The role of personal motives in determining car ownership and use: a literature review

ABSTRACT

Car dependency leads to a variety of societal problems and challenges, not least environmental ones. It is thus not only relevant but also critical to better understand the determinants of car ownership and use. Among those contributing factors, the role of subjective factors is often acknowledged to be important, yet not well understood. We conduct a systematic review of the literature regarding the effect of such motives on car ownership and usage. Based on the commonalities found in the reviewed articles, we identify and describe the five most relevant subjective factors in detail: (i) instrumental motives and autonomy, (ii) affective motives, (iii) symbolic motives, (iv) social norms, and (v) environmental motives. We synthesise these findings in a car ownership/use motives model, discuss implications for public policy and outline directions for future research.

KEYWORDS:

• Car ownership
• car use
• subjective variables
• motives
• attitudes

1. Introduction

Private and motorised travel is associated with high costs not only for the actual drivers who perform those but also for society as a whole, including in the form of greenhouse gas emissions. In fact, road transport accounts for more than three-quarters of the total emissions generated by the transportation sector in the US (US Environmental Protection Agency, 2020) and in Europe (European Commission, Directorate-General for Mobility and Transport, 2022). In addition, car use is also associated with negative health impacts and time wasted in traffic congestion, among others(Rojas-Rueda et al., 2011; Wee, 2013). Nevertheless, car use and ownership have been increasing in recent years in many OECD countries, and it has been estimated that by 2030 the majority of vehicles will actually be owned by residents of non-OECD countries (Dargay et al., 2007). Even in countries with high-quality public transport and bike infrastructure, such as the Netherlands, the majority of trips and trip-kilometres are performed by cars (Ton et al., 2020). It is therefore paramount to understand the underlying determinants of car ownership and usage-related choices.

Several review papers focused on specific aspects of car ownership and usage modelling. Anowar et al. (2014) focus on the different modelling approaches and variables considered to model vehicle ownership. They consider variables such as household demographics, individual characteristics, life events, and the built environment, among others. They highlight the relevance, but do not review, the importance of unobserved effects, such as a proclivity toward a particular mode, perception of comfort, and environmental consciousness. These variables are often omitted due to their more complex measurement. Given the likely correlation between some of these omitted variables and those actually measured, for example between environmental consciousness and education, the estimation of the effect of the non-omitted variables is inconsistent and biased. It is therefore evident that while there is a need to gain a better empirical underpinning of the impacts of subjective factors on car usage and ownership, the measurement thereof remains challenging and is not yet well understood.

A sociological perspective on mobility is provided in a review by Cairns et al. (2014). They highlight that promoting more sustainable travel patterns requires modifications at different social levels and not only in terms of individual decision making. In terms of car mobility, they focused on individual behaviour, car dependence, emotional dimensions, symbolic role in society, and concerns about consumption. While their review offers evidence for the existence and importance of those dimensions in car ownership decisions, they do not review the methods required for measuring their relevance, how these unobservable variables correlate with each other, and the heterogeneity of those across the population.

Bhagat-Conway et al. (2022) explore the explanatory effect of subjective variables in travel behavioural analysis. They reviewed 158 studies, which considered a total of 2,864 different subjective survey questions. Due to the large number of distinct questions, the authors were unable to draw specific conclusions regarding the impacts on particular modes or purposes. In particular, they claim vehicle ownership only relates to convenience and residential preference. Conversely, our aim is to conduct a targeted analysis of car ownership and usage.

The complexity of the different motives, both personal and societal, for either owning or using a car makes car dependency a complex topic to research (Zhao, 2011). In addition, attitudes regarding car mobility and behavioural decisions, such as owning or using a car, are tangled because of their inter-dependency, and the existing bidirectionality has been proved but not always considered in the analysis of car dependency (Kroesen et al., 2017; Moody & Zhao, 2019). de Vos et al. (2021). propose a comprehensive model linking attitudes, desires, and intentions with actual behaviour and subsequent satisfaction, while also comparing the different social-psychological attitude theories that provide a foundation for the relationship between them which is yet to be designed and applied for the context of car dependence, and thus in ownership and use.

Given the intricate nature of this problem, devoting more research efforts to unravelling some of the implicated dimensions is needed. To this end, we conduct a comprehensive and systematic review of the literature concerning the subjective determinants relevant to the decision to own a car and its usage frequency. In our review, we also examine the different methods employed for measuring these motives and critically assess their relevance for investigating car ownership and use decisions. The remaining of this paper is structured as follows: in Section 2, we elaborate on the review procedure and provide indications regarding the features of the papers included in the review. In section 3, we present some general description of the papers analysed. Then, Section 4 constitutes the core of the article, where we describe each of the five subjective dimensions we identify in detail. Section 5 synthesises the results by describing a car ownership-use motives model. Finally, we conclude our article in Section 6 by discussing policy implications and further research needs.

