Fragmented Realities: Middle-Class Perception Gaps and Environmental Indifference in Jakarta and Phnom Penh

Abstract

Rapid urbanization in Southeast Asia has created a paradox between severe environmental degradation and the often-muted concern of urban residents. The objective of this study is to explain this disconnect by comparatively analyzing the structural factors that shape environmental perceptions among the urban middle class in Jakarta and Phnom Penh. Drawing on survey data from over 2000 households, the study reveals two distinct narratives. In Jakarta, the middle class reports a surprisingly low frequency of environmental problems, suggesting a “perception gap” driven by physical and social insulation from the city’s harshest realities. Conversely, in Phnom Penh, residents report higher concern, but their widespread silence when asked for solutions points not to apathy but to a sense of powerlessness within a top-down governance system. We argue that apparent environmental indifference is not a uniform phenomenon but a product of distinct structural forces. The paper’s contribution is to illustrate how urban fragmentation, class-based insulation, and perceived political agency shape the relationship between environmental reality and citizen concern.

1. Introduction

Rapid urbanization in Southeast Asia has created a significant paradox: while cities like Jakarta and Phnom Penh face undeniable environmental crises, a corresponding sense of urgency is not always evident among their residents. This study investigates this disconnect, focusing on the rising urban middle class, a demographic often seen as a key agent of social change. While studies often analyze the environmental attitudes of the middle class in the Global North (e.g., [1,2,3]), the dynamics within the rapidly growing cities of the Global South remain comparatively under-theorized [4,5,6]. We use the lens of urban fragmentation to pose a central research question: How does the specific nature of urban fragmentation in different socio-political contexts shape the environmental perceptions and engagement of the urban middle class? We hypothesize that the apparent environmental indifference of this group is not a simple matter of apathy but is a phenomenon rooted in the socio-political fabric of each city. This paper will test the proposition that in Jakarta, these divisions create socio-spatial insulation, leading to a perception gap between documented reality and lived experience, while in Phnom Penh, they foster a sense of political powerlessness that stifles citizen engagement.

Jakarta and Phnom Penh serve as compelling and contrasting case studies to explore this hypothesis. Both capitals are epicenters of national economic growth and have experienced haphazard urban expansion leading to severe ecological strains [7,8]. Jakarta’s struggle with flood control is a centuries-old challenge exacerbated by unchecked modern sprawl [9], while Phnom Penh’s recent, poorly regulated development boom has created a landscape of land speculation and worsening pollution [10]. While macro-level theories of planetary urbanism [11] effectively describe the broad economic forces driving these transformations, they can obscure the granular realities of how residents navigate these environments. This study contributes by theorizing the specific mechanisms through which these macro-processes are translated into the subjective experiences that shape citizen engagement.

The different expressions of this fractured landscape are driven by distinct governance contexts. In Cambodia, an opaque, centralized system has facilitated elite control over urban development, systematically sidelining public interest and fostering political distance between the state and its citizens [12]. In post-Suharto Jakarta, despite democratic reforms, persistent clientelistic networks between political and business oligarchies continue to influence urban policy, eroding public trust [13]. These contrasting governance models provide a valuable analytical frame for understanding why outwardly similar urban challenges might produce different citizen responses. When citizens perceive governance as unresponsive or corrupt, their motivation to engage in collective action on public issues like environmental quality diminishes.

Within this fragmented system, the aspirations of the emerging middle class play an important role. Driven by material desires and upward mobility, this demographic often engages with the city primarily as a space for consumption and personal advancement [14]. A key part of our investigation, therefore, is to examine whether this focus on private accumulation and solutions (private cars, gated communities, bottled water, etc.) is a rational response to a system where collective amenities are unreliable, rather than inherent apathy. This inward turn deepens the socio-spatial fragmentation, as residents’ priorities become narrowly focused on navigating their immediate enclave, leading to a pragmatic detachment from larger, more abstract urban challenges.

