---
category: literaturenote
citekey: alonsoepeldehowcanwealleviate2026
title: How can we alleviate transport poverty? Insights from a cluster analysis for Spain
authors: "Alonso-Epelde, E.; García-Muros, X.; González-Eguino, M."
year: 2026
date: 2026-01-01 2026-01-01
doi: 10.1016/j.tranpol.2025.103863
publication: Transport Policy
url: "https://www.sciencedirect.com/science/article/pii/S0967070X25004068"
zotero_key: 7UEXIMME
zotero_storage: GGYF8NE2
collections: imporditud
folder: 001_artiklid
firstAuthor: "Alonso-Epelde, E."
status: converted
---
## Contents lists available at [ScienceDirect](www.sciencedirect.com/science/journal/0967070X)
## Transport Policy
journal homepage: [www.elsevier.com/locate/tranpol](https://www.elsevier.com/locate/tranpol)
# How can we alleviate transport poverty? Insights from a cluster analysis for Spain
E. Alonso-Epelde a,b,\* , X. García-Muros a,c , M. Gonz´ alez-Eguino a,b,c
- a *Basque Centre for Climate Change (BC3), Edificio Sede 1-1, Parque Científico de UPV/EHU, Barrio Sarriena s/n, Leioa, 48940, Spain*
- b *Universidad del País Vasco (UPV/EHU), Barrio Sarriena s/n, Leioa, 48940, Bizkaia, Spain*
- c *Ikerbasque, Basque Foundation for Science, Bilbao, Spain*
#### ARTICLE INFO
*Keywords:* Transport poverty Energy poverty Climate policy Vulnerability Mitigation
#### ABSTRACT
Addressing transport poverty in the context of rising energy prices is increasingly relevant, given the fundamental role that transportation plays in facilitating access to basic needs and rights. Identifying and characterising transport-poor households is crucial to developing targeted policies and compensation schemes. During the 2022 energy crisis, where such data was not available, many countries decided to implement large universal fuel subsidies. These measures, which were widely criticized by environmental NGOs and institutions such as the IMF and OECD as ineffective and costly, underscore the urgency of finding alternative policy designs. This paper addresses this shortcoming, using the Low-Income-High-Cost (LIHC) transport poverty indicator and hierarchical cluster analysis to identify and analyse six profiles of transport-poor households in Spain. We assess whether a universal discount of 20 cents per litre of fuel, a policy in place in Spain during 2022 and similar in many other countries, could be replaced by more targeted measures. Our results show that targeted interventions based on this type of cluster analysis could significantly reduce transport poverty levels more effectively and economically than universal subsidies, resulting in a just energy transition.
## **1. Introduction**
Since the 1980s, research on energy poverty has extensively explored the ways in which to identify and measure households that are vulnerable to this form of deprivation, which manifests in various ways: disproportionate energy expenditure, inability to maintain adequate home temperatures, accumulation of utility bill arrears or forced reduction of other essential needs to pay energy bills ([Adom et al., 2021](#page-13-0); [Adusah-Poku and Takeuchi, 2019;](#page-13-0) [Bednar and Reames, 2020](#page-13-0); [Boardman, 1991](#page-13-0); [Bradshaw and Hutton, 1983](#page-13-0); [Castano-Rosa](#page-13-0) ˜ et al., [2019;](#page-13-0) [Dong et al., 2021](#page-13-0); [Foster et al., 1984](#page-13-0); [Gonzalez-Eguino,](#page-13-0) ´ 2015; [Halkos and Gkampoura, 2021](#page-13-0); [Hills, 2012](#page-13-0)). However, until a few years ago, studies on energy poverty had focused on households struggling to access essential domestic energy services (such as electricity or heating), and neglected another critical dimension of this type of poverty: transport poverty [\(Martiskainen et al., 2021\)](#page-13-0).The failure to address transport poverty within the broader discourse on energy poverty1 represents a significant gap.
In recent years, transport poverty has attracted increasing attention in both academic and political spheres, given its social, economic, and environmental implications. This is especially pertinent in the current era of energy transition, where rising energy costs and the push for climate policies are making transport a critical component of household energy consumption. In the European Union (EU), for example, transport accounts for nearly 50 % of household energy expenditures.2 As transport is an essential means for accessing education, healthcare and
\* Corresponding author. Universidad del País Vasco (UPV/EHU), Barrio Sarriena s/n, Leioa, 48940, Bizkaia, Spain.
*E-mail address:* eva.alonso@bc3research.org (E. Alonso-Epelde). 1 Energy poverty has traditionally been understood in the literature as domestic energy poverty, encompassing various definitions that consistently describe insufficient domestic energy consumption to meet basic needs, such as heating, lighting, and cooking. Although transport poverty has not been widely recognised as a form of energy poverty in the literature, given that it is also directly related to energy, we have considered both dimensions (domestic energy poverty and transport poverty) as distinct yet interconnected aspects of a broader concept of energy poverty. Importantly, because the causes and consequences of transport poverty differ, our analysis remains specific to this dimension while acknowledging its place in the wider energy poverty framework.
2 We consider household energy expenditure to include both domestic energy (i.e. heating, cooling, electricity for appliances …) and transport energy (expenditure on fuels for private vehicles and public transport fares for daily mobility).
employment [\(Cass et al., 2005](#page-13-0); [Coote and Percy, 2020](#page-13-0)), households affected by transport poverty face significant limitations on their rights and opportunities, which can perpetuate social exclusion and deepen socioeconomic inequalities ([Kenyon et al., 2003](#page-13-0); [Lucas, 2019\)](#page-13-0).
For this reason, the emerging literature on transport poverty3 has focused on conceptualising, measuring, and analysing the phenomenon of transport poverty ([Alonso-Epelde et al., 2023;](#page-13-0) [Berry et al., 2016](#page-13-0); [Carruthers et al., 2005](#page-13-0); [Lovelace and Philips, 2014](#page-13-0); [Lucas, 2012, 2019](#page-13-0); [Lucas et al., 2016](#page-13-0); [Mattioli et al., 2016, 2018\)](#page-13-0). Although several authors have attempted to develop and establish a single definition for transport poverty, there is still no well-established standard definition. However, in this paper, we have adopted the one proposed by [Lucas et al. \(2016\)](#page-13-0), that considers a household or individual to be transport poor if they meet any of the following conditions: i) if they do not have mobility options adapted to their abilities/disabilities or physical conditions, ii) if they cannot maintain a reasonable quality of life due to the lack of transportation options that allow them to travel to places where they can carry out their daily activities, iii) if their disposable income is below the official poverty line due to high expenses to cover mobility needs, iv) if they have to spend excessive time commuting (with the risk of suffering time poverty or social isolation) or v) if they have to travel in dangerous, unsafe or unhealthy conditions regularly.
As far as transport poverty measurement is concerned, there is a growing body of literature that proposes and analyses indicators for transport poverty4 [\(Alonso-Epelde et al., 2023;](#page-13-0) [Berry et al., 2016](#page-13-0); [Car](#page-13-0)[ruthers et al., 2005; Dodson and Sipe, 2007; Lovelace and Philips, 2014](#page-13-0); [Mattioli et al., 2016\)](#page-13-0). While no single indicator is perfect and none is able to address all dimensions of transport poverty (affordability, mobility poverty, accessibility and exposure to transport externalities), some have the advantage of being easily replicable in different contexts and over time, and are particularly useful for policymakers. For this reason, this study uses the Low Income High Cost indicator proposed by [Alonso-Epelde et al. \(2023\)](#page-13-0) to measure transport poverty.
The metric selected focuses on the affordability dimension of transport poverty, which is expected to be one of the dimensions that will be most affected in the coming years, due to energy transition policies that will increase fossil fuel prices. While there is extensive debate in Europe about affordability challenges in the context of rising energy costs, in the United States (US), the concept of "transportation expenditure burden" is beginning to be used to address this dimension of transport poverty ([Vaidyanathan et al., 2021](#page-14-0)). In the US, this issue has also been analysed in conjunction with housing expenditure burden ([Salon et al., 2016](#page-14-0); [Schouten, 2022\)](#page-14-0), as exemplified by indicators like the Housing + Transportation (H + T) Index developed by the Center for Neighborhood Technology, which operates as an aggregate census-tract measure, reflecting neighborhood-level characteristics rather than individual household circumstances ([CNT, 2012](#page-13-0)). Although this approach is less common in Europe, some indicators, such as the LIHC, account for disposable income after deducting both transport and housing costs when determining whether a household falls below the poverty threshold.
In the context of energy transition policies, addressing transport poverty is not just a matter of social equity but also of environmental necessity. As governments implement policies to mitigate climate change, such as increasing taxes on fossil fuels to penalise private vehicle use, there is a risk that without careful design, these measures could exacerbate transport poverty and deepen existing inequalities. This underscores the importance of integrating social justice considerations into energy transition strategies to ensure that vulnerable households are not disproportionately burdened by these changes.
This study aims to contribute to understanding transport poverty by applying a Hierarchical Clustering Analysis to identify the socioeconomic and demographic profiles of the households most affected by transport poverty. In addition, we use the information obtained from the cluster analysis to analyse the potential of targeted policies to alleviate transport poverty from an intersectional perspective. This research aims to inform the development of evidence-based policies that promote social inclusion, equitable access to transportation, and overall well-being by providing a nuanced understanding of transport poverty profiles, particularly in the context of the ongoing energy transition.
The study focuses on the case of Spain, a country with diverse geographical and socioeconomic characteristics and one of the first countries to have carried out studies on transport poverty (see [Alon](#page-13-0)[so-Epelde et al., 2023](#page-13-0)), making it an ideal case for exploring this issue. Spain's urban areas face challenges with public transportation affordability, while rural areas depend on costly private vehicles due to limited public transit. Economic disparities and the nation's focus on sustainability further shape transport poverty, highlighting the need for targeted, sustainable solutions.
The paper is structured as follows: Section 2 describes the data used and the methodology followed to define the profiles of transport poverty, Section [3](#page-4-0) presents the results of the cluster analysis and assesses how the profiles we obtained can be used to design targeted policies to mitigate transport poverty in the current energy transition context, Section [4](#page-8-0) discusses the results and how they are relevant to policy and Section [5](#page-10-0) summarises the conclusions of the paper and its limitations and suggests new research areas for the future.
## **2. Methodology**
#### *2.1. Data*
This paper follows the framework proposed by [Alonso-Epelde et al.](#page-13-0) [\(2023\)](#page-13-0) to identify transport-poor households. This framework proposes several indicators of transport poverty that are politically useful due to their reliance on readily available data and their replicability across most countries. We focused our analysis on the Low Income High Cost (LIHC) metric, since it identifies those who are severely vulnerable (see Annex A) from the affordability perspective and these are the households that should be prioritised for aid to alleviate transport poverty. We used microdata from the Spanish Household Budget Survey (HBS) for2019,5 as provided by the National Institute of Statistics (INE). In 2019, this survey looked at a sample of 20,817 households that were considered to be representative of Spanish society. This dataset includes information on household expenditure on goods and services (including those related to private and public transport) and their demographic and socioeconomic characteristics.