2. Research method

Our initial interest encompassed determinants of car ownership and usage, especially in relation to shared mobility and the ability to use cars without owning one. Thus, two research queries were specified in Scopus on March 14th, 2022, as follows:

Table

Query 1.	TITLE-ABS-KEY ({car ownership} OR {car usage} OR {car dependence} OR {car independence})
	AND
	PUBYEAR > 1999
Query 2.	ALL ({car ownership} OR {car usage} OR {car dependence} OR {car independence})
	AND
	TITLE-ABS-KEY ({ridehailing} OR {ridesharing} OR {ride-hailing} OR {ride-sharing} OR {shared mobility} OR {car leasing} OR {car sharing})
	AND
	PUBYEAR > 1999

Our goal was to find articles focusing on the ideas of car ownership and usage (Query 1), but also including the concept of car dependency. We also wanted to include articles studying the aforementioned topics in relation to shared services (Query 2). The reasoning for doing so was in order to find articles that might explore car use without necessarily owning one, as in car-sharing or, to a lesser extent, in ride-hailing. A total of 2,277 articles were found after this comprehensive search, which was followed by a title and keywords screening leading to 383 relevant articles. This first filter gave us sufficient flexibility to consider also more general mode choice papers that are equally relevant for our purposes of studying car ownership and use. As the number of articles was found to be extensive and possibly prohibitive, we decided to narrow down the scope following the guidelines offered by Wee and Banister (2016) and de Vos and El-Geneidy (2021). To do so, we carried out a labelling process where we characterised the set of 383 papers in terms of the type of variables analysed, considering not only subjective factors but also the (i) built environment, (ii) urbanisation level, (iii) economic and equity considerations, (iv) other transportation modes, and (v) the role of life events. Note that these labels are not mutually exclusive, i.e. even when a dimension was studied but not as the main objective of the article, it was still indicated for the respective paper. By doing so, we aimed to find relevant interactions between these different dimensions. However, the number of articles at the intersection of several dimensions proved to be too limited to justify a review in its own right.

We focus on those articles that consider car ownership or use as their subject of analysis (dependent variable) rather than those where car ownership and use played an explanatory role (independent/explanatory variable). Finally, we decided to scope our review study to those articles that consider the effect of different subjective variables on car ownership and/or use. Furthermore, we excluded those articles or book chapters that do not analyse empirical data, either quantitative or qualitative. This process results in a final set of 54 papers after removing those written in a language other than English on the one hand and including additional relevant references cited in our set of papers but not yet included (backwards snowballing) on the other hand. This group of articles is carefully read to identify key features, commonalities and differences. To do so, we performed a frequency analysis of the different subjective dimensions considered in this set of papers in order to extract a set of key factors thereafter used for structuring the review of the substantive findings. Then, we synthesise the key findings and ideas and propose a conceptual model. The process is summarised in Figure 1.

Figure 1. Literature review research method summarised.

3. Descriptive analysis

Several aspects of the data and methods employed by the reviewed papers are first discussed before turning to the synthesis of the various subjective dimensions identified. We observe that most of the studies analyse data collected in (Western) Europe, the US, and China (Figure 2). There is scarcely an analysis of data from African and Latin American contexts, with only one relevant study for each. Several of the analysed articles refer to these socio-psychological constructs as attitudes in an overgeneralising way, as some of these constructs are not attitudes. Thus, we refer to these variables as subjective dimensions or motives in the remainder of the article. In the following section, five different subjective dimensions will be presented, considering the effect of the expectations from others and towards others, the feelings attached to (owning and using) cars, their perceived convenience, as well as other relevant dimensions. Despite our efforts to explore potential geographical differences, we did not uncover consistent variations across regions. Instead, the distinctions primarily arose from the unique characteristics and circumstances of each individual case study. Consequently, we present our findings without acknowledging possible but unstudied cultural differences, with the exception of cases where multiple cities are analysed and compared within a single study.

Figure 2. Geographical distribution of data employed in past studies.

The majority constitute quantitative studies: 16 out of the 56 reviewed studies conducted qualitative analysis, of which two are in combination with quantitative analysis. The most common approach adopted by the set of 42 quantitative studies involved some form of factor analysis or principal component analysis, in order to reduce the dimensionality and come up with latent constructs associated with the different subjective attributes, followed by an appropriate model and its analysis. The most commonly used modelling framework is Structural Equations Modelling, which is carried out by 17 out of the 42 studies that conducted a quantitative analysis. The dataset analysed varies from around 200 respondents to almost 3,000. Arguably, these large differences in sample size partially stem from the fact that some of the analyses involve national or regional surveys funded by governmental institutions, whereas others design and deploy their own specific data collection, often times resorting to convenience samples. A limitation of some of these samples is that it is quite common to find articles focused exclusively on the “young population” by sampling university students or even those students of specific university departments. As a matter of fact, among the 42 reviewed papers, 13 are focused on university students.

Most of the qualitative studies included in our review conducted individual interviews, and Thematic Analysis was the most common methodology for analysing the data collected, used in 10 out of the 16 studies. Unfortunately, the method applied by qualitative studies included in our review set was often not described in any detail. Samples ranged between 25 and 50 interviewees, with the exception of Steg et al. (2001), which had a sample of 185 participants.

