Group Size and Social Capital Impact on Free Riding Behavior in Padang’s Self-Guards Managed Railway Crossing

Railway level crossings are the most hazardous locations where roads and railways meet. This critical intersection needs strict safety measures to minimize accidents and ensure the community is safe. However, in Indonesia’s Padang, these crossings have seen disturbing increases in incident trends. Recent figures from the BTP Sumbagbar Railway Engineering Centre show that, between 2017 and 2022, 123 accidents occurred there, leading to 16 deaths, 49 injuries, and 39 minor injuries [1]. Risk is made worse by the swift increase of incidents in 2023, with 29 total accidents within a year, resulting in grief as all resulted in fatalities and many people were injured as well. This situation illustrates the urgent need for local safety initiatives integrated with transportation planning [2].

One of the main problems that contributes to the grievous number of accidents is the spread of illegal or unregistered level crossings. Out of the 388 crossings that were counted, only 42 reported guard presence, while 58 were registered with no guarding presence, and a staggering 288 were illegal crossings [3]. This situation grows more aggravated when level crossings and residential neighborhoods are considered in the region of Padang. Of the 30 residential locations, whose areas are cut across by railway lines, only 11 have guards who are provided by the residents, while 19 do not have guards. These spontaneous safety interventions show similarities to cooperative grassroots risk governance frameworks [4].

This scenario poses a significant risk to residents and road users. The prevalence of unlicensed or unguarded level crossings in Padang's residential areas forced local communities to adopt self-managed safety initiatives. Residents organized and funded railway crossing guards to reduce risks in this high-traffic area. This reflects a pattern of collective problem-solving in high-risk areas. In line with the broader findings on community participation in improving safety [5]. The challenge of unguarded level crossings is not unique to Padang. Research from other urban settings confirms that community collaboration is a valuable supplement to official transport safety systems [4, 6].

However, these grassroots efforts are often undermined by the problem of free-rider. In this context, “free-rider” benefits from safety improvements provided by publicly-funded safety officers but does not contribute financially to their maintenance. This has been discussed extensively in public goods literature. Free-riders tend to increase with group size [7] in the absence of mutual obligations [8, 9]. Additionally, behavioral research shows that free-riding is more likely when individuals perceive low personal gain or reduced visibility of their actions within a group [10]. Social cohesion can counteract the dilution of responsibility within a larger group. As such, it can encourage greater commitment to collective safety.

Several argue that as group size increases, monitoring individual contributions becomes more difficult, leading to more opt-outs [7]. Another supports this view, stating that larger communities face unique constraints due to the dispersion of responsibilities among residents [11]. In contrast, other studies have challenged this perspective. They propose that strong social networks within a community can ward off opt-outs [5]. This, in turn, can foster a greater sense of responsibility among residents [12], especially where trust networks are dense [13].

In residential areas, the effectiveness of community-managed safety measures depends not only on the size and composition of the group. Various internal social dynamics also determine the provision of informal railway crossing guards. These dynamics include social capital, neighborhood cohesion, and household characteristics [14]. Social capital plays an important role in addressing the issue of free-riding in community-driven safety initiatives [15]. Defined as the shared networks, trust, and norms that facilitate cooperation, it enables communities to mobilize resources at the grassroots level and fosters accountability [13, 16].

Research emphasizes the potential of social networks to complement technological solutions in transportation safety systems. It highlights their importance in encouraging collective action. Inclusive and cohesive environments, supported by strong social capital, are better equipped to sustain safety initiatives. Even in the face of resource limitations and increasing group size [5, 17]. Conversely, weak social capital may exacerbate free-rider behavior [18].

The residents may feel less obligated to contribute when trust and cooperation are lacking. Residents of a residential area who tend to stay for a longer period tend to develop deeper relationships with their community. This potentially fosters a stronger sense of attachment and responsibility [19]. These residents are more likely to participate in neighborhood activities. They have a higher intention to support public goods, and contribute to common goals [20, 21].

In contrast, newcomers or temporary populations may not have the time or opportunity to integrate into the community. This ultimately results in lower levels of participation and a greater tendency to jump on the bandwagon [22]. However, the positive impact of length of residence on social capital is not universal. Furthermore, in disadvantaged or declining neighborhoods, prolonged exposure to adverse conditions such as economic stagnation [23], social fragmentation, and psychological distress [24]. All of these conditions can weaken trust, reduce participation, and erode social cohesion [18].

Related to the concept of community size discussed earlier, the complexity of interactions is exacerbated. It is true that smaller groups–close relationships and more frequent interactions–facilitate effective monitoring and informal enforcement mechanisms [11]. This potentially reduces the prevalence of free-riding [12]. In contrast, larger groups face significant challenges in sustaining collective action due to increased anonymity and reduced accountability [17]. Nonetheless, strong endogenous social capital can mitigate these challenges by fostering norms of reciprocity and shared responsibility [15].