This study makes three primary contributions. Theoretically, it challenges and extends existing models of environmental concern by demonstrating how structural factors like urban form and governance mediate the relationship between lived experience and perception in the Global South. Methodologically, it showcases the value of a mixed-methods approach that combines a nuanced, multi-indicator definition of the middle class with statistical analysis of both quantitative and qualitative data. Empirically, it provides a granular, comparative analysis of Jakarta and Phnom Penh, revealing distinct, policy-relevant patterns of middle-class insulation and perceived powerlessness that characterize environmental indifference in these rapidly urbanizing contexts.

2. Study Context: Urban Fragmentation in the Southeast Asian City

Jakarta and Phnom Penh, while distinct in scale and history, share a trajectory of rapid, often chaotic urbanization built on speculative development, infrastructural strain, and the rise of an aspirational middle class. Central to understanding the disconnect between environmental reality and citizen perception in these cities is the concept of urban fragmentation. In urban theory, fragmentation refers to the process by which a city breaks down into a patchwork of disparate, poorly connected parts. This process is multidimensional, manifesting as socio-spatial segregation into enclaves [15], infrastructural divides in essential services [16], and political fractures in city-wide governance. As outlined in our introduction, this paper hypothesizes that it is the specific local expression of these intersecting forms of fragmentation that shapes the divergent environmental perceptions we observe. The following subsections detail the evidence of these processes at work.

2.1. Converging Crises: Flooding, Pollution, and Congestion

The environmental challenges in both capitals are not abstract threats but daily, lived realities that degrade the quality of urban life. Jakarta, a megacity of 30 million on a deltaic floodplain, faces an existential crisis from water. With an estimated 40% of the city already below sea level, it is exceptionally vulnerable to widespread flooding, which has intensified over the last decade [9,17]. Following a catastrophic flood in 2007, the government initiated massive, capital-intensive mitigation projects, including the World Bank-funded Jakarta Urgent Flood Mitigation Project (JUFMP). However, these efforts to normalize waterways often place the blame on the urban poor living in informal settlements, leading to controversial evictions without fully resolving the underlying infrastructural deficits. In addition, the city’s former Chinese and Christian governor Basuki Tjahaja Purnama (Ahok) aggressively pursued the relocation of slum-dwellers living beside rivers to mitigate floods [18]. Ahok would also expedite the approval process of land reclamation in Jakarta Bay to create 17 artificial islands which he claimed would help fix water contamination [19]. This created a maelstrom of allegations surrounding the corrupt practices of developers involved in the reclamation [18]. Ahok would eventually lose in the 2017 election and would be jailed for blasphemy, indicating how religious and racial divisions could be mobilized against the city’s first minority Chinese governor. Phnom Penh, similarly, is defined by its relationship with water. Located at the confluence of the Mekong, Tonle Sap, and Tonle Basaac rivers, the city’s natural floodplains have been systematically compromised by rampant construction on former lakes and wetlands, which are critical for natural drainage [7]. The city’s Master Plan 2035 has proven ineffective against this poorly regulated development, lacking the specific land-use schemes or regulatory power to prevent it [10].

Beyond flooding, the air and streets of both cities are choked with pollution and traffic. In Jakarta, chronic congestion cost the city an estimated USD 5 billion in productivity losses in 2015, with a passenger car population that has grown by approximately 5% annually since 2012 [20,21]. The resulting vehicle exhaust is the main source of hazardous air pollution, a situation so dire that it prompted a successful 2021 lawsuit by a citizen coalition against top government officials for their failure to ensure clean air [22]. Phnom Penh faces a parallel crisis of mobility and waste. The number of registered motorcycles has risen by at least 10% annually for over a decade, contributing to constant noise and traffic [23]. This is compounded by a waste problem spiraling out of control; solid waste nearly doubled from 2010 to 2017, with plastic bags growing from a fraction of waste in 1999 to over 21% by 2016 [24].