4. Subjective dimensions

In the following subsections, we describe in detail the findings pertaining to five different car subjective dimensions. To do so, the articles were analysed and coded, and frequency analysis was employed to identify the most common themes. Based on these findings, the five most frequently observed subjective variables were selected as the motives to focus on. These subjective factors pertain to (and their share over the reviewed papers) instrumental motives and autonomy (82%), affective dimension (38%), status symbol (59%), social norms (30%), and environment (41%). We also provide an additional subsection with other non-recurrent variables analysed by some of the papers. In addition, we include two complementary appendices. Appendix A provides a summary table with details about the data, methodologies, case study context and the presence of each subjective dimension as well as the respective findings. We also make the distinction between different objects of study, namely, car ownership, car use, or both simultaneously. Appendix B compiles statements that have been used to specify the analysed subjective factors in surveys. We have also categorised those statements into sub-topics within each factor and provide our personal recommendations based on our experience and our review. Both appendices can be found in the suplemental data of the article.

4.1. Instrumental motives and autonomy

Expectations concerning the level of service offered by different transport alternatives is a key determinant of travel and transport mode choices. Motives related to the instrumental motives of owning and using a car are often defined in terms of the sense of control over one’s own life, autonomy, and the freedom to use it when desired and to any destination (Barker, 2014; Beirão & Sarsfield Cabral, 2007; Chng et al., 2019; Haustein, 2021). For example, some common statements included in this category are: “cars give freedom to travel any time”, “driving makes my life more easy”, and “owning a car allows me to do whatever I want”. More examples can be found in Appendix B.

Furthermore, a higher comfort level compared to other modes of transport is linked to aspects such as travelling seated, isolated (from disturbances induced by fellow passengers), and protected from the environment(Maciejewska & Miralles-Guasch, 2019; Thøgersen et al., 2021; Verma et al., 2016), and the possibility to carry bulky luggage or heavy loads (Berg et al., 2015; Gebhardt, 2021; Heinonen et al., 2021; Ibrahim, 2005; Verma, 2015). Interestingly, in some studies, these motives are also considered a facilitator of socialising and maintaining friendships, most relevant for younger groups (Delbosc & Currie, 2014; Hopkins, 2016; Kim et al., 2020; van Acker et al., 2014).

As evident, there are multiple features that are bundled together within this dimension, namely instrumental or easiness aspects, autonomy or freedom, and comfort or safety. In fact, there is no clear trend in the literature in terms of whether to account for these aspects separately or jointly, often referred to as convenience or instrumental motives. For example, when looking at the 46 reviewed articles that study some of these sub-topics, 19 of them do so separately, while 27 combine at least two of them. Interestingly enough, there are seven papers that combine the three aspects but four of them call the entire factor “purely instrumental”, without acknowledging autonomy and comfort, indicating the absence of a consensus.

Moreover, this dimension is the one closer to quantitative attributes commonly included in transportation modelling, such as access, waiting, and travel times. Consequently, these motives are present in the vast majority of reviewed papers (44 out of 54), as can be seen in Appendix A. Interestingly, this subjective variable is rated as more important by habitual drivers (Belgiawan et al., 2014; Ikezoe et al., 2021; Steg, 2005; Verma et al., 2016), suggesting that using the car creates a positive experience (Sigurdardottir et al., 2013) which in turn increases the likelihood of buying a car (He & Thøgersen, 2017; Pojani et al., 2018). This phenomenon is in line with the evidence for a bidirectional relation between car ownership intentions and car usage, as previously described in the introduction (Kroesen et al., 2017; Moody & Zhao, 2019).

As mentioned, some studies account for this dimension by means of an instrumental factor only, which is associated with the same attributes described in this section. Past findings suggest, for example, that men have a higher tendency to drive more because of a more positive appreciation of these instrumental characteristics (Bergstad et al., 2011; Van & Fujii, 2011), whereas the relationship with age is inconclusive (Van & Fujii, 2011; Zhou & Wang, 2019). However, the use of the “instrumental” expression is conflictive, as it has been previously described as the degree to which a car serves exclusively in terms of moving someone from their origin to their destination and nothing else (Magdolen et al., 2021; Steg, 2005). This type of instrumental definition is usually negatively correlated with higher car use. As we only focus on those aspects related to the relative convenience of the cars, we leave the articles referring to the original definition of instrumental in Appendix A, and note these under the “Less explored factors” section.

Moody et al. (2021) found that people own and use vehicles because they perceive an added value (which mostly comes from the attributes referred to in this subsection) that compensates for the costs incurred. However, it remains unknown to what extent some of these attributes are associated with travelling by car as-such, as opposed to being associated with being the owner and driver of the vehicle. Even though these dimensions do not necessarily happen in conjunction, past research has not clearly distinguished between them and therefore does not allow for disentangling their respective motives. For example, the easiness of transporting belongings as well as the comfort level could be, at least, partly emulated through car-sharing or ride-hailing services, whereas the feelings of freedom and autonomy are likely to more clearly connect to the notion of ownership, i.e. a car that is reserved for an exclusive use rather than shared fleets and services which are bounded in terms of the availability of vehicles or drivers around. A related yet distinctive set of characteristics associated with the emotions that occur when being behind the driving wheel or as a car passenger is described in the following sub-section.