This is what enables larger communities to sustain collaborative efforts. The findings highlight the duality of the length of residence as a potential driver and inhibitor of community contributions [24]. Even if most of the members of these larger communities live in communities with strong social capital built up over a long period of residence, this potentially facilitates the act of contributing collectively [19]. What emerges more strongly are the positive effects of length of residence. The reverse is also true [20].

The length of residency also plays a pivotal role in shaping social capital and its relationship to collective action. The study finds that longer durations of stay in disadvantaged neighborhoods are associated with reduced social capital [23]. Impact, weakening its membership in social clusters. This suggests that prolonged residence in such areas leads to social fragmentation [25]. Further erodes the strength of community ties and hampers collective action. These findings are in line with recent research on neighborhood dynamics. It emphasizes that longer residency in deprived areas can foster disengagement rather than cohesion [18].

These dynamics underscore the importance of understanding how social capital, group size, and length of stay interact to shape contributions to grassroots initiatives. Particularly in high-risk contexts such as unguarded railroad crossings. While there is substantial research on social capital and the provision of public goods, limited attention has been paid to how these dynamics work. It specifically affects community-driven safety measures at railroad crossings [2]. Existing studies often focus on traditional public goods, such as parks or roads. It ignored the unique challenges posed by high-risk safety initiatives in residential areas.

This study addresses this gap by examining the role of exogenous social capital and group size in moderating the impact of length of residence on contributions to self-funded railway crossing guards in Padang. Technically using two-stage probit regression, this study investigates how factors such as neighborhood cohesion and length of residence shape participation in community safety initiatives. Based on collective action theory, the study hypothesized that group size strengthens the relationship between social capital and contribution, with length of residence acting as an important moderating factor.

The research design of this study is structured to investigate the dynamics of free-rider behavior concerning the provision of public goods, particularly in the context of railway crossing safety, within the urban setting of Padang city. The research focused on three main sub-districts in Padang that are predominantly traversed by level railway crossings: Padang Timur, Padang Utara, and Koto Tangah. These areas have a significant number of new residential developments intersected by railway tracks.

Koto Tangah sub-district is particularly notable, as it has the largest population among the sub-districts. Given the substantial population and extensive railway crossings in Koto Tangah, this area was chosen for a detailed study. The residential areas in Koto Tangah, as observed through Google Earth, are characterized by long stretches of railway tracks passing through them. To narrow down the sample, the analysis was further focused on individual neighborhood within these areas. The use of satellite and participatory mapping tools has been widely used in transport studies where official data is scarce [4].

Some neighborhoods in Koto Tangah have multiple residential areas intersected by railway tracks. In selecting specific neighborhood, those closest to the railway tracks were prioritized. Due to the unavailability of official data neighborhood intersected by railway tracks, we collected data through our resources. This data collection involved taking photos using Google Earth to generate maps of residential areas, railway tracks, and the locations of level crossings.

Neighborhoods were stratified based on demographic characteristics, socioeconomic status, and proximity to railroad crossings. Specifically, we categorized individuals as free-rider enthusiasts or avoiders based on their contributions towards self-help railway crossing guards. Only neighborhoods with level railroad crossings that residents pay for and have unofficial crossing guards were included in the study. We excluded residents who did not contribute and neighborhoods without resident-provided or PT KAI (Kereta Api Indonesia)-provided crossing guards.

This stratification ensures a diverse sample, allowing for broader generalizations about free-rider behavior in various contexts. Within each stratum, a systematic random sampling technique was used to select 20 households. This method minimizes selection bias and ensures each household has an equal chance of being included in the study. Each neighborhood was carefully selected to represent Padang's diverse urban landscape. This took into account factors such as population density, economic diversity, and the prevalence of unofficial railroad crossings [5].

3.1. Measurement of variables

Dependent Variable (Y). The dependent variable used in this study is the tendency of free riders. Categorizing free rider tendencies based on contribution behavior facilitates the assessment of individual involvement in collective endeavors. Free rider tendencies use dummy variables. Using dummy variables to operationalize free rider tendencies allows for quantitative analysis of individual behavior within the community.

The operationalization of free-rider behavior draws from the conceptual framework in the International Encyclopedia of Ethics, which distinguishes between passive and active free-riding [30]. That is a distinction reflected in our binary classification of contribution behaviors. In addition, Kantian perspectives offer normative reasoning on why participation in communal safety efforts is a moral obligation [31].

