for consumption per capita in difference
C.3. DOLS and PMG estimation before and after an average break
C.4. Descriptive statistics