4.2. Affective dimension

The set of motives and emotions associated with having access to a car as-such as well as those associated with the driving task, is what is usually referred to as the affective dimension of car ownership and use. For instance, statements such as “I love my car”, “driving a car is fun”, or “driving gives me a kick” are quite common to find in structural models or as quotes of participants when describing this attribute. For more examples, please refer to Appendix B. These ideas are typically also referred to as descriptors of being “pro-car” (Heinonen et al., 2021; Kalter et al., 2020). Having positive motives towards cars is found to have a positive effect on car use (Abrahamse et al., 2009; Choocharukul et al., 2008; de Vos & Alemi, 2020; Kalter et al., 2020; Lois & López-Sáez, 2009) and ownership (Heinonen et al., 2021; Ikezoe et al., 2021). In addition, it has been found that this set of affective motives is also related to a higher number of cars owned and also to using cars for the majority of the trips (Magdolen et al., 2021). The relation in the opposite direction also holds, as a study focusing exclusively on people aged between 25 and 40 years old found out that those with less “pro-car” motives are more likely to have a car-free lifestyle (Heinonen et al., 2021).

Some studies focus exclusively on how car affection develops in the younger segments of the population. This is interesting as they have not driven much yet and their subjective value towards cars are often shaped as passengers. Even though we didn’t focus our review in terms of future car intentions, we consider it is relevant in this context as the affective dimension develops over time, and become then an explanatory variable in adulthood. These studies conclude that among all of the different effects, a positive car experience amongst the youth is the strongest predictor of their future car ownership and use (Pojani et al., 2018; Sigurdardottir et al., 2013). This suggests that car motives can actually be passed on and inherited from parents to their children based on their driving stances. Interestingly, Pojani et al. (2018) also report that even adolescents who do not enjoy their car experience express their intention to drive or at least learn how to drive in the future.

The set of affective motives toward cars are specifically related to the car experience and could only partially be transferred to other modes of transport, such as pro-bike motives, where people also have strong emotions towards bikes and riding them. The higher the attachment to one’s cars, the harder it is to attract someone to consider other alternatives (Beirão & Sarsfield Cabral, 2007; Kalter et al., 2020). In this sense, car affection needs to be acknowledged when designing or improving public transport alternatives, as it is not sufficient to only consider the level of service offered.

As mentioned, this affective dimension encompasses all emotions and thoughts a car generates in one’s life. Notwithstanding, car ownership and use is also considered as part of a signal that people wish to communicate to family, friends, and society at large. These two subjective factors are sometimes studied jointly as symbolic-affective motives. The specifics about the symbolic value of cars is detailed in the following subsection.

4.3. Car as a status symbol

A different dimension has to do with the symbolic value someone associates with owning a car in terms of their social status and also as self-identity in terms of placing themselves as successful individuals. This idea is typically referred to as the symbolic value of cars. This is reflected, for example, in ideas associated with an image of progress, which is also usually associated with how someone wants to signal success. For example, Steg (2005) finds out that ideas such as “a car provides status and prestige”, “my car shows who and what I am”, and “you can know a person by looking at his or her car” are associated with an increase in car use for commuting trips. This dimension has been included in more than half of the reviewed studies (see summary table in the Appendix) and was almost always found to be associated with an increased likelihood of owning and using a car (Moody & Zhao, 2020). The only partial exception is reported by studies in China, where younger respondents considered the car as a way of emotional expression more than as a mode of transportation, resulting in a higher car ownership but at a reduced usage rate when compared to older generations (Zhou & Wang, 2019). Furthermore, it was also found that while people living in rural parts of China acquire vehicles because of the status symbolic value, this effect is smaller than in urban areas (Ao et al., 2019).

Another aspect which has gained some consideration in this domain pertains to how different types and models of cars alter status signalling. It was found, for example, that the brand and cost of a car can strengthen the symbolic value of cars (Zhao & Zhao, 2020). This aspect gains additional importance nowadays, considering the rapid adoption of electric and, later on, self-driving vehicles, with Tesla being a notable example. How different the status image of an owner of such cars will be is still unclear as, to the best of our knowledge, this has not been studied. We hypothesise that it will make the interactions between the different motives more complex because of the relation (e.g. common underlying socio-demographics) between tech-savviness and environmental concerns motives in terms of fossil fuel usage (Haustein & Jensen, 2018). In fact, this symbolism associated to electric cars not only varies within a society (Morton et al., 2017) but also across different cultures (Ashmore et al., 2018).