The LIHC metric identifies households that dedicate a large proportion of their income to mobility needs and are also vulnerable from an income perspective (covering the affordability dimension of transport poverty). Therefore, households would be vulnerable to transport poverty if they meet the following two conditions: i) they expend more, in relative terms, than the median on transportation6 and ii) their disposable income after subtracting housing and transport costs is below the poverty threshold (in Spain, this is set at 60 % of the national
3 See Annex A for a detailed literature review on the conceptualisation and measurement of transport poverty.
4 See [Alonso-Epelde et al. \(2023\)](#page-13-0) for a review of the literature on the concept and the different metrics for analysing transport poverty.
5 We focused our analysis on 2019 because the lockdown measures resulting from the COVID-19 pandemic made 2020 and 2021 exceptional years in terms of mobility and because the 2022 survey was not yet available when we started the study.
6 Transport expenditure takes spending on goods and services related to private transport (fuels) into account, along with short and medium-distance public transport (bus, metro, commuter train, etc.). Air and maritime transport are excluded because they are not considered necessary for carrying out basic daily household activities.
**Fig. 1.** Methodology workflow.
median):
*T* ≥ *medc*(*T*)
(*I* − *H* − *T*) *<* 0*.*6*med*(*IAHC*)
where *T* is the equalised transport expenditure7 for each household, *I* is the equalised household income,8 *H* is the equalised housing cost of each household, *medc* is the median calculated on households that expend on transport goods and services, *med* refers to the national median taking into account the whole population and *IAHC* is the income after housing costs.
As far as the strengths of the dataset are concerned, on the one hand, using the HBS microdata was an advantage as they were available and provided an accessible database that covered a broad time series for a large number of European and non-European countries ([Oseni et al.,](#page-13-0) [2021\)](#page-13-0). This makes the study replicable and comparable for various countries or regions. Likewise, the great granularity of the data provided by the HBS makes it possible to use a number of socioeconomic and demographic characteristics of transport-poor households to identify different profiles.
The choice of the LIHC as the primary transport poverty indicator for the study also provides some other advantages, since it is an objective indicator based on quantitative information that identifies severely vulnerable households, thereby eliminating false positives collected by some of the other indicators for measuring transport vulnerability (see [Alonso-Epelde et al., 2023](#page-13-0)). Furthermore, using this indicator facilitates the interrelation with domestic energy poverty measures, so that, in future research, combined profiles could be created that analyse both perspectives of energy poverty, assuming we understand transport poverty as another dimension of energy vulnerability.
However, measuring transport poverty through the LIHC indicator also provides some limitations to the study. While there is no perfect indicator, the main limitation of the LIHC is that it does not cover all the dimensions of transport poverty, as it focuses solely on affordability. While affordability is a critical aspect, transport poverty is a multifaceted issue that also includes dimensions such as accessibility, mobility poverty and exposure to transport externalities. In this sense, hidden transport poverty should also be analysed, since a household might have low transport expenditure not because they are financially secure but due to restricted access to viable transport options, making it difficult to meet basic mobility needs. Consequently, it would be interesting to develop this type of cluster analysis for a composite indicator covering multiple dimensions of transport poverty or for an indicator covering hidden transport poverty. However, to the best of our knowledge, such a comprehensive metric has not yet been developed in a way that is both rigorous and easily replicable across different countries and large time series. Additionally, like domestic energy poverty metrics that exclude purchase and maintenance costs of energy systems (e.g., heating system costs), our approach focuses on recurrent fuel expenditures rather than vehicle ownership costs, a conscious choice for policy relevance and
comparative consistency, though we recognise this omits a significant cost component that could be analysed in depth in further research.
The use of data from the Household Budget Survey (HBS) to calculate metrics like the LIHC has the significant advantage of being easily replicable for longitudinal and cross-country comparisons, which makes these data particularly useful for policymaking purposes. However, the HBS only provides information on household expenditures and characteristics, without capturing whether households have access to viable mobility options or even require them. This limitation restricts the analysis of other dimensions of transport poverty, such as accessibility or the adequacy of available transport options. Nevertheless, we acknowledge that addressing hidden transport poverty and other overlooked dimensions requires additional data and methodologies. Developing composite metrics or conducting primary data collection would be a significant step forward, but would also introduce challenges, such as reduced comparability across contexts and time periods. These limitations highlight the importance of creating a system of complementary indicators to provide a more holistic understanding of transport poverty.
#### *2.2. Hierarchical clustering analysis*
Although 3.6 %9 of Spanish households and 5.2 % of Spanish households that consume transport goods and services would be transport poor according to the LIHC metric, our objective is not to measure transport poverty (this has been done previously by [Alonso-Epelde et al.,](#page-13-0) [2023\)](#page-13-0), but to understand the characteristics of these households in order to facilitate the design of targeted policies (with an intersectional approach) to mitigate transport poverty more efficiently.
To do this, a Hierarchical Clustering Analysis (HCA) was carried out to identify the different vulnerability profiles10 once the transport-poor households had been identified in the database. [Fig. 1](#page-2-0) summarises the methodological workflow. Clustering or cluster analysis is an unsupervised (without any predefined choice) statistical machine learning technique through which observations are grouped based on their similarities [\(Feczko and Fair, 2020; Gao et al., 2023\)](#page-13-0). Since cluster analysis began to be implemented in computer algorithms in the 1960s [\(Forgy,](#page-13-0) [1965;](#page-13-0) [Johnson, 1967](#page-13-0); [Ward, 1963\)](#page-14-0), numerous increasingly complex clustering algorithms have been developed [\(Gao et al., 2023](#page-13-0); [Jain,](#page-13-0) [2010\)](#page-13-0).
In this study, we apply a bottom-up HCA, known as agglomerative clustering [\(Gao et al., 2023;](#page-13-0) [Kaufman and Rousseeuw, 2009\)](#page-13-0), through which a cluster tree is established based on distances between subclusters. Agglomerative clustering is considered a bottom-up approach, because each observation is initially considered to be a single cluster, which is then iteratively merged until all observations form one large cluster. The Gower distance metric was chosen to measure the dissimilarity between households, due to its suitability for mixed datatypes ([Gower, 1966\)](#page-13-0). We use the complete linkage method to produce compact, well-separated clusters, which calculate the distance between two clusters as the maximum distance between any pair of observations from the two clusters. The sample used for the analysis was made up of 714 observations from the HBS, representing 670,521 Spanish households.
[Table 1](#page-4-0) lists the socioeconomic and demographic variables selected for the cluster analysis and the categories associated with each variable. These variables were selected based on the characteristics that have
7 Equalised transport expenditure, equalised income and equalised housing cost are calculated by applying the OECD's modified equivalence scale, thus taking the economies of scale generated in households based on their size into account. The modified OECD scale values the reference person in the household at 1, the remaining members of the household aged 14 or over at 0.5 and the remaining members aged 14 or under at 0.3.
8 Equalised consumption expenditure was used instead of income, as it is considered a better proxy for permanent household income, since it fluctuates less in the long run [\(Goodman and Oldfield, 2004](#page-13-0)) and because in Spain, as in most European Union countries, the survey used to study energy or transport poverty (Household Budget Survey) was highly detailed and reliable in terms of expenditure, but less so for income, which was collected via a single question. Detailed income data was collected in the Survey on Income and Living Conditions (SILC).
9 If the affordability perspective was measured using the LIHC indicator, transport poverty in Spain would amount to 670,521 households or 1,840,232 individuals.
10 Although such analyses do not yet exist for transport poverty, profiling analyses have been carried out for energy poverty using different clustering methods – Qualitative Comparative Analysis (QCA), Multiple Correspondence Analysis (MCA) and Ascending Hierarchical Classification (AHC) ([Belaïd, 2018;](#page-13-0) [Boeri et al., 2020](#page-13-0); [Primc et al., 2019](#page-14-0)).
**Table 1** Characteristics for identifying transport vulnerability profiles.
| | Variable | Category |
|---------------------------|--------------------------------|------------------|
| Household characteristics | Quintile (Q) | Q1 – Lowest |
| | | income level |
| | | Q2 |
| | | Q3 |
| | | Q4 |
| | | Q5 – Highest |
| | | income level |
| | Level of rurality of the | Urban |
| | household's area of residence | Intermediate |
| | (Z) | Rural |
| | | |
| | Household type (TH) | Older single |
| | | persons |
| | | Single person |
| | | Single parent |
| | | Older couples |
| | | Couples without |
| | | children |
| | | Couples with |
| | | children |
| | | Others |
| | Tenure status of the | Rented |
| | household's primary residence | Ownership with |
| | (R) | mortgage |
| | | Ownership |
| | | without mortgage |
| | | Relinquish |
| | Occupational situation of the | Employed |
| | household (OS) | One employed |
| | | Unemployed |
| | | Not provided |
| Characteristics of the | Age of the reference person in | Young |
| reference person of the | the household (A) | Adult |
| household | | Older |
| | Gender of the reference person | Woman |
| | of the household (G) | Man |
| | Country of birth of the | Spain |
| | reference person of the | EU27 |
| | household (C) | Other Europe |
| | | Rest of world |
| | Education level completed by | No education |
| | the reference person of the | Primary |
| | household (S) | education |
| | | Secondary |
| | | education |
| | | Post-secondary |
| | | |
| | | education |
been most studied in energy ([Primc et al., 2019](#page-14-0)) and transport poverty literature, particularly the analysis carried out by [Alonso-Epelde et al.](#page-13-0) [\(2023\).](#page-13-0) They identified the variables that influence the probability of falling into transport poverty using a logit. These variables include key socioeconomic and demographic factors, such as income, area of residence, gender, age, origin and education of the reference person, which we incorporated into our study.
We examined the dendrogram (see [Figure B1](#page-11-0) in Annex B) and used the silhouette assessment (see [Figure B2](#page-11-0)) to determine the optimal number of clusters, which is found by minimising intra-cluster observation distances and maximising the distance between each unique cluster. The dendrogram shows the attribute distances between each pair of sequentially merged classes and makes it possible to choose the optimal number of clusters based on its hierarchical structure. The silhouette assessment is also used to identify the optimal number of clusters, by selecting the solution that produces the most cohesive and well-separated clusters ([Gao et al., 2023](#page-13-0); [Rousseeuw, 1987](#page-14-0)). The highest average silhouette coefficients indicate that the quality of the solution is appropriate, i.e., the clusters are well-defined and separated. Consequently, we chose 6 clusters, based on the largest vertical difference between nodes in the dendrogram and one of the highest coefficients in the silhouette assessment, the characteristics of which will be analysed in the results section.
## **3. Results**
## *3.1. Characterisation of transport poor households*
In this section, we examine the distribution of relevant variables within each cluster to identify patterns and commonalities among transport-poor households. As mentioned in the previous section, we identified six different profiles for transport-poor households based on the characteristics of various households and the household's reference person.11
The profiles of transport-poor households in Spain are presented below, taking the common socioeconomic characteristics of the different households belonging to each cluster into account. In addition, [Fig. 2](#page-5-0) shows how the characteristics mentioned earlier were distributed over the different groups (more detailed data can be found in Annex C). [Table 2](#page-6-0) includes a table with the share of households that fall into each category within each variable and profile.