2.2. Contexts of Distrust: Governance, Politics, and Public Cynicism

The persistence of these environmental problems is inextricably linked to systemic governance failures that foster deep-seated public cynicism. In Phnom Penh, urban development is largely driven by opaque partnerships between the state and foreign investors, particularly from China, which accelerated after the 2018 election [25]. With small government budgets and a reliance on foreign aid, megaprojects, high-rises, and satellite cities are often implemented with little to no public consultation [12,26]. This model has led to inflated land prices and, most critically, forceful land grabs. The infamous case of Boeung Kak Lake, where ten thousand residents were forcibly evicted from their homes after the government disregarded their land titles to grant a concession for commercial development, epitomizes a system where citizen rights are secondary to elite interests, breeding a sense of powerlessness among the populace [12].

In Jakarta, governance is characterized by a different, yet equally disempowering, dynamic. In the post-Suharto era of decentralization, a lack of local government oversight has allowed for the proliferation of superblocks and other large-scale developments that symbolize the enduring informal networks between developers and government officials [27]. Urban management is also intensely politicized. The tenure of former governor Basuki Tjahaja Purnama (Ahok) is a case in point. His aggressive environmental policies, including relocating slum-dwellers to mitigate floods and expediting controversial land reclamation projects, became entangled in a maelstrom of corruption allegations and ethno-religious politics that led to his election loss and subsequent jailing for blasphemy [18]. For the public, such events demonstrate that urban policy is often a tool of political contestation rather than a pathway to collective problem-solving. Ultimately, the Indonesian government’s decision to relocate the national capital to East Kalimantan can be seen as the ultimate act of capitulation, an admission of defeat in the face of Jakarta’s overwhelming environmental woes, reinforcing a sense of abandonment among its citizens.

2.3. The Rise of an Aspirational Middle Class

It is within this challenging urban landscape that a new middle class has emerged. Defining this demographic is far from straightforward, as researchers have employed a variety of approaches. Macro-economic definitions often rely on absolute income or consumption thresholds, such as households with daily per capita expenditures between $2 and $10 [28] or incomes between $10 and $100 [29]. While useful for tracking global trends, these broad economic frameworks are often criticized for failing to capture the sociological distinctions of status, security, and aspiration that shape behavior. Consequently, a growing body of research, particularly in the Global South, argues for multi-indicator frameworks that move beyond income. These approaches treat the middle class as a group defined by a combination of consumer capacity, educational attainment, and lifestyle markers [4,6]. For example, South Africa’s Living Standards Measure prioritizes asset ownership (cars, appliances) over raw income [5], while studies in Vietnam often use a hybrid definition combining income with occupation and education level. This context-sensitive, multi-faceted approach is relevant for providing a robust analytical foundation for a study concerned with perception and civic engagement.

This demographic’s political and social influence is growing; in Indonesia, the appeal of President Joko Widodo to the urban middle class was a key factor in his 2014 election, signaling a shift in national politics [13]. This group is central to the urban environmental paradox. On one hand, their aspirational lifestyle, characterized by rising disposable incomes and material desires [14], creates a direct tension with key international development goals, particularly the Sustainable Development Goals (SDGs). The explosive growth of private vehicles, the demand for housing in sprawling suburbs and patterns of consumption that generate significant waste are all hallmarks of middle-class life that directly challenge the objectives of SDG 11 (Sustainable Cities and Communities) and SDG 12 (Responsible Consumption and Production) [30].

This places them in a state of tension, caught between pro-social environmental values and the individualistic imperatives of a consumer culture [31]. While often educated, globally connected, and aware of environmental issues, their daily priorities are shaped by their immediate surroundings. In contexts where progress toward SDG 16 (Peace, Justice, and Strong Institutions) is lagging—characterized by unreliable public services and endemic political distrust—the rational choice for many middle-class households is to turn inward. They invest in private solutions to public problems: private cars to bypass inadequate public transport, bottled water to avoid unsafe tap water, private schools for better education, and gated communities for security. Their engagement with the city becomes less about collective well-being and more about navigating a fragmented landscape to secure a comfortable and safe life for their families. This pragmatic detachment, often perceived as indifference, is arguably a logical response to an urban system where the social contract feels broken and personal accumulation is presented as the primary path to a better life.

3. Methodology

This study employs a mixed-methods approach, analyzing a subset of data from a large-scale, multi-city comparative urban study. By repurposing relevant quantitative and qualitative data from this parent study, the current paper provides a focused, in-depth analysis of environmental perceptions among middle-class residents in Jakarta and Phnom Penh.