In terms of sample characteristics and their relation with this subjective factor, the most common and relevant result is that men attach greater importance to the status symbol value of cars than women (Hiscock et al., 2002; Ikezoe et al., 2021; Pojani et al., 2018). Interestingly, income is commonly found to exhibit no relation with the importance attached to the status symbol value of cars. Notwithstanding, it has been found that low-income groups that acquire a vehicle as a notion of progress find it harder to stop using it as it was difficult to get in the first place(Beirão & Sarsfield Cabral, 2007).

For the younger population (i.e. between 17 and 23 years old), the symbolic value is also related to social norms, with owning and using a car being connected to ideas such as adulthood, freedom, and financial independence (Ashmore et al., 2020; Delbosc & Currie, 2014). This is in line with results from a qualitative study which identifies how owning a car is connected to young people’s aspirational status symbols (Zhu et al., 2012) and cultural icons, thereby providing a framework to enhance identity (Barker, 2014).

A limitation found in several studies is the grouping of this subjective factor with others related to the emotion a car triggers in oneself(Choocharukul et al., 2008; de Vos & Alemi, 2020; Ikezoe et al., 2021; Kalter et al., 2020; Luke, 2018; Magdolen et al., 2021; Pojani et al., 2018; Steg et al., 2001). This construct is typically named affective-symbolic motives (or vice-versa). However, the majority of these studies focus mostly on the symbolic aspect, as they typically include only one or two affective statements such as “I like driving”. This is a clear example of the existing interactions between motives, as those statements relate to the excitement subjective factor we already described in the previous subsection.

4.4. Social norms

Social norms encompass what is commonly practised (descriptive norms) or (dis)approved (injunctive norms) within a society or group. They reflect the beliefs, attitudes, and behaviours of others, shaping individuals’ understanding of what is acceptable or expected (Keizer & Schultz, 2018). Social norms also play a significant role in someone’s decisions about travelling by car. Overall, the stronger someone considers that buying and car and travelling by car is the normal and expected thing to do, the higher the chance that they do it (Barker, 2014; Belgiawan et al., 2017; Donald et al., 2014; Heinonen et al., 2021; Paijmans & Pojani, 2021; Steg, 2005; Steg et al., 2001). Past studies have used a variety of statements to operationalise this dimension, such as “my family/friends think I should commute by car”, “getting your own car is part of growing up”, “cars are a part of modern life”. More examples are presented in Appendix B. Thøgersen et al. (2021) report that the set of social norms is also reinforced by positive motives towards car use, in terms of instrumental motives, affection, and symbolic value.

It has been found that adolescents perceive learning to drive as a ritual linked to reaching adulthood, as they also tend to not have more additional motivations to do so (Hopkins, 2016). When an adolescent’s social network favours driving more, their car intentions are stronger (Luke, 2018). These results are in line with the findings from Van and Fujii (2011), who point out that the effect of social norms is higher for older age segments, as several of the ideas associated with social expectations are related to becoming an adult and having a normal adult lifestyle.

The social norms dimension is the mirror image of the previously reviewed “car as a status symbol” dimension. Interestingly, some studies have analysed this dimension in relation to a recurrent travel choice, with statements such as “My colleagues would think it is peculiar not to commute by car”. In some cases, those living voluntarily without cars are considered to be outsiders (Heinonen et al., 2021). It is also common to associate those living a car-free lifestyle with not having the money to afford it and deprivation (Heinonen et al., 2021; Paijmans & Pojani, 2021). Thus, it is evident that there exists a complex relationship between social norms and the symbolic value of cars, as someone might try to avoid being labelled poor by owning a car. A well-thought methodological set-up needs to be devised to identify both effects separately.

One could also consider that, among the set of existing social norms, there is a subgroup that pertains to all the ideas associated with environmental concerns. This dimension has gained more relevance in recent years because of concerns associated with climate change. To some extent, it could be argued that environment-related expectations from society constitute part of social norms. However, given the specificity, relevance, and also prevalence in past studies, we have decided to consider it separately in the following subsection.

4.5. Environment

Environmental concerns about travel decisions and related motives have a significant effect on car ownership and use. There is a variety of environmental subjective variables that have been studied in the context of car ownership and use, such as environmental norms, behaviours, and awareness, to name a few. As mentioned in the previous sub-section, these motives are prone to social pressure, which means that if one’s social circle has an interest in climate concerns, then one is more likely to develop pro-environment motives in terms of their mobility choices, environmental impact, and climate change actions (Thøgersen et al., 2021). According to the literature, those who lean towards more environmentally conscious lifestyles also reduce their car ownership and use (Paijmans & Pojani, 2021; Thøgersen et al., 2021). Environmental motives can also reduce driving indirectly as it reduces the habitual intention to drive (Donald et al., 2014). Furthermore, stronger environmental motives are associated with higher use of other transportation alternatives, such as active modes and public transport (Anable, 2005; Kalter et al., 2020; Vredin Johansson et al., 2006).