## *3.1.1. Cluster 1: Rural poor couples with children*
This cluster constitutes 31 % of the sample. These households tend to be made up of couples with children who reside mainly in rural areas with a low population density. Their economic situation is particularly vulnerable as they are low-income families, but at least they have the advantage of owning a house without having to pay a mortgage. The person of reference of the household tends to be an adult (between the age of 30 and 65) male born in Spain, who, in the vast majority of cases, is the only person in paid employment in the household. As far as education is concerned, they tend to have completed compulsory secondary education.
## *3.1.2. Cluster 2: Mortgaged houVseholds*
This group represents 22 % of the sample. These households are characterised as lower middle-income households that, despite owning a house, still have to pay mortgage costs. They are usually made up of single-person or childless couples and typically reside in areas of intermediate density. The person of reference of the household tends to be a man between the age of 30 and 65, born in Spain, with secondary or post-secondary education.
### *3.1.3. Cluster 3: Rural older couples*
Cluster 3 constitutes 17 % of the sample (i.e., transport-poor households according to the LIHC metric). These households are characterised as lower-middle income households residing in rural areas and are mainly made up of older couples without dependent children. In addition, the vast majority reside in a home they own and have no mortgage. The person of reference of the household is usually a man over the age of 65 born in Spain. In addition, most people have only primary education, although they may have also completed secondary education in some cases.
## *3.1.4. Cluster 4: Lower-middle income couples with children*
This group constitutes 15 % of the sample. These households are usually characterised as lower-middle-income households. They are usually couples with children who live in a property without a mortgage located in a disseminated area. In this case, the person of reference of the household is usually a woman between the age of 30 and 65, born in Spain, with a Bachelor's degree or secondary education. Likewise, the person of reference is the only member of the household with paid employment.
11 The person of reference of the household is the member of the household that contributes most to the household consumption budget.
**Fig. 2.** Main characteristics across clusters.
#### *3.1.5. Cluster 5: Immigrants*
This group constitutes 9 % of the sample. These households are characterised by immigrant couples, generally with children who live in rented accommodation in a city. In this case, even though more than one person is employed within the household, the family's income is still medium-low. The person of reference is usually an adult man from a non-European country and has completed secondary or post-secondary education.
## *3.1.6. Cluster 6: Single-parent families*
This group constitutes 6 % of the sample. These homes are characterised by single-parent families, i.e., households made up of a single parent and their children. These households are especially vulnerable to transportation because, in addition to being very low-income households, they usually live in areas of intermediate density (where mobility makes it difficult to carry out both productive and reproductive tasks),
and pay a monthly rent. The person of reference of the household is usually an adult woman born in Spain who has completed secondary education.
When analysing the profiles of transport-poor households, we can see certain factors that determine their vulnerability. This vulnerability can be related to the composition of the household, the economic situation or the dwelling's location. Most groups are made up of low-income or lower-middle-income households that live in rural or intermediate areas. Rural households tend to dedicate a larger proportion of their income to transport, as they have greater mobility needs. This explains why rurality is a key factor in explaining transport poverty in most clusters.
The only group where urban households predominate are immigrant households (cluster 5). This suggests that socioeconomic factors rather than demographics mainly determine this group's vulnerability. Although there are usually several employed people in the household,
**Table 2** Total expenditure (€) and share of expenditure on transport goods and services (%).
| | % of
sample | Average
expenditure
(€) | %
transport | % private
transport | % public
transport |
|-------------------------------------------------------|----------------|-------------------------------|----------------|------------------------|-----------------------|
| Rural poor
couples
with
children | 31 | 19,735 | 16.9 | 15.4 | 1.5 |
| Mortgaged
households | 22 | 17,683 | 15.3 | 13.2 | 2.0 |
| Rural older
couples | 17 | 17,498 | 15.0 | 14.5 | 0.5 |
| Lower-middle
income
couples
with
children | 15 | 18,787 | 15.7 | 14.1 | 1.5 |
| Immigrants | 9 | 22,287 | 16.4 | 12.0 | 4.4 |
| Single-parent
families | 6 | 15,314 | 16.7 | 13.8 | 2.9 |
| All
Households | | 30,860 | 4.5a | 3.9 | 0.5 |
a Analysis of Spanish HBS data reveals that 70 % of households' report transportation expenses, comprising 61 % spending on private transport fuels and 37 % using public transport, with notable geographic variations: urban areas show 58 % private fuel consumption versus 45 % public transport usage, while rural areas demonstrate substantially lower public transport usage at 26 %. Additionally, among transport-poor households identified in our cluster analysis, housing expenditures represent between 24.0 % (rural poor couples with children) and 33.0 % (mortgaged households) of total household expenditure, illustrating the compound affordability challenges faced by these vulnerable groups.
they are lower-middle-class families who spend a significant part of their income on rent. Moreover, this cluster (immigrants) dedicates a higher proportion of their income to public transport and less to private transport (see Table 2), showing that their vulnerability as far as transport is concerned is driven by the cost of public transport and the need they have to move in an urban environment to meet their day-today needs.
Essential housing-related expenses such as rent or mortgage payments considerably reduce households' disposable income, aggravating
**Table 3** Policy scenarios to mitigate transport vulnerability.
| Scenario | Description | Aid by household |
|--------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------|
| General measures | | |
| Reference | A fuel subsidy of €0.20 per litre
at the pumpa | Average reduction in
diesel and petrol prices
of 11 % and 12 % |
| Transfers-all | Lump sum for all Spanish
households, regardless of their
consumption of fuels. | €146.69 |
| Targeted measures | | |
| Low-Income | Lump sum for the lowest 20 % of
low-income households (Q1) | €733.55 |
| Transport-poor | Lump sum for households
identified as vulnerable to
transport poverty | €4100.08 |
| Income&Transport
poor | 70 % of the budget is allocated
based on income and the
remaining 30 % expands this
transfer based on the
sociodemographic
characteristics of households
identified in the transport
poverty cluster analysis. | Base: €1482.68
+ additional amount
depending on
household
characteristics |
a This real measure reduced the average final price of diesel and gasoline by 11 % and 12 % and cost the Spanish government €2.749 billion in 2022. This data is based on the analysis by Gonz´ [alez-Eguino et al. \(2023\).](#page-13-0)
their vulnerability. This is especially relevant in cluster 2, which we have defined as "mortgaged households". These are lower-middle income households, most of which do not live in rural areas (households in intermediate areas predominate), but their situation worsens as a result of their expenses from paying the mortgage (on average 34 % of their income). These households are usually made up of a single person or couples with children. This group is especially relevant from a policy perspective, as they consist of households in the low-middle-income brackets (not necessarily poor), whose vulnerability is directly driven by their expenditure on mobility. Therefore, if aid policies were based solely on economic criteria, they would be excluded, highlighting the importance of including transport vulnerability as a significant factor in addressing broader vulnerable situations.
Households with children predominate in most groups (clusters 1, 2, 4 and 5), indicating that the composition of households also plays an important role in transport vulnerability. This is often linked to increased mobility needs, as mobility is necessary to carry out both productive (e.g., commuting to work) and reproductive tasks (e.g., travelling to take children to school or shopping) and, therefore, the larger the number of members in a household, the more likely they are to experience transport poverty.
Especially vulnerable among these groups are single-parent households (cluster 6), where only one parent has to carry out both types of tasks, their income is very limited, and the vast majority of them have to pay rent for housing located in an area of intermediate density (which tend to be far away from their place of work). Despite representing only 6 % of transport-poor households, these households are extremely vulnerable, because they have to devote a large proportion of their income to transport due to their socioeconomic and demographic characteristics. In fact, it is likely that in many cases these factors limit their ability to afford transport consumption to such an extent that they cannot afford it and fall into hidden transport poverty.
Finally, cluster 3 (which we have defined as rural older couples) also deserves special attention since the combination of rurality and age (over 65) can not only lead these households to situations of transport poverty but also to social exclusion. Lack of accessibility to transport can make it difficult for older people to access essential services such as health care and grocery stores.
In short, a cluster analysis allows us to understand the vulnerability of households to transport poverty from an intersectional perspective, identifying a set of characteristics and patterns repeated within each group. However, all of these clusters also have similarities in sharing their status as transport poor. Table 2 shows that in all clusters the average annual expenditure per household is below the average expenditure of a Spanish household (€30,860). Single-parent households stand out in particular, with yearly average expenditure (€15,314) representing half of the expenditure of an average Spanish household. Furthermore, all households dedicate a disproportionate share of their income (between 15 % and 17 %) to the consumption of transport goods and services, which amounts to approximately four times more than the average Spanish household spends (4.5 %).
## *3.2. Policy insights to alleviate transport poverty*
In this section, we show how the cluster analysis information we obtained can be used to design targeted policies to mitigate transport poverty. In other words, we use the characteristics identified in the different transport poor profiles and their intersections to design more targeted policies, taking the different levels of vulnerability observed in the previous section into account. To do so, we used one of the main
**Table 4** Design of Feasible targeted aid.
| Concept | Amount | |
|-------------------------------------------------------------------------------------------------------|--------------------------|--|
| Base aid to:
Low-income transport usersa | Base amount:
€1482.68 | |
| If any of the following conditions are met, the basic amount is
increased by the amount indicated: | Additional
amount: | |
| Live in a rural area and are under the age of 65 | x1.15 | |
| Live in a rural area and are over the age of 65 | x1.97 | |
| Being an immigrant | x1.07 | |
| Having children | x1.16 | |
| Being a single parent | x2.32 | |
| Live in a rented house | x1.14 | |
| Own a house with a mortgage | x1.06 | |
a Low income transport users are households with an income below €14,255 whose expenditure on transport goods and services is higher than the national median transport expenditure of transport consuming households, i.e. households spending more than €976.45 per year on transport.
policies implemented in Spain during the energy crisis12 in 2022 as a case study: a universal transport fuel subsidy of €0.20 per litre at the pump, which had an overall cost of €2.749 billion.13
Although this measure helped to reduce households' energy burdens and contain inflation [\(García-Miralles, 2023\)](#page-13-0), it was not targeted at vulnerable or low-income households ([OECD, 2022\)](#page-13-0) and, therefore, it was regressive. As a result, many national and international organisations, such as the IMF [\(Arregui et al., 2022](#page-13-0)), were very critical of these kinds of measures, as they subsidised the consumption of fossil fuels and were not targeted to help the most vulnerable communities. Therefore, our objective is to design different policy scenarios using the same budget as the one used to subsidise fuel during the energy crisis and explore the impact they may have on the transport poverty dimension.
