This study has provided the idea and process of harvest leakage quantification and tracing of the leakage through forest sector modelling, and which is valuable and applicable to other countries for assessing the same issue. The key to the tracing of the leakage is to investigate the response of domestic demand (of raw materials) and trading (of both raw material and final products), and the changes in the trading will determine the size of the leakage.
The results of the modeling indicate that reducing harvest levels in one Nordic country will be offset by varying levels of increased harvest in the other countries, thereby raising the importance of sustainable and legal harvest practices in these countries. This emphasizes the imperative for robust international communication and collaboration to develop multilateral forest policy. On the other hand, a harvest policy that is supportive in one Nordic country could potentially offset unsustainable harvesting in other countries through the harvest leakage effect. This policy may also benefit industries in the country that adopts it, as well as those in surrounding countries, by leading to lower timber prices. For example, increasing harvest levels in Norway, Finland, and Sweden could raise sawnwood production in Sweden. Similarly, sawmills in Norway may require stable harvest levels to remain competitive. These results are consistent with a prior study by (García et al., 2018), that indicated how an increase in supply and reduced prices of managed forests may decrease incentives for harvesting in unmanaged forests. It is important to note that encouraging harvest policies should prioritize long-lived harvested wood products, as compared to biomass for energy, to prolong carbon sequestration in timber.
The study reveals that the modeling of price for both raw material and final products are vital for assessing the way of leakage (raw materials or final products). In this study the leakage of sawnwood production is evident when harvest decreases in Norway, i.e., sawnwood is imported to Norway. However, (A. M. I. Kallio & Solberg, 2018)’s study found that Norway would import more sawlogs from Sweden to maintain the original production level. In addition, Norwegian sawmills are expected to maintain the original production level and export extra sawlogs to other countries (mainly to Sweden) when harvest increase, while (A. M. I. Kallio & Solberg, 2018)’s results indicate the increased sawnwood production level in Norway. This is relating to the modelling of the relative price relationship in the models, the goods always flow from the low-price region to high-price region, and the direction of the trading flow in the reference year will not change if the original relative price relationship is still the same. Using the example of Norway and Sweden, the net trade of roundwood flows from the low-price region in Norway to the high-price region in Sweden, while the sawnwood flows from the low-price region in Sweden to the high-price region in Norway in the reference year. When changes in harvesting occur (within the most of tested constraints), it is unlikely that the direction of trading flow will change due to the consistent nature of the original relative price relationship, although the amount will vary in response to these changes. Thus, the net flow of roundwood will continue to move directly from Norway to Sweden, albeit in a reduced amount, and the net flow of sawnwood will continue to move from Sweden to Norway, albeit in an increased amount. The utilization of two different models is the primary cause for the disparate outcomes, and the determining factors for such outcomes include model parameters, such as the estimation of transportation cost, labor, and energy cost, in addition to the refinement of geographical resolution.
The results of the modelling analysis demonstrate that the price of raw materials and final products, mobility in the global market, transportation costs, production costs, and the ability to increase production capacity are critical factors contributing to harvest leakage, and their combined influence results in a dynamic flow of leakage. Sweden, due to its high ramping-up capacity in sawmills, strong connections to neighboring countries (Norway, Finland, and Denmark), and lower production costs, plays a pivotal role in the flow of raw materials and final products, consequently impacting the scale and direction of harvest leakage in Nordics. Additionally, the region-level model reveals several dynamic aspects when analyzing regional harvest leakage. Regions with the highest leakage are not necessarily those with the highest levels of harvesting and production but are also influenced by the competitiveness of the local industries. For instance, the region S1 in Sweden, despite not having the most abundant forest resources, possesses the highest ramp-up capacity, which drives up regional harvest the most.
The present study does not offer a simple solution to mitigate the negative impacts of harvest leakage; rather, it contributes to our understanding of how leakage occurs, and the various dynamics associated with leakage in different countries. The study examines two possible directions of harvest changes in a country and finds that both leakage (change in the opposite direction) and spillover effects (changes in the same direction) are present in other countries when harvest changes occur in one country. Therefore, it highlights the need for a more sustainable approach to forest resource usage, rather than focusing solely on harvest volume, we must also consider the impact of harvesting methods on biodiversity and ecosystem preservation. E.g., adopting selective harvesting rather than clear-cutting, should be more valued when the goal relates to biodiversity and ecosystem preservation(Brunet et al., 2010). Meanwhile if a relatively stable domestic timber supply is maintained the leakage and spillover effect can be minimized in open markets.
The analysis presented in this study is of considerable importance to policy makers, investors, and industries, as forest policies can have varying effects on different players due to the leakage effect. For instance, a forest conservation policy implemented in a country may lead to a decrease in timber supply, particularly at certain times. However, the impact on various stakeholders will vary depending on factors such as the location of the industry, the level of policy strictness, and the role of the player. For example, a forest owner may benefit from the increased timber price, while sawmills in different countries may face either a shutdown or an expansion of production. Ultimately, the market effects of the policy can have an impact on social welfare. As such, it is crucial to be aware of the market effects, including harvest leakage, to avoid unintended consequences.
However, it should be noted that this model has limitations in quantifying the leakage rates of forest harvest. Firstly, the leakage rates in this study are solely focused on harvest leakage and may not correspond to carbon leakage, as they overlook carbon emissions during transport and forest industry production, substitution effects, and differences in forest growth. Furthermore, this study omits the carbon sequestration effects in soil and surface albedo effect (Schwaiger & Bird, 2010). Secondly, the regions in the model are still relatively large compared to actual timber transportation ports, which may lead to an underestimation of the inter- and intra-transportation costs, possibly resulting in an overestimation of the leakage. The trade of timber is directly linked to price differences and transportation costs between regions. Moreover, this study primarily focuses on the short-term effects of leakage since the model was run for a single year, while the long-term leakage effect is likely to be more significant. Additionally, non-Nordic countries are simplified as one area (ROW), although they are the primary leakage area for the studied Nordic countries. As such, further geographic refinement is necessary to capture more detailed trading behavior. Finally, the model is also limited to the leakage of the boreal forest, and the leakage of hardwood is underestimated.