There is empirical evidence to suggest that it is attitudes rather than factual knowledge that promote more pro-environment behaviours (Nilsson & Küller, 2000). Conflicts between attitudes and reasoning lead to counterintuitive outcomes, as people expressing being worried about the effects of their car use may still use them with no noticeable change in their travel choices (Bamberg & Schmidt, 2003; Delbosc & Currie, 2014). Also, the more they drive to work, the more responsible they tend to feel about the negative effects thereof (Abrahamse et al., 2009). For adolescents, the effect of parents’ education and behaviour is significant, meaning that norms related to the environment (Pojani et al., 2018), as well as the willingness to accept car restrictions (Sigurdardottir et al., 2013), are stronger among members of families making conscious pro-environmental choices. Nevertheless, these subjective factors are found to play a limited role in the future car ownership decision of adolescents (Belgiawan et al., 2016). It has been found, however, that exposure to more sustainable modes, such as transit, can increase the future usage of kids when growing up (Smart & Klein, 2018).

Living a car-free lifestyle is positively correlated with more pronounced pro-environmental attitudes (Heinonen et al., 2021). Those who exercise a car-free pro-environment lifestyle are disproportionally more educated, employed, and higher income individuals (Anable, 2005; Brown, 2017; Kalter et al., 2020; Magdolen et al., 2021). The environmental concerns amongst this specific group of people also interact, in some case studies, with their symbolic and affective motives, because of their high education and income. Magdolen et al. (2021) speculate that this may be the case that these other motives outweigh the negative effect of environmental concerns in terms of car use, resulting in a positive effect overall. Brown (2017) also find out that people whose motives and behaviours are associated with a pro-environmental lifestyle are significantly more likely to hold a carshare membership.

Overall, the relationship between environmental concerns and car ownership and usage is a complex one. On one side, there is evidence to suggest that environmental concerns are associated with a reduced number of vehicles per household as well as multimodal transportation choices. On the other side, the effect in the opposite direction also exists as people who are more concerned about the environment are also more educated, which commonly relates to higher income and being more mobile. The ultimate effect is thus uncertain, especially when considering the interaction with zero-emission vehicles and the perceived status of the most luxury electric cars available in the market, as discussed in section 4.3.

4.6. Less explored factors

Lastly, we examine those subjective variables that we find in some articles but were either not frequently mentioned or significant enough to detail them in separate subsections. Among those, the most recurrent had to do with ideas associated with the monetary costs of travelling by car, considered by 12 of the reviewed works. Transportation costs are a critical element in explaining mode choice behaviour and are therefore considered a key variable in modal choice model, but they can hardly be considered a subjective factor. Past evidence shows that travellers are aware that car costs are higher than those of the alternatives, this being most often a relevant factor for non-car owners (Beirão & Sarsfield Cabral, 2007; Hagman, 2003). In fact, the majority of zero-car households report cost to be the main reason for not owning a car (Brown, 2017). However, it has been found that instrumental motives (section 4.1) and cost-related ideas are of complex nature, as people own cars regardless of their economic situation when they consider cars to be necessary (Ikezoe et al., 2021). Gardner and Abraham (2007) also point out that people consider some of the costs associated with cars to be sunk costs, which incentivise car use even further in order to avoid not benefiting from their purchase. All these facts make it complicated to differentiate cost-related motives from economic constraints.

Several studies investigate the effect of motives towards other transportation modes on car ownership and use. Typically, these studies are focused on public transport (Belgiawan et al., 2014; Heinonen et al., 2021; Luke, 2018; Nilsson & Küller, 2000; Verma, 2015) and cycling (de Vos & Alemi, 2020; Sigurdardottir et al., 2013). No correlation is found between favourable motives and perceptions towards public transport and car purchase because the effect of the instrumental motives is higher (section 4.1) (Belgiawan et al., 2014). It has also been reported that pro-cycling motives are associated with a bicycle-oriented vision of the future (Sigurdardottir et al., 2013), which is associated with pro-environment motives (section 4.5). Hence, even though a negative relationship between pro-cycling and car ownership and use can be found, it is difficult to estimate the direct effect apart from the one of environmental concerns.

An additional stream of research studies the influence of reasons that enable or hinder choice realisation, referred to as perceived behavioural control (PBC) over car-related choices (Anable, 2005). It has been found that PBC does not have a direct effect on car use, but rather an indirect one through car attitudes (Donald et al., 2014). People with a sense of high control over their choices might already be using different alternatives to cars, whereas it is the people with a low sense of control who only reduce car use when feeling a moral obligation or the influence of their social circle (Abrahamse et al., 2009). Perceived behavioural control also relates to habits, as some people might perceive to have a limited ability to change their habits. The effect of habit is found in several studies, finding out that the more habitual car use is, the easier it is to develop positive motives towards cars, which ends up in a self-reinforcing cycle (Domarchi et al., 2008). Additional motives studied by the reviewed papers can be found under the column “Less explored factors” in Appendix A.

5. Synthesis: car ownership/use motives model

We synthesise the interactions between the different subjective constructs reviewed in Figure 3. The width of the links reflects the relative strength of the factors, based on both the overall presence in the literature and the relative importance as reported in the reviewed papers. We see that instrumental motives are found the strongest determinants in explaining car ownership and use, followed by symbolic motives. Both affective and environmental motives are found to have a lesser relation, especially when studying car ownership exclusively.