[Table 3](#page-6-0) presents the scenarios. The *Reference* scenario is the policy implemented by the Spanish government: a universal fuel subsidy of 20 cents per litre. Since this policy only benefited those households that consumed fuel for transport, in scenario *Transfers-all,* we explored the impact of a broader measure: a lump sum transfer of €147 to every Spanish household, regardless of their consumption of transport fuels. This lump sum is the result of using the total cost of the *Reference* scenario policy to compensate all Spanish households. This policy is in line with other climate measures and crisis measures that have compensated every household in a country, such as the Canada Carbon Rebate introduced in Canada in 2019, or the general lump sum in the US during the financial crisis.
The previous scenarios represent broader measures that do not specifically target reducing inequality. However, many governments often implement social programmes for low-income households. Therefore, the *Low-Income* scenario distributes the available budget among the lowest 20 % of the income distribution (Q1), resulting in a transfer of €733 per household.
Although general transfers, particularly those targeting income such as the *Low-Income* scenario, may reduce inequality and represent progressive measures, they may not be able to resolve issues of transport poverty, especially for households whose vulnerability is primarily driven by mobility costs. Therefore, in the final two scenarios, we used the information provided by our transport poverty indicators to design targeted measures that address transport poverty.
**Table 5** Transport poverty index (LIHC) with and without mitigation measures.
| Scenario | LIHC index |
|---------------------------------|------------|
| Without measures | 3.58 % |
| General: Reference | 3.08 % |
| General: Transfers to all | 3.41 % |
| Targeted: Low-Income | 3.51 % |
| Targeted: Transport-poor | 0.93 % |
| Targeted: Income&Transport-poor | 2.08 % |
Consequently, the *Transport-poor* scenario shares all the policy costs between the households identified as transport poor. Since the number of households covered by this scenario is much lower than in the previous scenarios (the 670,488 households considered to be transport poor), this scenario results in significantly higher transfers per household (€4100/household). Although the amount of this transfer may seem very high, it is in line with other specific aid measures implemented to promote the energy transition in Spain, such as the aid allocated to promote sustainable mobility.14
Among our scenarios, the *Transport-poor* scenario represents the best policy for reducing transport poverty, as it targets resources directly to the households identified as transport poor, ensuring that aid is directed exclusively to those most in need. However, although the government wants to provide these transfers to households experiencing transport poverty, it is difficult to identify them from an administrative perspective. The cluster analysis information we obtained can be used to design a targeted and "feasible" policy15 to mitigate transport poverty. Therefore, in the *Income&Transport-poor* scenario, 70 % of the budget is allocated to a general transfer based on income (for those earning less than €14,255), which corresponds to a base aid amount of €1483 for potential beneficiaries (see Table 4). The remaining 30 % is distributed to complement this transfer, based on the socioeconomic characteristics (and their intersections) identified in the cluster analysis, the number of potential beneficiaries and the vulnerability levels of the different groups presented in Section [3.1](#page-4-0). The specific multiplier values were calculated to ensure the total distributed budget in this scenario matches exactly the budget allocated in the other policy scenarios, maintaining comparability across approaches. Table 4 outlines how the aid increases if specific socioeconomic characteristics are met. These amounts are cumulative; if more than one condition is met, the aid increases by an additional amount for each condition met. The values for the additional characteristics were calculated by considering the intersections of the characteristics within the various transport-poor profiles, ensuring that beneficiaries with higher levels of vulnerability (i.e., those with a larger gap to overcome transport poverty) receive greater support than those with lower vulnerability. This approach also ensures that the aid is sufficient to lift households out of transport poverty. However, these allocations are illustrative, demonstrating how policymakers could weight support while recognizing real-world programs may adjust these parameters.
Table 5 shows the impact of the different scenarios on transport poverty levels in Spain. The transport poverty level in 2022 (according to the LIHC indicator) was 3.58 %. The general fuel subsidies (*Reference* scenario) reduce the transport poverty levels to 3.08 %. The lump-sum
12 After the price of petrol and diesel exceeded the €2/litre barrier in 2022 (P´[erez, 2022\)](#page-14-0), the Spanish government applied a transitional universal transport fuel subsidy of €0.20 per litre (c/l)) ([Jefatura de Estado del Gobierno de](#page-13-0) [Espana,](#page-13-0) ˜ 2022).
13 This amount does not take into account the support to professionals that was part of the same legislative package. We estimate that, in total (with the cost of households and sectors), the cost of the policy would amount to approximately €5.911 billion.
14 In 2021 the Spanish government implemented a programme to increase the number of electric vehicles in Spain called Plan Moves III. The policy incorporated a maximum aid package equal to €7000 to purchase an electric car.
15 The government can provide these transfers based primarily on income, and then complement the aid based on other sociodemographic characteristics associated with transport poverty, based on the result from the cluster analysis. This has already been implemented in Spain in the case of the subsidies on the final price of electricity and gas for vulnerable households (bono social – a subsidised rate), which was then complemented based on other information (such as climatic zones).
transfer to all households (*Transfers-all*) reduces the transport poverty by just 0.17 %–3.41 %. This is mainly due to two reasons: i) lower-income households are not the primary users of private transport (in fact, the majority do not have a car), and ii) for those who are transport users, the amount they receive is not enough to lift them from transport-poverty. Furthermore, this policy only alleviates poverty in transport-poor households that have a lower level of vulnerability, leaving those households with a higher level of vulnerability unprotected.
While the *Reference* scenario reduces the number of transport-poor households by 0.5 %, it does not seem like the most cost-efficient or sustainable way to reduce transport poverty. [Table 5](#page-7-0) shows that the targeted measures have a significantly higher potential to alleviate transport poverty.
The measure that would reduce transport poverty levels the most would be to dedicate all the resources to transport-poor households (*transport-vulnerable scenario*). The *transport-poor* scenario can reduce transport poverty levels to 0.93 %. However, while transferring aid to transport-poor households is technically unfeasible due to the inability of public administrations to identify these households, this analysis shows that a well-designed policy that takes the characteristics of transport-poor households into account can be more effective than generalised policies or income-based policies to mitigate transport poverty in the groups most affected. Therefore, the *Income&Transportpoor* targeted policy could reduce the LIHC rate by up to 2.08 %. Moreover, the results of the *income-vulnerable* scenario are interesting, as they show that giving money to low-income households may reduce poverty in general terms but does little to reduce the transport poverty problem. In fact, it will only be reduced by 0.07 % to a level of 3.51 %.
#### **4. Discussion and policy implications**
In the current context, identifying and analysing the profiles of transport-poor households is essential. While transport is a necessary means for satisfying certain basic needs or rights, it is, in turn, one of the most polluting sectors from a climate and public health perspective. In fact, transport emissions represent around 25 % of the total greenhouse gas emissions in the EU and are one of the main sources of atmospheric pollutant emissions ([European Environment Agency, 2022](#page-13-0)). Additionally, emissions from road transport have continued to increase in recent years. As a result, experts have repeatedly pointed out the need to transform the energy system to reduce greenhouse gas emissions and comply with the commitments made in international negotiations, such as the Paris Agreement16 or the European Green Deal17 ([Rama et al.,](#page-14-0) [2022\)](#page-14-0).
Consequently, policies that penalise the use of private transport, by increasing prices, will be implemented in the coming years to promote the energy transition, particularly the decarbonisation of the transport sector. However, if these policies are not designed with a social justice perspective, we run the risk of severely disadvantaging the most vulnerable households and increasing energy and transport poverty
levels ([Alonso-Epelde et al., 2023](#page-13-0); Bohringer ¨ [et al., 2022;](#page-13-0) Tom´ [as et al.,](#page-14-0) [2023\)](#page-14-0). This implication is fundamental from a perspective of equal opportunities and social justice, since vulnerable households also have the greatest difficulties in taking advantage of the opportunities of the energy transition, primarily due to financial constraints (e.g. limited access to financing) that prevent initial investments in clean technologies, informational gaps, and housing conditions incompatible with new infrastructure requirements, such as EV charging [\(Maestre-Andr](#page-13-0)´es et al., [2019\)](#page-13-0). Moreover, improving the design of these policies by integrating mechanisms to mitigate their social impact is important from a moral and ethical point of view and as a strategy to maximise their public acceptability. The transition has to be fair, and needs to be perceived as such. Therefore, including transport-poor households in the design of policies can help prevent situations such as what happened in France with the "Yellow Vests" movement [\(Nature, 2018](#page-13-0)).
In this context, our study has identified different profiles of transport-poor households, which, as demonstrated in section [3.2](#page-6-0)., can help policymakers define potential beneficiaries or set criteria for policies and interventions aimed at reducing, or preferably eliminating, transport poverty more effectively. In fact, by identifying the specific characteristics and challenges faced by different transport-poor profiles, policymakers will be able to design policies focused on solving the specific problems of each group and allocate resources (which are limited) efficiently to improve the transportation infrastructure and services that are most likely to reduce transport poverty. For instance, rural households, such as those in Cluster 1 (rural poor couples with children) and Cluster 3 (rural older couples), would benefit most from investments in rural mobility infrastructure, aid for purchasing electric vehicles and the promotion of carpooling or shared mobility solutions. The fact that most of these households require private transport to meet their mobility needs highlights the dual challenge of maintaining accessibility to basic services while pursuing environmental goals. By contrast, measures such as free or subsidised public transport passes would be a better choice for Clusters 5 (immigrants) and 6 (single-parent families), who often reside in urban or intermediate-density areas. By leveraging the findings from our cluster analysis, policymakers could design actions that go beyond generic solutions and address the multidimensional nature of transport poverty. This would lead to more equitable and efficient use of resources while ensuring that vulnerable populations are prioritised in the transition towards sustainable and inclusive transport systems.
More specifically, at European level, the results presented in the study may be of special interest within the framework of the "Fit for 55" package, which contains a number of legislative proposals for reviewing the entire EU climate and energy framework by 2030 and adapting it to the new emissions reduction target18 approved by the European Commission (EC). Among them, the proposal for a new Energy Taxation Directive (ETD) or the extension of the Emissions Trading Scheme (ETS) to the building and transport sector (also known as ETS2) [\(European](#page-13-0) [Commision, 2021a; 2021b](#page-13-0)) are expected to increase transport prices. As far as the ETD is concerned, according to the EC, the Energy Taxation Directive 2003/96/EC is outdated and does not reflect the EU's climate and energy commitment and, therefore, it has been proposed that tax rates be updated. The proposal introduces a new minimum tax rate structure based on the energy content and environmental performance of fossil fuels and electricity. The tax base will also be expanded to include some products that had previously escaped the EU energy tax framework and some of the current exemptions and reductions will be eliminated. In this way, the new system will ensure that the most polluting fuels are the most taxed after the transition period. Meanwhile, the ETS2 aims to reduce greenhouse gas emissions by extending the
16 At COP21, the Paris Agreement was adopted, the first universal and legally binding climate agreement. It agreed to keep the increase in the average temperature of the earth below 2 ◦C and to do everything possible to limit this increase to 1.5 ◦C. The agreement, which came into force in 2016, sought to reach the maximum emissions as soon as possible to achieve a balance between emissions and absorptions by 2050. To this end, countries would have to present national climate action plans (NDCs) every 5 years and a mechanism was established to reinforce these voluntary contributions from the countries. Likewise, the Paris Agreement addressed issues such as adaptation, aid to developing countries and the transfer of technology and capabilities.
