The global imperative to reduce greenhouse gas (GHG) emissions to limit climate change impacts necessitates a significant transport transition. The European Union's aim of net-zero emissions by 2050, embodied in the European Climate Law (2021) and the Fit for 55 package, sets ambitious targets. These targets include policies such as increasing renewable energy, implementing a carbon border adjustment mechanism, mandating zero-emission vehicles by 2035, and introducing a new Emissions Trading Mechanism (ETS2) by 2027. However, the economic and social challenges of this transition are substantial, particularly considering recent events such as the COVID-19 pandemic, inflation, and the energy crisis. Public acceptance of these policies is low, largely due to perceptions of regressive impacts, disproportionately affecting lower-income households. The 2018 Yellow Vest protests in France, triggered by a carbon tax increase, exemplify this public discontent. This study aims to identify those most negatively impacted by transport transition policies by defining and quantifying transport poverty in Germany, to enable the design of just and efficient policies that accelerate the transition.

The literature lacks a unified definition of transport poverty. While attempts have been made to define it considering factors like income, fuel expenditure, access to transport, and mobility practices (Verhorst et al., 2023; Banister, 2018; Lucas et al., 2016; Berry et al., 2016), no consensus exists. Existing indicators often focus on single dimensions (e.g., the 10% income threshold for transport expenditure) and fail to capture the multidimensional nature of the problem (Berry et al., 2016). This paper adopts Sen's Capability Approach (1992), which emphasizes the importance of individual capabilities and freedoms, not just resources, to achieve well-being. This approach has been successfully applied to energy and fuel poverty studies (Charlier et al., 2021; Charlier and Legendre, 2018; Day et al., 2016; Berry et al., 2016). Building on previous work (Lucas et al., 2016; Berry et al., 2016), this study considers various dimensions such as mobility, accessibility, affordability, and exposure to externalities, utilizing the capability approach to develop a multidimensional transport poverty indicator.

The study employs a two-step methodology to identify and classify transport poverty. First, principal component analysis (PCA) is used to reduce the dimensionality of a dataset comprising eleven variables identified from the literature as potential indicators of transport poverty. These variables include monthly distance traveled, monthly trips taken, monthly time traveled, extra travel time (public vs. private transport), car use restriction, poor spatial matching, lack of alternative transport, car ownership, income, levelized cost of driving one kilometer (LCOKm), and number of trips for personal needs. PCA identifies three principal components: (1) travel behavior conditional on spatial matching; (2) driving restrictions; and (3) resources available (income, LCOKm, transport alternatives). The PCA results unify the often disparate variables and dimensions used in previous transport poverty studies. Second, a latent class model (LCM) is used to classify households into different transport poverty classes based on the three principal components. The LCM determines the optimal number of classes, avoiding arbitrary thresholds used in previous studies. The authors test models with two, three, and four classes, selecting the four-class model as the best fit using Akaike (AIC) and Bayesian (BIC) criteria. The LCOKm is an original contribution representing the average cost of driving one kilometer over a vehicle's lifetime, incorporating fuel, operation, maintenance, and purchase price. The data used is from the German Mobility Panel (MOP) survey (2004-2019).

The PCA reveals three key dimensions of transport poverty: travel behavior conditional on spatial matching, driving restrictions, and resources available. The LCM analysis identifies four classes within the German population: (1) Independent (47.8%): households with good spatial matching and diverse transport options; (2) Sufficient (34.44%): households with high income and high transport expenditures, despite poor spatial matching; (3) Car-dependent (15.55%): households with poor spatial matching and reliance on cars; and (4) Transport-poor (2.21%): households with low income, poor spatial matching, and reliance on expensive public transport. An estimated 14.7 million German households (1,838,057 transport-poor and 12,932,935 car-dependent) are classified as transport-poor or car-dependent. The analysis of cost elasticity of driving demand shows that the transport-poor and car-dependent groups exhibit inelastic demand, meaning they are highly sensitive to price changes and unable to easily reduce their driving. Conversely, the sufficient and independent groups show elastic demand, allowing them to adjust their driving behavior in response to price increases. The study demonstrates that using fuel expenditure alone underestimates the impact of pricing on household decision-making. The LCOKm provides a more accurate measure of the total cost of car ownership.

The findings highlight the heterogeneous impacts of transport transition policies across different socioeconomic groups. The car-dependent and transport-poor are particularly vulnerable to increases in driving costs, facing a trade-off between essential travel and other budgetary needs. The study's results address the research question by identifying the specific groups most likely to experience hardship due to the transport transition. This granular understanding of transport poverty is crucial for policy design. The use of the capability approach and multidimensional indicators allows for a more nuanced understanding than previous, simpler approaches. The study's results are relevant to the field by providing a robust methodology for defining and measuring transport poverty, offering a valuable tool for policy makers across various contexts.

This paper presents a novel transport poverty scale based on a two-step methodology, providing a clear framework for identifying those most affected by the transport transition. The identification of four distinct classes of transport poverty—independent, sufficient, car-dependent, and transport-poor—allows for targeted policy interventions. The findings strongly suggest focusing on the car-dependent and transport-poor classes with redistributive policies to ensure a just transition. Future research could explore the use of behavioral economics to enhance policy effectiveness, expand the analysis to other EU countries using comparable datasets, and investigate the long-term impacts of various policy interventions on transport poverty.

The study is based on German data from the MOP survey, potentially limiting the generalizability of the findings to other countries or contexts. The analysis focuses on short-term responses to price changes; long-term behavioral adaptations to transport transition policies may differ. The LCM relies on the assumption of conditional independence among the chosen variables, which may not fully capture the complexities of real-world interactions.