3.1. Data Source and Parent Study

The data is drawn from a comprehensive survey project, “Future of Asian Cities—Citizens Urban Science: Policy and Practice”, conducted between 2019 and 2022 across major Southeast Asian cities. The primary aim of the parent study was to develop a comparative understanding of urban livability, social cohesion, and infrastructural access in rapidly growing metropolitan areas. The original survey was extensive, comprising six modules that covered a wide range of topics. This paper analyzes a specific subset of questions from the neighborhood conditions and socio-demographic modules, which, while not explicitly designed to measure environmentalism, could be repurposed to assess residents’ perceptions of their immediate environmental challenges.

The authors were part of the local research teams that contributed directly to the contextual adaptation of the survey. The Jakarta team (RUJAK Center for Urban Studies) provided suggestions for survey locations to ensure representation across different settlement types, while the Phnom Penh team (Center for Khmer Studies) helped refine key terminologies to ensure cultural and local appropriateness, as detailed below. This direct involvement informs the contextual descriptions of the two cities. Portions of this rich dataset have been used in prior research on different topics, such as Benita [32].

3.2. Sampling Strategy and Site Selection

The parent study employed a multi-stage sampling strategy that combined stratified and cluster sampling. The population in each city was first stratified by geographical zones to ensure broad representation. Within each stratum, specific neighborhoods or administrative units were selected as clusters, and households were then randomly selected within these clusters. The primary inclusion criterion for respondents was an age of 18 years or older. While the sampling was systematic, it is important to note limitations regarding non-response. In Phnom Penh, it was reported that a few potential respondents declined to participate due to security or political sensitivity concerns; these were excluded from the final analysis. For Jakarta, specific household-level non-response rates were not specified in the parent study’s documentation, a limitation inherent in this secondary analysis. However, efforts were made to ensure the sample was broadly representative; for instance, the gender proportion in the Phnom Penh sample (46.3% male, 53.7% female) was found to closely match the commune-level database (48.3% male, 51.7% female).

The study sites were purposively selected to represent a spectrum of middle-class urban living (see Figure 1). In Jakarta, the neighborhoods analyzed—Tanah Merah (North Jakarta), Bendungan Hilir or “Benhil” (Central Jakarta), and Cengkareng Barat (West Jakarta)—were chosen to represent three distinct settlement types: legally informal housing (kampongs), vertical high-rise apartments, and conventionally planned neighborhoods, respectively. In Phnom Penh, four sangkats (communes) were selected to ensure broad geographical representation: two upper-middle-income areas in the urban core (Beung Trobek and Toul Sangke) and two middle- to lower-middle-income suburban areas (Kraing Thnong and Dangkao). This selection allows for a comparative analysis of perceptions between residents in the more developed city center and those in the rapidly expanding urban fringe.

3.3. Data Collection and Survey Instrument

Data collection was conducted through face-to-face interviews administered by trained local enumerators in the respective national languages (Bahasa Indonesia and Khmer). The process was technologically standardized, with surveys administered on iPads using the Qualtrics platform. Each interview was comprehensive, lasting approximately 30 to 45 min. Crucially, the study adhered to strict ethical protocols, receiving full ethical approval from the Institutional Review Board of the Singapore University of Technology and Design (protocol code IRB-19-00210).

To ensure the cultural appropriateness and clarity of the survey instrument, pilot testing was conducted. In Phnom Penh, the instrument underwent two rounds of piloting with local residents. This process was used to sharpen ambiguous questions into close-ended formats and to ensure a common understanding among the enumerators. A key outcome of this pre-testing was the careful, context-specific definition of the term “neighborhood” provided to respondents to minimize confusion. While the Jakarta team contributed to the overall design, a similar formalized pilot test of the full survey instrument prior to deployment was not explicitly documented in the parent study’s report.