17 The Green Deal promoted by the European Union, sets out a detailed vision to make Europe the first climate-neutral continent by 2050, safeguard biodiversity, establish a circular economy and eliminate pollution, while boosting the competitiveness of European industry and ensuring a fair transition for the regions and workers affected [\(European Commission, 2019\)](#page-13-0).
18 The new commitments acquired propose increasing the net emissions reduction target from 40 % to at least 55 % compared to 1990 levels [\(European](#page-13-0) [Commission, 2021\)](#page-13-0).
current ETS and effectively putting a price on carbon emissions. Consequently, under the ETS2, entities in the transport sector must purchase allowances for their emissions, creating a financial incentive to adopt cleaner technologies and reduce carbon output.
Although these are some of the tools required to promote the decarbonisation of the European energy system, the EC is aware that these types of policies may have negative impacts on households' wellbeing and increase inequalities (Bohringer ¨ [et al., 2022;](#page-13-0) [Feindt et al.,](#page-13-0) [2021; OECD, 2011;](#page-13-0) [Piketty and Saez, 2014](#page-14-0)), as they are likely to lead to higher fuel prices as the costs will be passed on to consumers. Therefore, the EC has proposed the creation of the Social Climate Fund19 (SCF), which aims to channel resources to alleviate the adverse effects of these policies on some of the most vulnerable groups, including vulnerable transport users. However, the European Parliament recognises the difficulty in effectively channelling these resources towards transport-vulnerable groups, due to the knowledge gap in transport poverty. Therefore, this study constitutes a timely contribution, as it provides the information required to design Social Climate Plans effectively20, which all countries will have to prepare to access the SCF resources available. In this sense, although the study is focused on the specific case of Spain, it is valuable internationally as it presents a methodology that is easily replicable in other countries.
Our analysis in section [3.2.](#page-6-0) is particularly interesting in this context, as the SCF opens the door to channelling resources to vulnerable households through direct aid.21 Our findings not only demonstrate that targeted actions are more effective in alleviating transport poverty, but also provide a concrete example of how to design politically feasible support measures by accounting for household characteristics and their intersections. In the case of Spain, the results from the cluster analysis highlight the profiles of households that should be prioritised as beneficiaries of direct aid. These include older couples living in rural areas, lower- and middle-income families with children in rural settings, single-parent households and immigrant families residing in urban environments.
Moreover, our results show that well-designed aid policies can reduce inequality and transport poverty and this can be very beneficial in the short and medium term due to their ability to reduce social opposition to unpopular measures (such as raising taxes on fuel). In fact, as shown in the study carried out by Tomas ´ [et al. \(2023\)](#page-14-0), designing policies that include compensation for the most affected groups may not only be beneficial from a distributional point of view, but may also help the different socioeconomic agents not to position themselves directly against policies that seek to decarbonise the transport sector. As we show, identifying different transport poverty profiles helps to design targeted aid, thereby significantly reducing transport poverty from an affordability perspective. In addition, from the results in section [3.2](#page-6-0). we can conclude that these policies (e.g., those analysed in *Transport-poor* and *Income&Transport-poor scenarios*) are more efficient, as they achieve more significant transport poverty reduction than generalised policies
(such as those analysed in the *Reference* and *Transfer-all* scenarios) or policies that target households by income alone (*Low-Income scenario*), while also ensuring appropriate uses of the public budget. Although economic factors are essential when it comes to addressing transport poverty, this phenomenon is also largely determined by other socio-demographic characteristics of households.
However, in the current context of the energy transition, it is also essential to implement policies, plans, and strategies that contribute to eliminating the structural factors driving transport poverty so that in the medium-long term, direct aid is no longer needed, and their elimination can proceed without harming the households that were receiving this aid. Some households could become dependent on this aid, so the goal should be to offer political alternatives which, in the medium-long term, reduce the vulnerability of these households through other means.
These alternatives could come from income, for example, from attempting to reduce income inequality, introducing redistributive policies such as higher minimum wages (which would be especially beneficial for the groups identified in the study, as their disposable income, which is shown in [Tables 2](#page-6-0) and is significantly below the average of Spanish households), or implementing policies to reduce unemployment. Furthermore, addressing transport poverty's root causes requires tackling the housing-transport nexus. In Spain, where gentrification has concentrated affordable housing in car-dependent peripheries, coordinated policies are critical - combining anti-displacement measures (e.g., rent controls, social housing quotas in transit-accessible areas) with transport solutions to break the cycle of energy vulnerability. Likewise, alternatives could come from transforming the transport system. In the transition process from a model focused on private vehicles to a more sustainable model, the new transportation system should be designed to build a more affordable, accessible model that responds to the needs of all citizens. Some of the effective courses of action could include promoting public transport (through price discounts, free passes for the vulnerable transport users identified in the paper or from extending and/or improving bus, tram or commuter lines), implementing public bicycle services or shared transportation initiatives (carsharing or carpooling).
Transforming the transport system can be challenging in rural environments due to the geographic and demographic characteristics of each region. At the same time, it is essential to help mitigate transport poverty since, as we have identified in this study, it is a characteristic shared by most profiles identified as transport-poor households. Therefore, the various national and European funds associated with the energy transition represent an opportunity to begin to meet this challenge. However, to successfully address it from a social justice perspective, it will be necessary to reorient public investments22 towards more sustainable transportation alternatives appropriate to the needs of the groups identified, such as transport accessibility improvement programmes, social leasing programmes and aid for purchasing electric vehicles in rural areas.
In conclusion, the fight against transport poverty requires diverse policies and interventions, some aimed at mitigating the vulnerability of households in the short term and others at transforming the structural factors that reproduce this type of poverty. It is important to have a framework or strategy that defines the national objectives and associated lines of action in order to implement this range of measures effectively. In Spain, the Ministry of Transport has opened the public consultation period for creating a National Strategy to Eliminate
19 The SCF aims to provide funding to Member States to support policies that address potential social impacts. This must be achieved through temporary direct aid and investments aimed at reducing dependence on fossil fuels in the medium and long term. The SCF will have a financial endowment of 72.2 billion euros and will be financed with part of the collection from the ETS2.
20 The document of the European Union Council setting out the criteria for creating the SCF [\(European Commission, 2022](#page-13-0)) for the period 2025–2032 states that Member States must develop their own Social Climate Plans. These plans will include (article 4.1.C) an "estimation of the possible effects of this increase in prices on households and, in particular, on the incidence of energy poverty, on microenterprises and on transport users, which includes, in particular, an estimation and identification of vulnerable households, vulnerable micro-enterprises and vulnerable transport users".
21 While fuel subsidies could theoretically increase driving and undermine climate goals, real-world programs like the EU Social Climate Fund avoid this by mandating support for sustainable mobility options.
22 For example, public investment in the transport sector in Spain in recent decades has been heavily skewed towards expanding and improving infrastructure. More specifically, large public investments have been allocated to the construction of high-speed networks (motorways or the AVE railway network) or the expansion of airports, thus reducing the resources allocated to the services that the majority of the population use on a daily basis: local and/or regional public transport lines (bus, tram, metro, train, etc.).
Transport Poverty, similar to the one already in place for energy poverty. Consequently, this study could also be useful when it comes to defining the target audience of the policies and when thinking about specific lines of action to mitigate the vulnerability of different groups.
Although identifying the profiles of households most affected by transport poverty could help in the development of more effective policies and strategies to act against this phenomenon, the development of complementary research that expands knowledge and debate regarding its causes and consequences is essential. In the same way, it is essential to study what types of policies are most effective in reducing the vulnerability of transport-poor households in each context, as the strategies that work in the Spanish or European contexts (in which our study is framed) may not be equally impactful in other regions or socioeconomic environments.
## **5. Conclusions**
Addressing issues related to transport poverty in the current context of energy transition and high prices of energy goods (particularly those driven by climate policies targeting private transport23) is increasingly relevant, given its fundamental role in facilitating access to basic needs and rights of citizens. In this sense, identifying and analysing the profiles of transport-poor households is crucial to promoting climate and energy policies that help the most vulnerable households take advantage of the opportunities provided by the energy transition and the transformation of the transport system. In fact, the design of fairer decarbonisation measures promotes their political and social acceptability, avoiding conflicts such as those that occurred in France due to the increase in fuel taxes (i.e., the yellow vest protests).
In this study, six profiles of transport-poor households were identified for the case study of Spain through a hierarchical clustering analysis, based on the LIHC indicator (i.e., households that spend a high proportion of their income on their transport needs and who also have low income to cover these expenses). The profiles were identified based on the socioeconomic and demographic characteristics (both of the household and the person of reference of the household) collected in [Table 1.](#page-4-0) The following groups were identified among the transport-poor households: 1) rural poor couples with children (31 %), 2) mortgaged households (22 %), 3) rural older couples (17 %), 4) lower-middle income couples with children (15 %), 5) immigrants (9 %) and 6) singleparent families (6 %).
The analysis of the groups shows how certain factors can determine their condition of being transport-poor and how their intersection causes the household's vulnerability to increase. These factors include income, location and housing tenure, or other social factors such as composition of the household, origin, or gender of the person of reference of the household. Thus, low or lower-middle-income households, those located in rural areas, those that have mortgage or rental expenses, those that are made up of couples with children or single parents, or those that are made up of immigrants are especially vulnerable to transport poverty.
Furthermore, the study shows how the vulnerability of some of these groups is significantly higher due to the intersection of various socioeconomic and demographic characteristics, such as low-income couples with children living in rural environments or single-parent families. Thus, the study confirms that, due to the characteristics of transport poverty, the most efficient way to mitigate household vulnerability is by designing targeted policies that go beyond income and include additional criteria that seek to alleviate the vulnerability of the groups identified in the study. In fact, through the case study proposed in Section [3.2.](#page-6-0), we conclude that a well-designed, technically feasible policy, which takes the characteristics of transport-poor households into account, can significantly reduce transport poverty, as it is more efficient than generalised policies or income-based policies when it comes to mitigating transport poverty in the most affected groups.
Although these results are relevant, there are points that could be improved in future research, such as reducing the limitations of the indicators to measure transport poverty mentioned in Section [2.1.](#page-1-0) or try to reduce biases that arise from the data. For example, in future research it would be interesting to analyse hidden transport poverty, since many low-income households find their access to education, health and work limited because they cannot cope with the high transport costs that they need to pay to allow them to perform these activities daily. This is a reality that disproportionately affects women, and if we only use indicators such as the LIHC, we can underestimate the vulnerability of women to this type of poverty.
