Estimating the consumption-based carbon footprint: a small area model as a tool for place-based policies

Abstract

The Carbon Footprint (CFP), derived from household consumption survey data, is a crucial indicator for assessing the impact of human consumption on greenhouse gas emissions. The definition and measurement of personal CFP rely on data that require the comparability between classifications of firm production and household consumption. In this article, we address three key aspects related to the definition and estimation of personal CFP. First, to compute the CFP, we build upon the methodology developed by Pang et al. (2020, Urban carbon footprints: A consumption-based approach for Swiss households. Environmental Research Communications, 2(1), 011003.) by incorporating a conversion factor matrix into the formulas, using official Eurostat data. This matrix serves as a bridge between macroeconomic data across different statistical classifications of production and consumption. Second, aiming to conduct inferential procedures on CFP, we select the probability distribution that best fits CFP empirical distribution. The Generalized Beta Distribution of the Second Kind (GB2) provides the best fit. Third, in order to map local CFP through reliable estimates we propose a Small Area Estimation (SAE) model based on Generalized Additive Models for Location, Scale, and Shape (SAE-GAMLSS), assuming CFP follows a GB2 distribution. Finally, we emphasize the significance of mapping per-capita CFP based on reliable local estimates to support the implementation of effective place-based policies.