If a person's contribution is below the average contribution of residents of each housing, it is categorized as low, and then given a value of 1. Where the person can be categorized as tending to behave free rider (free rider lover). Conversely, if a person's contribution is above the average contribution of residents of each housing, it is categorized as high, then given a value of 0. Where the person can be categorized as tending to behave not free-rider (free-rider averter). Previous studies show that communal norms reduce free-riding. Previous research has shown that communal norms significantly reduce free riding behavior. This supports the inclusion of social factors in behavioral modeling [8, 15]. By discerning between low and high contributors, the study captures variances in behavior and perceptions concerning collective action and public goods provision.

Main Independent Variables (X) is group size, operationalized as the average household size within the residential area. Researchers often employ various measures, including demographic data, social network structures, and qualitative insights, to assess group size accurately. Group or community size refers to the number of individuals within a social unit, such as a neighborhood or organization. Researchers assess group size using various measures, including demographic data and social network structures. The definition and operationalization of community size directly impact the analysis and interpretation of free-rider tendencies in safety initiatives at railway crossings [11]. So, in this case, the group size is derived from the average family size in each residential area, and serves as a variable. This variable, representing the number of family members in sampled households, reflects the community's scale [10].

Our measurement of social capital builds upon a rigorous examination of homeownership's effects on social capital formation [16]. Their findings validate our use of residency duration as a proxy for social capital accumulation [28]. Particularly, their demonstration of how stable housing tenure strengthens community ties and collective action capacity. The concept of social capital is widely recognized as the networks, norms, and trust that facilitate coordination and cooperation for mutual benefit.

Scholars have argued that individuals who live in a community for a longer period tend to accumulate more social capital through social interactions and the establishment of trust-based relationships. The length of stay in a neighborhood is positively correlated with the formation of more stable and stronger social ties [19]. It is essential for collective action and social cohesion. Other researchers argue that residents who have lived relatively long in a neighborhood often participate more actively in community organizations. They essentially engage in volunteer work and contribute to the social fabric of their community [21].

This participation helps individuals integrate into social networks. In turn enhances their social capital. Individuals who stay longer in one place are more likely to develop personal relationships with neighbors. Then, participate in local events and build mutual trust that reinforces community bonds [14]. In contrast, short-term residents may face barriers in accumulating social capital because of a lack of stability in relationships and social connections. Communities with higher turnover rates often experience lower levels of social capital [23].

The research employs Binary Probit Regression as the primary data analysis method. This is chosen due to the nature of the data, where the dependent variable is non-metric and the independent variables are a mix of continuous and categorical. Probit regression is suitable for binary outcomes, making it ideal for predicting the probability of free-rider behavior based on various independent variables. The Probit regression model is represented as:

$Y_i=\beta_0+\beta_1 X_1+\beta_2 X_{2+} \beta_i \sum_i^n X_i+\varepsilon_i$    (1)

Other control variables are shown below in Table 1.

3.2 Stage probit regression

Using the duration of residency in a neighborhood as an instrumental variable (IV) is a theoretically grounded and empirically robust approach to analyzing the interplay between social capital, group size, and collective action. Residency duration is a natural proxy for social capital. Longer-term residents tend to form stronger trust networks, deeper social ties, and greater accountability within their communities [16].

These dynamics significantly enhance participation in public goods provision. The public good such as community-driven safety initiatives at railway crossings. Its provision may be by facilitating sustained engagement and reducing free-riding behavior. Empirical evidence supports this relationship. It highlights how extended tenure in a neighborhood strengthens social cohesion and fosters norms of reciprocity and shared responsibility [27].

The relevance of residency duration as an IV lies in its strong correlation with social capital. However, this relationship is not uniform across contexts. In stable and affluent neighborhoods, long-term residency generally enhances collective action. Thus, creating environments conducive to community safety efforts. Conversely, in disadvantaged or declining neighborhoods, prolonged exposure to adverse conditions such as economic stagnation or social fragmentation. It can erode trust, reduce participation, and weaken social cohesion [29].

This dual nature of residency duration is a potential enhancer and inhibitor of social capital. It will provide a nuanced lens for examining variations in collective behavior. Residency duration is also exogenous to the specific outcome of free-riding behavior, satisfying the exclusion restriction criterion for IV validity. While it influences social capital, it is not directly determined by individual decisions to participate in or free-ride on community safety initiatives. Instead, residency duration is shaped by external factors such as housing markets, employment opportunities, and family dynamics [32]. Those facts are unrelated to free-riding behavior. Furthermore, residency duration affects free-riding only indirectly through its impact on social capital, ensuring the instrument's validity.

$Y_i=\beta_0+\beta_1 X_1+\beta_2\left(X_2=X_z\right)+\beta_i \sum_i^n X_i+\varepsilon_i$       (2)

where, $\beta_2$ represents the coefficient predicted social capital $\left(X_2\right)$ from residency duration variable $\left(X_z\right)$. A significant coefficient confirms that social capital, instrumented by residency duration, influences the likelihood of participation or free-riding.