Moreover, the results presented in the study are of particular interest at European level within the framework of the "Fit for 55" package and specifically for developing Social Climate Plans, in which Member States must specify the measures they will take to be able to access the SCF. However, the methodology can be replicated for various countries or regions, which means that it can be useful in very diverse frameworks. Likewise, in Spain, it can be useful when it comes to developing a National Strategy to Eliminate Transport Poverty, which could act as a framework for implementing cross-sector policies and interventions in a coordinated manner. Specially, it could be used to design measures or aid to mitigate the vulnerability of households in the short term. Nevertheless, eliminating transport poverty requires the implementation of medium- and long-term measures to address the structural factors that perpetuate this issue. In this regard, public investments aimed at promoting modal shifts (such as investments in public transport and developing and enhancing public cycle networks) and improving accessibility (for example, by promoting carpooling or car-sharing programs in rural areas) are critically important. These measures must genuinely benefit the groups identified in this study to ensure a sustainable and equitable transformation.
In conclusion, identifying transport poverty profiles is essential when it comes to developing targeted interventions that can mitigate transport vulnerability and improve the quality of life of those affected by transport poverty most efficiently. In this sense, the political will to develop short- and long-term cross-cutting policies that address transport poverty in different ways will be essential.
## **CRediT authorship contribution statement**
**E. Alonso-Epelde:** Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. **X. García-Muros:** Writing – review & editing, Validation, Supervision, Project administration, Funding acquisition. **M. Gonzalez-Eguino:** ´ Writing – review & editing, Validation, Supervision, Project administration, Funding acquisition.
## **Declaration of generative AI and AI-assisted technologies in the writing process**
During the preparation of this work the authors used DeepSeek in order to improve language and readability. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
### **Declaration of competing interest**
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
23 While fuel prices may fluctuate due to multiple factors (geopolitical conflicts, market dynamics …), this study focuses specifically on policy-induced increases from climate measures like the EU's ETS2 and Energy Taxation Directive revision. These structural drivers are most relevant for designing equitable compensation mechanisms, though our methodology remains applicable regardless of price determinants.
#### **Acknowledgement**
This research has received funding from the European Union's Horizon Europe Research and Innovation Programme under grant agreement No 101069880 – AdJUST, Advancing the understanding of challenges, policy options and measures to achieve a JUST EU energy transition; by María de Maeztu Excellence Unit 2023–2027 Ref. CEX2021–001201-M, funded by MCIN/AEI/10.13039/ 501100011033; and by the Basque Government through the BERC 2022–2025 program.
## **Annex A.**
The concept of transport poverty lacks a universally accepted definition, largely due to the complex and context-dependent nature of transport needs. Transport poverty is inherently linked to mobility, which affects access to essential services and socioeconomic activities. This relationship is often shaped by individual circumstances, geographical context and temporal factors. For example, transport poverty can disproportionately affect individuals within a household, with a notable gender component, highlighting its intricate and personal nature ([Booth et al., 2000;](#page-13-0) [Robinson and](#page-14-0) [Thagesen, 2017](#page-14-0)).
Early studies, such as those by [Wachs and Kumagai \(1973\),](#page-14-0) emphasised transport accessibility as a crucial factor in determining quality of life, highlighting it as an indicator of social and economic inequality. The Social Exclusion Unit (2003) further underscored the role of transport in social exclusion, stressing its importance in accessing education, employment, healthcare and other essential activities necessary for full societal participation.
However, the academic and policy discourse has struggled with the consistent use of terminology related to transport poverty. To address this inconsistency, [Lucas et al. \(2016\)](#page-13-0) proposed a unified lexicon to conceptualise it. They introduced a comprehensive definition of transport poverty, outlining several conditions under which an individual may be considered transport vulnerable, including the lack of accessible mobility options, inability to maintain a reasonable quality of life due to transport inadequacies, excessive expenditure on mobility, time poverty and exposure to unsafe travel conditions. The framework presented by [Lucas et al. \(2016\)](#page-13-0) also highlighted key notions that underpin transport poverty, such as transport affordability, poor mobility options, poor accessibility options and exposure to transport externalities. Each of these aspects addresses different dimensions of transport poverty, from financial constraints to the lack of adequate infrastructure and the adverse effects of transport systems.
Additionally, a number of studies have proposed metrics to quantify transport poverty, typically focusing on affordability, mobility choices and accessibility [\(Allen and Farber, 2019](#page-13-0); [Benevenuto and Caulfield, 2020; Berry et al., 2016; Carruthers et al., 2005; Currie et al., 2010; Dodson and Sipe,](#page-13-0) [2007; Gomide et al., 2005; Kamruzzaman and Hine, 2012](#page-13-0); [Lovelace and Philips, 2014](#page-13-0); [Mattioli, 2017](#page-13-0); [Mattioli et al., 2016](#page-13-0), [2018](#page-13-0); [Salon and Gulyani,](#page-14-0) [2010; Shen, 1998](#page-14-0); [Sustrans, 2012;](#page-14-0) [Tao et al., 2020\)](#page-14-0). However, most existing measures are highly specific to particular contexts and are difficult to replicate over time or across different regions, thereby limiting their usefulness for broader policy-making. Consequently, there is a growing need for standardised, replicable metrics that can track transport poverty over time and in various contexts using publicly available data.
The Low Income High Cost (LIHC) metric proposed by [Alonso-Epelde et al. \(2023\)](#page-13-0) emerges as a particularly effective measure of transport poverty by addressing these challenges. The LIHC metric assesses transport poverty by evaluating both the financial burden of transport costs and the remaining disposable income after these costs have been considered. This makes it a valuable tool for identifying households that are not only burdened by transport costs but are also at risk of falling into poverty because of them. Additionally, the use of the HBS, with its detailed, standardised data across European countries, makes it possible to apply the LIHC metric consistently across different regions and over time [\(Alonso-Epelde et al.,](#page-13-0) [2023\)](#page-13-0). This replicability significantly enhances its usefulness for policymakers, enabling them to monitor transport poverty trends effectively and develop targeted interventions to mitigate transport-related social inequality and improve access to essential services.
## *Annex B.*
**Fig. B1.** Dendrogram: Agglomerative clustering
**Fig. B2.** Silhouette assessment: Agglomerative clustering
#### **Annex C.**
**Table C1** Distribution of households in each cluster across different socioeconomic and demographic variables
| | Cluster 1 | Cluster 2 | Cluster 3 | Cluster 4 | Cluster 5 | Cluster 6 |
|-------------------------------|-----------|-----------|-----------|-----------|-----------|-----------|
| Q_Q1 | 35.4 % | 69.8 % | 26.5 % | 56.0 % | 20.4 % | 80.6 % |
| Q_Q2 | 57.7 % | 29.8 % | 58.3 % | 41.6 % | 77.6 % | 16.7 % |
| Q_Q3 | 6.9 % | 0.4 % | 13.6 % | 2.4 % | 2.0 % | 2.8 % |
| Q_Q4 | 0.0 % | 0.0 % | 0.8 % | 0.0 % | 0.0 % | 0.0 % |
| Q_Q5 | 0.0 % | 0.0 % | 0.8 % | 0.0 % | 0.0 % | 0.0 % |
| Z_Urban | 22.3 % | 28.1 % | 32.6 % | 29.6 % | 91.8 % | 8.3 % |
| Z_Intermediate | 29.2 % | 17.4 % | 41.7 % | 28.0 % | 0.0 % | 69.4 % |
| Z_Rural | 48.5 % | 54.5 % | 25.8 % | 42.4 % | 8.2 % | 22.2 % |
| TH_Elder alone | 15.4 % | 0.0 % | 0.0 % | 0.0 % | 0.0 % | 0.0 % |
| TH_Single parent | 7.7 % | 2.9 % | 6.8 % | 10.4 % | 12.2 % | 61.1 % |
| TH_Others | 10.8 % | 12.0 % | 5.3 % | 16.0 % | 4.1 % | 11.1 % |
| TH_Couples with children | 19.2 % | 65.3 % | 28.0 % | 44.0 % | 69.4 % | 2.8 % |
| TH_Older couples | 42.3 % | 0.0 % | 0.0 % | 4.0 % | 2.0 % | 0.0 % |
| TH_ Couples without children | 0.0 % | 15.3 % | 23.5 % | 14.4 % | 10.2 % | 2.8 % |
| TH_Single person | 4.6 % | 4.5 % | 36.4 % | 11.2 % | 2.0 % | 22.2 % |
| R_Rented | 5.4 % | 24.0 % | 21.2 % | 12.0 % | 67.3 % | 38.9 % |
| R_Relinquish | 0.0 % | 2.1 % | 2.3 % | 2.4 % | 0.0 % | 2.8 % |
| R_Ownership with mortgage | 6.2 % | 21.5 % | 57.6 % | 36.8 % | 24.5 % | 36.1 % |
| R_ Ownership without mortgage | 88.5 % | 52.5 % | 18.9 % | 48.8 % | 8.2 % | 22.2 % |
| OS_Not provided | 15.4 % | 7.4 % | 40.2 % | 20.0 % | 14.3 % | 77.8 % |
| OS _Employed | 0.0 % | 22.7 % | 38.6 % | 27.2 % | 65.3 % | 2.8 % |
| OS _Unemployed | 80.0 % | 17.4 % | 8.3 % | 17.6 % | 4.1 % | 19.4 % |
| OS_One employed | 4.6 % | 52.5 % | 12.9 % | 35.2 % | 16.3 % | 0.0 % |
| A_Adult | 7.7 % | 97.1 % | 98.5 % | 96.0 % | 93.9 % | 80.6 % |
| A_Young | 0.0 % | 1.2 % | 1.5 % | 4.0 % | 4.1 % | 11.1 % |
| A_Elder | 92.3 % | 1.7 % | 0.0 % | 0.0 % | 2.0 % | 8.3 % |
| G_Man | 83.1 % | 94.2 % | 80.3 % | 21.6 % | 83.7 % | 5.6 % |
| G_Woman | 16.9 % | 5.8 % | 19.7 % | 78.4 % | 16.3 % | 94.4 % |
| C_Spain | 99.2 % | 78.9 % | 82.6 % | 99.2 % | 28.6 % | 91.7 % |
| C_Rest Europe | 0.8 % | 0.0 % | 0.8 % | 0.0 % | 2.0 % | 0.0 % |
| C_Rest of world | 0.0 % | 16.5 % | 12.9 % | 0.8 % | 63.3 % | 5.6 % |
| C_EU27 | 0.0 % | 4.5 % | 3.8 % | 0.0 % | 6.1 % | 2.8 % |
| S_Post-secondary education | 10.0 % | 24.0 % | 44.7 % | 36.8 % | 38.8 % | 25.0 % |
| S_Secondary education | 31.5 % | 50.0 % | 31.8 % | 36.8 % | 30.6 % | 44.4 % |
| S_Primary education | 50.0 % | 19.4 % | 6.8 % | 7.2 % | 6.1 % | 11.1 % |
| S_No education | 6.2 % | 1.2 % | 2.3 % | 0.0 % | 2.0 % | 2.8 % |
| S_Higher education | 2.3 % | 5.4 % | 14.4 % | 19.2 % | 22.4 % | 16.7 % |
Note: S – Education level completed by the reference person of the household, C – Country of birth of the reference person of the household, G – Gender of the reference person of the household, A – Age of the reference person of the household, OS – Occupational situation of the reference person in the household, R – Tenure status of the household's primary residence, TH – Household type, Z – Level of rurality of the household's area of residence, Q – Income quintile.