Figure 3. Car ownership-use motives model.

The figure synthesises several key observations. First, an individual develops a sense of relative convenience of cars, based on the different perceived and experienced car instrumental motives and autonomy, as well as an affective dimension. Second, individuals’ motive interaction with society is bi-directional: on one hand, people communicate a symbolic value of what a car means and on the other hand social norms shape the extent to which people own and use a car. Third, the subset of social norms related to environmental concerns plays a role in someone’s perception of cars (and other transportation modes). Fourth, there are also other motives that influence car ownership and use decisions, but to a lesser extent.

Based on our review of the substantive findings reported in the literature, we conclude that car instrumental motives and autonomy are typically found to constitute relevant explanatory variables, which is in line with its frequent consideration in the reviewed papers. In addition, the status symbol and social norms are typically modelled together. Lastly, environmental concerns have received increasing attention during the past decade, which we hypothesise is driven by the increased attention to carbon emissions and climate change. The combination of these different aspects determines the overall subjective value towards the intention to buy (or continue owning) a car and thereby travel by car, which we frame as the car ownership-use motives model.

As noted in every of the articles reviewed, car use and ownership are complex phenomena. The quantitative studies measuring the type of subjective characteristics reviewed in this article differ both in the set of motives considered as well as in how these motives interact with each other. First, they do not always consider the same subjective factors, and also, sometimes motives are mixed into larger overarching constructs as the definition of each individual one is not universal. Second, the reviewed studies don’t always test the same interactions between each motive as well as with car ownership and usage. For example, it can be the case that a study tests the effect affective motives have over the symbolic value, and then the effect of the symbolic value over car ownership, when also a different one could test and find significant and direct impacts of both subjective factors separately. It is then possible to study multiple ways these motives interact as there is still no evidence to prefer direct or indirect effects.

Many of the reviewed studies are grounded in general psychological frameworks, with the Theory of Planned Behaviour (Ajzen, 1991) being the most commonly used one. This theory proposes that individual decisions are mediated through Perceived Behavioural Control (PBC) and intentions, which are mediated also through subjective norms and attitudes. However, this theory has been criticised for not considering unconscious influences on choices as well as the role of emotions (Sniehotta et al., 2014). Rather than placing our conceptual motives model as part of one particular psychological theory, we argue that there is significant empirical evidence to suggest that the different subjective factors reviewed act as relevant explanatory variables of car ownership and use. This set of motives may also be integrated into a Travel Mode Choice Cycle (TMCC), which links attitudes and transport behaviour with travel satisfaction and intentions (de Vos et al., 2021).

Lastly, personal motives are by no means the sole determinants of car ownership and use. Other relevant and non-subjective factors include the role of the residential location and the built environment, the effect of life events such as moving to a different city or country, having a child, or retiring, the impact of different policies such as no-car zones and tax increments, as well as the effect of transportation costs, mobility patterns, and socio-demographic characteristics. Reviewing studies which have estimated the impacts of both subjective and non-subjective factors, we find the reported explanatory power to be of similar magnitude. Of which, only six (Domarchi et al., 2008; Heinonen et al., 2021; Komarov, 2004; Vredin Johansson et al., 2006; Zhao, 2011; Zhao & Zhao, 2020b) explicitly analyse the effect of incorporating subjective factors in addition to socioeconomic ones. Interestingly, all six papers concluded that the inclusion of subjective factors significantly improves the overall model fit. In addition, all these different variables are not independent from each other, as for example, subjective factors have a role in residential location and acceptance of different policies and vice-versa. This makes it difficult to propose a Structural Equation Model or a qualitative analysis that simultaneously accounts for all these variables, as noted in the papers that attempt that.

6. Discussion

In the following, we discuss different policy implications that originate from a better understanding of the role of subjective factors towards car ownership and use and thereafter outline possible future research directions.

6.1. Policy implications

Measuring the number of cars per household is essential when aiming for car ownership reduction. The location and the characteristics of the residents living zero-car lifestyles is relevant, as they could be perceived as successful examples and could guide policy towards expanding the circumstances that facilitate a car-free lifestyle. In doing so, it is critical to distinguish between two groups of people who live with no cars. On the one hand, there are people that make a deliberate choice to live without cars, which are typically referred to as car-free, and are typically associated with environmental motives. On the other hand, some people do not own a car due to personal restrictions, such as monetary or physical, referred to as car-less. In California, for example, based on the 2012 California Household Travel Survey, it was found that the car-less group accounts for the majority (80%) of zero-car households (Brown, 2017). The situation is thus rather complex, as policies targeted at increasing the wealth of lower-income families, or even general economic growth, might enable them to afford a private vehicle. It is furthermore important to couple such policies with improved car alternatives or car-sharing policies.

