Endogenising capital in multi-regional input-output models: implications for sustainability analysis
MetadataVis full innførsel
Reducing anthropogenic greenhouse gas (GHG) emissions is one of the defining challenges of our time. A prerequisite for designing national or international carbon mitigation policies is the availability of comprehensive methods for GHG emissions accounting. However, in a globalised world where trade volumes keep growing and goods travel long distances from production sites to end consumers, the accounting of GHG emissions is becoming increasingly difficult. Consumption-based (CB) accounting captures the direct and indirect impacts associated with the production of goods and services and allocates them to the final consumers rather than producers, and the impacts calculated according to CB principles are often referred to as footprints. Environmentally extended (EE) multi-regional input-output (MRIO) analysis has emerged as the tool of choice for calculating footprints as it enables practitioners to calculate a variety of environmental and social impact indicators that take into account the upstream impacts of final products. While current MRIO models effectively account for the upstream impacts associated with intermediate goods, they do not treat capital goods as inputs to production processes but as exogenous components of the inter-industrial system. Capital goods are, per definition, produced in order to be utilised in further production processes, and not treating them as such implies that footprints as they are currently calculated underestimate the impacts of goods and services for final consumption, and thereby also the impacts embodied in international trade. This thesis therefore aims to investigate how capital goods can be better integrated in MRIO models. A preliminary study was performed to obtain an understanding of how capital contributes to GHG emissions. Using the EE MRIO database EXIOBASE2, we analysed the size, structure and carbon footprint of the gross fixed capital formation (GFCF) for the 48 available countries and regions, and found that in 2007 (the year of study) the GFCF stood for 24% of the global final demand of goods and services but contributed to 30% of the global GHG emissions, with large variations observed across the analysed countries. Furthermore, by comparing the structure of the GFCF in different countries, we concluded that developed countries tended to invest in less carbon-intensive assets than countries at low and intermediate levels of development, and that the overall carbon intensity of GFCF varied substantially. These results pointed to the importance of integrating capital in MRIO models based on detailed and consistent auxiliary data, and models presented in this thesis are therefore constructed using approaches that have substantial data requirements. The flow matrix method described in paper II entails that the capital goods currently in use are disaggregated over assets and industries to create a capital use matrix. This disaggregation was done using capital use proxies from various external sources such as the KLEMS and WORLDKLEMS databases, which were harmonised against the EXIOBASE classification so that capital could be endogenised into the inter-industry system of EXIOBASE, thereby closing the IO system for capital. Using this capital-augmented IO framework, we applied standard Leontief demand-pull calculus to compute footprints that included the upstream impacts associated with both current and capital production requirements. Our results showed that endogenising capital in MRIO models substantially increased the carbon and material footprints of final consumption, and that this increase varied a lot across countries. We also noted increases in total emissions embodied in trade, and found that current disparities between CB and production-based measures of GHG emissions increased further for most countries. The product-level results showed important differences between product categories, with the increases in the footprints of service categories being substantially larger than for non-services, indicating that service sectors – which account for an increasingly large share of the global economic output (particularly for wealthier countries) – contribute much more to various environmental problems than previously thought. While the results confirm that the endogenisation of capital has substantial implications for CB accounting, it must be noted that the models used in this thesis still rely on many assumptions that impinge on the robustness of the model. One of these was analysed in depth in the fourth paper of this thesis, with the conclusion that an explicit temporal resolution is needed to consolidate the capitalaugmented IO framework, including detailed age cohort composition of the current capital stock as well as longer time series than currently available. Nevertheless, we hope that the analytical approaches adopted in this thesis as well as the models themselves could help the further development of inputoutput and industrial ecology methods in answering some of the key sustainability questions of our generation.
Består avPaper 1: Södersten, Carl-Johan; Wood, Richard; Hertwich, Edgar G.. Environmental Impacts of Capital Formation. Journal of Industrial Ecology 2018 ;Volum 22.(1) s. 55-67 https://doi.org/10.1111/jiec.12532
Paper 2: Södersten, Carl-Johan; Wood, Richard; Hertwich, Edgar G.. Endogenizing capital in MRIO models - the implications for consumption-based accounting. Environmental Science and Technology 2018, 52, 22, 13250–13259 https://doi.org/10.1021/acs.est.8b02791 Copyright © 2018 American Chemical Society
Paper 3: Södersten, Carl-Johan Hugo; Wood, Richard; Wiedmann, Thomas O. The capital load of global material footprints. - Final published version in Resources, Conservation and Recycling 2020 ;Volum 158. https://doi.org/10.1016/j.resconrec.2020.104811 This is an open access article under the CC BY license
Paper 4: Södersten, Carl-Johan Hugo; Lenzen, Manfred. A supply-use approach to capital endogenization in input-output analysis. - Final published version in Economic Systems Research 2020 https://doi.org/10.1080/09535314.2020.1784852