## **Data availability**
Data will be made available on request.
#### **References**
- Adom, P.K., Amuakwa-Mensah, F., Agradi, M.P., Nsabimana, A., 2021. Energy poverty, development outcomes, and transition to green energy. Renew. Energy 178, 1337–1352. [https://doi.org/10.1016/j.renene.2021.06.120.](https://doi.org/10.1016/j.renene.2021.06.120)
- Adusah-Poku, F., Takeuchi, K., 2019. Determinants and welfare impacts of rural electrification in Ghana. Energy Sustain. Dev. 52, 52–62. [https://doi.org/10.1016/j.](https://doi.org/10.1016/j.esd.2019.07.004) [esd.2019.07.004](https://doi.org/10.1016/j.esd.2019.07.004).
- Allen, J., Farber, S., 2019. Sizing up transport poverty: a national scale accounting of low-income households suffering from inaccessibility in Canada, and what to do about it. Transp. Policy 74, 214–223. [https://doi.org/10.1016/j.](https://doi.org/10.1016/j.tranpol.2018.11.018) [tranpol.2018.11.018.](https://doi.org/10.1016/j.tranpol.2018.11.018)
- Alonso-Epelde, E., García-Muros, X., Gonz´ alez-Eguino, M., 2023. Transport poverty indicators: a new framework based on the household budget survey. Energy Policy 181, 113692. .
- [Arregui, N., Celasun, O., Iakova, D., Mineshima, A., Mylonas, V., Toscani, F., Wong, Y.C.,](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref5) [Zeng, L., Zhou, J., 2022. Targeted, Implementable, and Practical Energy Relief](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref5) [Measures for Households in Europe. IMF.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref5)
- Bednar, D.J., Reames, T.G., 2020. Recognition of and response to energy poverty in the United States. Nat. Energy 5, 432–439. [https://doi.org/10.1038/s41560-020-0582-](https://doi.org/10.1038/s41560-020-0582-0)
- Belaïd, F., 2018. Exposure and risk to fuel poverty in France: examining the extent of the fuel precariousness and its salient determinants. Energy Policy 114, 189–200. .
- Benevenuto, R., Caulfield, B., 2020. Measuring access to urban centres in rural Northeast Brazil: a spatial accessibility poverty index. J. Transport Geogr. 82, 102553. [https://](https://doi.org/10.1016/j.jtrangeo.2019.102553) [doi.org/10.1016/j.jtrangeo.2019.102553.](https://doi.org/10.1016/j.jtrangeo.2019.102553)
- Berry, A., Jouffe, Y., Coulombel, N., Guivarch, C., 2016. Investigating fuel poverty in the transport sector: toward a composite indicator of vulnerability. Energy Res. & Soc. Sci. Energy demand mobil. domest. life: new insights energy just. 18, 7–20. [https://](https://doi.org/10.1016/j.erss.2016.02.001) [doi.org/10.1016/j.erss.2016.02.001](https://doi.org/10.1016/j.erss.2016.02.001).
- [Boardman, B., 1991. Fuel Poverty: from Cold Homes to Affordable Warmth. Belhaven](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref10) [Press, London](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref10).
- Boeri, A., Gianfrate, V., Boulanger, S.O.M., Massari, M., 2020. Future design approaches for energy poverty: users profiling and services for No-Vulnerable condition. Energies 13, 2115. .
- Bohringer, ¨ C., García-Muros, X., Gonzalez-Eguino, ´ M., 2022. Who bears the burden of greening electricity? Energy Econ. 105, 105705. [https://doi.org/10.1016/j.](https://doi.org/10.1016/j.eneco.2021.105705) [eneco.2021.105705](https://doi.org/10.1016/j.eneco.2021.105705).
- [Booth, D., Hanmer, L., Lovell, E., 2000. Poverty and Transport -. Report for the World](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref13) [Bank. Overseas Development Institute.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref13)
- Bradshaw, J., Hutton, S., 1983. Social policy options and fuel poverty. J. Econ. Psychol. 3, 249–266. [https://doi.org/10.1016/0167-4870\(83\)90005-3.](https://doi.org/10.1016/0167-4870(83)90005-3)
- [Carruthers, R., Dick, M., Saurkar, A., 2005. Affordability of Public Transport in](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref15) [Developing Countries. World Bank, Washington, DC.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref15)
- Cass, N., Shove, E., Urry, J., 2005. Social exclusion, mobility and access. Sociol. Rev. 53, 539–555. [https://doi.org/10.1111/j.1467-954X.2005.00565.x.](https://doi.org/10.1111/j.1467-954X.2005.00565.x)
- Castano-Rosa, ˜ R., Solís-Guzman, ´ J., Rubio-Bellido, C., Marrero, M., 2019. Towards a multiple-indicator approach to energy poverty in the European union: a review. Energy Build. 193, 36–48. [https://doi.org/10.1016/j.enbuild.2019.03.039.](https://doi.org/10.1016/j.enbuild.2019.03.039)
- CNT, 2012. Housing+Transportation Affordability Index, Center for Neighborhood Technology. .
- [Coote, A., Percy, A., 2020. The Case for Universal Basic Services. John Wiley](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref18) & Sons. Currie, G., Richardson, T., Smyth, P., Vella-Brodrick, D., Hine, J., Lucas, K., Stanley, Janet, Morris, J., Kinnear, R., Stanley, John, 2010. Investigating links between transport disadvantage, social exclusion and well-being in Melbourne – updated results. Res. Transport. Econ. Reform. Publ.Transp. through. World 29, 287–295. [https://doi.org/10.1016/j.retrec.2010.07.036.](https://doi.org/10.1016/j.retrec.2010.07.036)
- Dodson, J., Sipe, N., 2007. Oil vulnerability in the Australian city: assessing socioeconomic risks from higher urban fuel prices. Urban Stud. 44, 37–62. [https://](https://doi.org/10.1080/00420980601023810) [doi.org/10.1080/00420980601023810](https://doi.org/10.1080/00420980601023810).
- Dong, K., Ren, X., Zhao, J., 2021. How does low-carbon energy transition alleviate energy poverty in China? A nonparametric panel causality analysis. Energy Econ. 103, 105620. [https://doi.org/10.1016/j.eneco.2021.105620.](https://doi.org/10.1016/j.eneco.2021.105620)
- [European Commision, 2021a. Proposal for a COUNCIL DIRECTIVE Restructuring the](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref22) [Union Framework for the Taxation of Energy Products and Electricity \(Recast\) COM/](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref22) [2021/563 Final](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref22).
- European Commission, 2021b. Possible extension of the EU Emissions Trading System (ETS) to cover emissions from the use of fossil fuels in particular in the road transport and the buildings sector – Final report. Publications Office of the European Union, 2021. .
- [European Commission, 2022. Ad Hoc Working Party on the Social Climate Fund \(AHWP](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref24) [SCF\).](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref24)
- European Commission, 2021. 2030 climate & energy framework [WWW Document]. [htt](https://climate.ec.europa.eu/eu-action/climate-strategies-targets/2030-climate-energy-framework_en) [ps://climate.ec.europa.eu/eu-action/climate-strategies-targets/2030-climate](https://climate.ec.europa.eu/eu-action/climate-strategies-targets/2030-climate-energy-framework_en)[energy-framework\\_en](https://climate.ec.europa.eu/eu-action/climate-strategies-targets/2030-climate-energy-framework_en), 10.13.23.
- [European Commission, 2019. Communication on the European Green Deal.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref26) [European Environment Agency, 2022. Transport and Environment Report 2022](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref27).
- Feczko, E., Fair, D.A., 2020. Methods and challenges for assessing heterogeneity. Biol. Psychiatry 88, 9–17. .
Feindt, S., Kornek, U., Labeaga, J.M., Sterner, T., Ward, H., 2021. Understanding regressivity: challenges and opportunities of European carbon pricing. Energy Econ. 103, 105550. [https://doi.org/10.1016/j.eneco.2021.105550.](https://doi.org/10.1016/j.eneco.2021.105550)
- [Forgy, E.W., 1965. Cluster analysis of multivariate data : efficiency versus](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref30) [interpretability of classifications. Biometrics 21, 768](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref30)–769.
- Foster, J., Greer, J., Thorbecke, E., 1984. A class of decomposable poverty measures. Econometrica 52, 761–766. [https://doi.org/10.2307/1913475.](https://doi.org/10.2307/1913475)
- Gao, C.X., Dwyer, D., Zhu, Y., Smith, C.L., Du, L., Filia, K.M., Bayer, J., Menssink, J.M., Wang, T., Bergmeir, C., Wood, S., Cotton, S.M., 2023. An overview of clustering methods with guidelines for application in mental health research. Psychiatry Res. 327, 115265. [https://doi.org/10.1016/j.psychres.2023.115265.](https://doi.org/10.1016/j.psychres.2023.115265)
- García Miralles, E., 2023. Medidas de apoyo frente a la crisis energ´etica y al repunte de la inflacion: ´ un an´ alisis del coste y de los efectos distribucionales de algunas de las actuaciones desplegadas según su grado de focalizacion. ´ Boletín Economico ´ 15. [https://doi.org/10.53479/29651.](https://doi.org/10.53479/29651)
- Gomide, Alexandre, Leite, Sabina Augusta Kauark, Rebelo, Jorge M., 2005. Public transport and urban poverty : a synthetic index of adequate service (Vol. 1 of 2) : Main report (English). World Bank, Washington, DC. [http://documents.worldbank.](http://documents.worldbank.org/curated/en/804921467998799407) [org/curated/en/804921467998799407.](http://documents.worldbank.org/curated/en/804921467998799407)
- Gonzalez-Eguino, ´ M., 2015. Energy poverty: an overview. Renew. Sustain. Energy Rev. 47, 377–385..
- Gonzalez-Eguino, ´ M., García-Muros, X., Moyano-Reina, M., Alonso-Epelde, E., Rom´ an, M.V., Tom´ as, M., 2023. An´ alisis del impacto distributivo de las medidas para contener la factura energ´etica de los hogares. Observatorio de la Transicion ´ Energ´etica y la Accion ´ Clim´ atica (OTEA). [https://doi.org/10.5281/](https://doi.org/10.5281/zenodo.10454691) [zenodo.10454691](https://doi.org/10.5281/zenodo.10454691).
- Goodman, A., Oldfield, Z., 2004. Permanent differences? Income and expenditure inequality in the 1990s and 2000s (Research Report No. R66). IFS Report. [https://](https://doi.org/10.1920/re.ifs.2004.0066) [doi.org/10.1920/re.ifs.2004.0066](https://doi.org/10.1920/re.ifs.2004.0066).
- Gower, J.C., 1966. Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53, 325–338. [https://doi.org/10.1093/biomet/](https://doi.org/10.1093/biomet/53.3-4.325) [53.3-4.325](https://doi.org/10.1093/biomet/53.3-4.325).