As evident from the literature review, people use cars for different motives. That is to say that what makes one individual shift to lower car usage will not necessarily have the same effect on others. Policies focusing on improving travel alternatives, such as public transport, more competitive in terms of travel time and comfort relate to some extent to how convenient the car is. However, that does not necessarily also address the other aspects associated with the instrumental motives and autonomy of cars, such as the sense of freedom and carrying heavy or multiple loads, and does not influence ideas related to the symbolic value of cars and social norms. While making bike use more attractive and micro-mobility services more widely available might tackle an increased sense of freedom, it is more complicated to address the dimensions associated with the society (from a symbolic and expectations points of view). One could argue that those with high symbolic and affective motives to use cars will only reduce their usage when facing restrictions such as no-car zones, reduced speed zones, or congestion charges. A different approach will require efforts to change these existing norms, as for example in the case of flight shame which has gained some popularity in recent years (Cats et al., 2022; Chiambaretto et al., 2021; Gössling et al., 2020).

Similarly, some people refer to the combination of all the facilitators of car mobility as a car culture. In our literature review, we saw modelling examples of this idea when combining multiple motives into one single “pro-car” motive. This car culture embraces individual mobility, which poses a challenge for shared modes. It is thus critical to keep in mind the long-term position of cars in society. If authorities aim to design policies towards reducing motorisation rates, then it is imperative also to consider the factors that influence and shape car culture – such as advertisements and popular culture representations. In addition, it is paramount to keep in mind that these factors might evolve over time, both driven by simmering societal trends as well as a consequence of more disruptive events, as during the past pandemic when safety-related motives became more relevant and explained shifts toward private mobility (Aaditya & Rahul, 2021).

Lastly, in terms of environmental motives, there is evidence to support that environmental behaviour rather than expressing an agreement with environmental statements is a better predictor of sustainability-minded motives towards mobility (Nilsson & Küller, 2000). Examples of these statements are found in Appendix B. This is explained by the fact that the word environment is associated with social desirability notions. Thus, policies associated with facilitating and promoting a pro-environmental conduct, such as recycling, second-hand acquisition, and repairing personal items, are expected to be more effective than informational campaigns about the effects of the different transportation modes on the environment.

6.2. Future research

The abovementioned car-free versus car-less distinction should be the subject of further research. As car-free households tend to be more affluent, the more their share grows, the harder it will become to explain car ownership rates based on income and education. However, it might also be the case that someone who cannot afford a car might not have the desire to buy one. It might also be the case that someone who can afford a car and would like to own one is limited because of health-related barriers. Thus, when analysing travellers’ segments and the determinants of car ownership and frequency of use, it is relevant to design a particular set of questions to differentiate the reasons for not owning a car. For example, this set of questions could be coupled with in-depth interviews or focus groups to understand better their reasons to not owning a car to then classify each respondent into different categories along the car-free to car-less spectrum. This information could then be used in supervised clustering methods to link the collected information with such classes.

Shared services such as on-demand mobility and car-sharing are also making the relation between owning and using a car less straightforward. Thus, it would be relevant to study in further detail how different motives associated with car ownership and car use evolve when car use is done in the context of shared vehicles. Even though they have been promoted as potential solutions for reducing car ownership, it has been argued in the literature that their users tend to be those who are younger, have higher incomes, and those having reduced car availability (von Behren et al., 2019). Also, the habitual use of these alternatives is fragile and could lead to car ownership driven by life events (Doody et al., 2022). Notably, there is a lack of knowledge about the role subjective factors might play in considering using these types of mobility more frequently. We could hypothesise that instrumental motives play a significant role in explaining the use of these modes, followed by the emotions associated with driving as well as with the environment. However, such claims must be tested in experiments specifically designed for this purpose.

Specifically, regarding autonomous and electric vehicles, as mentioned in section 4.3, some of the available alternatives in the market have placed themselves for the tech-savvy, early adopters, and affluent segment of the population. Interestingly, by combining these different aspects into this new car models, motives associated with the symbolic value – the position of someone in the society – and pro-environment motives – the desire to reduce direct greenhouse gas emissions – become intertwined. The outcome of this combination can lead to an increased use of cars because of the apparent innocuous effect on the environment of such electric vehicles. Future research should investigate the degree to which the different subjective factors explored in this review explain such vehicle acquisition and changes in the number of trips by car per household.

Finally, there are numerous subjective variables that play a role in explaining car ownership and usage and they are subject to complex relations and interactions. Next to those, socio-demographics characteristics, life events, the built environment, and residential location decisions are also determinants of car ownership and usage. The causality relation between each pair of variables could be bidirectional, which leads to endogeneity issues. For example, as mentioned in the introduction, there is evidence to suggest that the effect of motives on choices and the effect of choices on motives coexist. Thus, a model that aims to control for all these possible interactions simultaneously needs to be of complex nature and require sophisticated techniques to estimate each individual effect separately. If the analysis is aimed at guiding public policy, it is paramount to understand better which modelling assumptions can be reasonably made in order to avoid biased results.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was supported by the Amsterdam Institute for Advanced Metropolitan Solutions (AMS Institute).