- Halkos, G.E., Gkampoura, E.-C., 2021. Coping with energy poverty: measurements, drivers, impacts, and solutions. Energies 14, 2807. [https://doi.org/10.3390/](https://doi.org/10.3390/en14102807) [en14102807](https://doi.org/10.3390/en14102807).
- [Hills, J., 2012. Getting the Measure of Fuel Poverty: Final Report of the Fuel Poverty](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref40) [Review \(Monograph No. 72\). Centre for Analysis of Social Exclusion, London, UK](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref40).
- Jain, A.K., 2010. Data clustering: 50 years beyond K-means. Pattern Recog. Lett. Award winn. paper. the 19th Int. Conf. Pattern Recog. (ICPR) 31, 651–666. [https://doi.org/](https://doi.org/10.1016/j.patrec.2009.09.011) [10.1016/j.patrec.2009.09.011.](https://doi.org/10.1016/j.patrec.2009.09.011)
- [Jefatura de Estado del Gobierno de Espana,](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref42) ˜ 2022. Real Decreto-ley 6/2022, De 29 De [Marzo, Por El Que Se Adoptan Medidas Urgentes En El Marco Del Plan Nacional De](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref42) [Respuesta a Las Consecuencias Economicas](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref42) ´ Y Sociales De La Guerra En Ucrania.
- Johnson, S.C., 1967. Hierarchical clustering schemes. Psychometrika 32, 241–254. .
- Kamruzzaman, Md, Hine, J., 2012. Analysis of rural activity spaces and transport disadvantage using a multi-method approach. Transp. Policy 19, 105–120. [https://](https://doi.org/10.1016/j.tranpol.2011.09.007) [doi.org/10.1016/j.tranpol.2011.09.007.](https://doi.org/10.1016/j.tranpol.2011.09.007)
- [Kaufman, L., Rousseeuw, P.J., 2009. Finding Groups in Data: an Introduction to Cluster](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref45) [Analysis. John Wiley](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref45) & Sons.
- Kenyon, S., Rafferty, J., Lyons, G., 2003. Social exclusion and transport in the UK: a role for virtual accessibility in the alleviation of mobility-related social exclusion? J. Soc. Pol. 32, 317–338. [https://doi.org/10.1017/S0047279403007037.](https://doi.org/10.1017/S0047279403007037)
- Lovelace, R., Philips, I., 2014. The 'oil vulnerability' of commuter patterns: a case study from Yorkshire and the Humber, UK. Geoforum 51, 169–182. [https://doi.org/](https://doi.org/10.1016/j.geoforum.2013.11.005) [10.1016/j.geoforum.2013.11.005](https://doi.org/10.1016/j.geoforum.2013.11.005).
- Lucas, K., 2019. A new evolution for transport-related social exclusion research? J. Transp. Geograp. Celebrat. 25 years J. Transp. Geograp. 81, 102529. [https://doi.](https://doi.org/10.1016/j.jtrangeo.2019.102529) [org/10.1016/j.jtrangeo.2019.102529.](https://doi.org/10.1016/j.jtrangeo.2019.102529)
- Lucas, K., 2012. Transport and social exclusion: where are we now? Transport policy. Urban Transp. Initiat. 20, 105–113. [https://doi.org/10.1016/j.tranpol.2012.01.013.](https://doi.org/10.1016/j.tranpol.2012.01.013) Lucas, K., Mattioli, G., Verlinghieri, E., Guzman, A., 2016. Transport poverty and its adverse social consequences [WWW Document]. [https://doi.org/10.1680/jtran.15.](https://doi.org/10.1680/jtran.15.00073) [00073.](https://doi.org/10.1680/jtran.15.00073)
- Maestre-Andr´es, S., Drews, S., van den Bergh, J., 2019. Perceived fairness and public acceptability of carbon pricing: a review of the literature. Clim. Policy 19, 1186–1204. [https://doi.org/10.1080/14693062.2019.1639490.](https://doi.org/10.1080/14693062.2019.1639490)
- Martiskainen, M., Sovacool, B.K., Lacey-Barnacle, M., Hopkins, D., Jenkins, K.E.H., Simcock, N., Mattioli, G., Bouzarovski, S., 2021. New dimensions of vulnerability to energy and transport poverty. Joule 5, 3–7. [https://doi.org/10.1016/j.](https://doi.org/10.1016/j.joule.2020.11.016) [joule.2020.11.016](https://doi.org/10.1016/j.joule.2020.11.016).
- Mattioli, G., 2017. "Forced car ownership" in the UK and Germany: socio-spatial patterns and potential economic stress impacts. Soc. Incl. 5, 147–160. [https://doi.org/](https://doi.org/10.17645/si.v5i4.1081) [10.17645/si.v5i4.1081](https://doi.org/10.17645/si.v5i4.1081).
- Mattioli, G., Wadud, Z., Lucas, K., 2018. Vulnerability to fuel price increases in the UK: a household level analysis. Transport. Res. Pol. Pract. 113, 227–242. [https://doi.org/](https://doi.org/10.1016/j.tra.2018.04.002) [10.1016/j.tra.2018.04.002](https://doi.org/10.1016/j.tra.2018.04.002).
- [Mattioli, G., Wadud, Z., Lucas, K., 2016. Developing a Novel Approach for Assessing the](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref55) [Transport Vulnerability to Fuel Price Rises at the Household Level.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref55)
- Nature, 2018. Wanted: a fair carbon tax. Nature 564, 161. [https://doi.org/10.1038/](https://doi.org/10.1038/d41586-018-07717-y) [d41586-018-07717-y](https://doi.org/10.1038/d41586-018-07717-y), 161.
- [OECD, 2022. Why Governments should Target Support Amidst High Energy Prices](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref57). [OECD, 2011. Divided We Stand. Why inequality keeps rising](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref58).
- Oseni, G., Palacios-Lopez, A., Mugera, H.K., Durazo, J., 2021. Capturing what Matters: Essential Guidelines for Designing Household Surveys. World Bank. [https://doi.org/](https://doi.org/10.1596/36763)
- P´erez, A., 2022. Fin al descuento de los 20 c´entimos en carburante: el Gobierno solo lo mantiene para profesionales [WWW Document]. [https://www.autobild.es/noticias](https://www.autobild.es/noticias/fin-descuento-20-centimos-carburante-gobierno-solo-mantiene-profesionales-1175906) [/fin-descuento-20-centimos-carburante-gobierno-solo-mantiene-profesionales](https://www.autobild.es/noticias/fin-descuento-20-centimos-carburante-gobierno-solo-mantiene-profesionales-1175906) [-1175906,](https://www.autobild.es/noticias/fin-descuento-20-centimos-carburante-gobierno-solo-mantiene-profesionales-1175906) 3.10.23.
- Piketty, T., Saez, E., 2014. Inequality in the long run. Science 344, 838–843. [https://doi.](https://doi.org/10.1126/science.1251936) [org/10.1126/science.1251936.](https://doi.org/10.1126/science.1251936)
- [Primc, K., Slabe-Erker, R., Majcen, B., 2019. Constructing energy poverty profiles for an](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref62)
- [effective energy policy. Energy Policy 128, 727](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref62)–734. [Rama, H.O., Roberts, D., Tignor, M., Poloczanska, E.S., Mintenbeck, K., Alegría, A.,](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref63) Craig, M., Langsdorf, S., Loschke, ¨ S., Moller, ¨ [V., Okem, A., Rama, B., Ayandale, S.,](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref63) [2022. Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref63) [Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref63) [Climate Change. Cambridge University Press.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref63)
- Robinson, R., Thagesen, B., 2017. Road Engineering for Development, second ed. CRC Press, London. .
- Rousseeuw, P.J., 1987. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. J. Comput. Appl. Math. 20, 53–65. [https://doi.org/10.1016/0377-](https://doi.org/10.1016/0377-0427(87)90125-7) [0427\(87\)90125-7](https://doi.org/10.1016/0377-0427(87)90125-7).
- [Salon, D., Blumenberg, E., Thomas, T., Williams, J., Tuil, J.L., 2016. A spatial analysis of](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref66) [housing and transportation affordability in Los Angeles county. Presented at the](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref66) [Transportation Research Board 95th Annual MeetingTransportation Research Board.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref66)
- Salon, D., Gulyani, S., 2010. Mobility, poverty, and gender: travel 'Choices' of slum residents in Nairobi, Kenya. Transp. Rev. 30, 641–657. [https://doi.org/10.1080/](https://doi.org/10.1080/01441640903298998) [01441640903298998.](https://doi.org/10.1080/01441640903298998)
- Schouten, A., 2022. Residential location and household spending: exploring the relationship between neighborhood characteristics and transportation and housing costs. Urban Aff. Rev. 58, 1554–1584. [https://doi.org/10.1177/](https://doi.org/10.1177/10780874211028814) [10780874211028814.](https://doi.org/10.1177/10780874211028814)
- Shen, Q., 1998. Location characteristics of inner-city neighborhoods and employment accessibility of low-wage workers. Environ. Plann. Plann. Des. 25, 345–365. [https://](https://doi.org/10.1068/b250345) [doi.org/10.1068/b250345](https://doi.org/10.1068/b250345).
- Sustrans, 2012. Locked Out: Transport poverty in England [WWW Document]. [htt](https://www.sustrans.org.uk/our-blog/research/all-themes/all/transport-poverty-in-england/) [ps://www.sustrans.org.uk/our-blog/research/all-themes/all/transport-poverty](https://www.sustrans.org.uk/our-blog/research/all-themes/all/transport-poverty-in-england/)[in-england/,](https://www.sustrans.org.uk/our-blog/research/all-themes/all/transport-poverty-in-england/) 8.12.22.
- Tao, S., He, S.Y., Kwan, M.-P., Luo, S., 2020. Does low income translate into lower mobility? An investigation of activity space in Hong Kong between 2002 and 2011. J. Transport Geogr. 82, 102583. [https://doi.org/10.1016/j.jtrangeo.2019.102583.](https://doi.org/10.1016/j.jtrangeo.2019.102583)
- Tomas, ´ M., García-Muros, X., Alonso-Epelde, E., Arto, I., Rodríguez-Zúniga, ˜ A., Monge, C., Gonz´ alez-Eguino, M., 2023. Ensuring a just energy transition: a distributional analysis of diesel tax reform in Spain with stakeholder engagement. Energy Policy 177, 113558. [https://doi.org/10.1016/j.enpol.2023.113558.](https://doi.org/10.1016/j.enpol.2023.113558)
- [Vaidyanathan, S., Huether, P., Jennings, B., 2021. Understanding Transportation Energy](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref74) [Burdens. American Council for an Energy-Efficient Economy \(ACEEE\), Washington,](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref74) [DC.](http://refhub.elsevier.com/S0967-070X(25)00406-8/sref74)
- Wachs, M., Kumagai, T.G., 1973. Physical accessibility as a social indicator. Soc. Econ. Plann. Sci. 7, 437–456. [https://doi.org/10.1016/0038-0121\(73\)90041-4.](https://doi.org/10.1016/0038-0121(73)90041-4)
- Ward, J.H., 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 58, 236–244. .