This paper focuses on a specific set of questions drawn from the neighborhood conditions module. Respondents were asked to rate the frequency of six environmental problems (“Air Pollution”, “Flooding”, “Noise pollution”, “Traffic congestion”, “Water pollution”, and “Foul smells”) sing a 5-point Likert-type scale with detailed frequency anchors as follows: “Never”, “Seldom (1–2 times per 5 years)”, “Infrequently (1–2 times per year)”, “Often (50–75% per year)”, and “Always (everyday)”. The use of a 5-point Likert scale is a common approach in environmental perception research and has been shown to be useful for capturing differences in residents’ experiences across urban and suburban contexts as documented by Huddart-Kennedy et al. [33] and Ambrosius & Gilderbloom [34]. To complement this quantitative data, the survey also included an open-ended question asking what residents would change to improve their neighborhood if they were in a position to make change.

3.4. Operationalizing Middle Class

Recognizing that “middle class” represents a multifaceted socio-economic identity not reducible to income alone, this study employed a multi-indicator framework to classify households. This method provides a more robust and contextually sensitive definition than a single metric. A household was identified as middle class if it met criteria across income, education, and asset ownership, based on established definitions and markers for the Southeast Asian region.

Income. We used the income brackets defined by the Asian Development Bank Institute (ADBI) as a baseline. In this framework, households earning between USD 4–20 per day are considered to encompass the lower- to upper-middle class strata [35]. This provided a consistent economic measure across both cities.

Education. In the Indonesian context, education is a particularly salient marker of social mobility and class identity. Following guidance from the World Bank, the completion of at least junior high school was used as a key educational indicator of middle-class status [36].

Assets and Occupation. Asset ownership and occupation type were used as further proxies for economic standing. In Cambodia, middle-class status is often closely associated with occupations such as small business owners and entrepreneurs. Across both urban contexts, vehicle ownership is a significant indicator. While motorcycle ownership is nearly ubiquitous, car ownership, in particular, signals a higher level of disposable income and is a strong indicator of middle-class demand in a city like Jakarta [37].

3.5. Data Analysis

For the quantitative analysis, a Mann–Whitney U test was first employed to assess whether statistically significant differences exist between the environmental perceptions of middle-class and non-middle-class respondents. The logic of this non-parametric test is to rank the combined ordinal data for each environmental problem and then determine if the rank totals for the two groups are significantly different, indicating divergent perceptions.

For the intra-city analysis, we employ a two-step hypothesis testing procedure to compare mean perception scores across the different surveyed neighborhoods. An Analysis of Variance (ANOVA) is conducted for each environmental problem. The null hypothesis $\left(H_0\right)$ for the ANOVA is that there is no statistically significant difference in the mean perception scores among the different neighborhood groups. The alternative hypothesis $\left(H_1\right)$ is that there is a statistically significant difference in the mean perception scores for at least one neighborhood group. If the null hypothesis of the ANOVA is rejected (p-value < 0.05), we proceed to Tukey’s Honestly Significant Difference (HSD) post hoc test. This test is necessary to determine which specific pairs of neighborhoods are different from each other. For each pairwise comparison (e.g., Neighborhood A vs. Neighborhood B), the null hypothesis ($H_0$) is that the mean perception scores of the two groups are equal. The alternative hypothesis ($H_1$) is that their mean scores are not equal. This two-step process ensures a robust and detailed comparison of intra-urban perceptions.

The qualitative data from the open-ended question was analyzed using thematic content analysis. The initial step involved calculating the frequency of key words and terms to identify the most commonly cited problems and solutions, which were then grouped into broader themes. To statistically test whether the distribution of these themes varied significantly by neighborhood, Fisher’s Exact Test was employed. This test was selected over the Chi-Squared test as it provides a more accurate calculation of the p-value for contingency tables with small or sparsely populated cells [38], which was a possibility in this dataset.

3.6. Methodological Justifications and Limitations

This study’s design has several inherent limitations which have been carefully considered and must be acknowledged to properly frame the findings. First, the reliance on self-reported data means our findings are about the perception of environmental problems, not their objective, scientifically measured reality. These perceptions are subject to individual biases, including recall and social desirability bias. However, this is a deliberate methodological choice. Since the research aims to understand the subjective experiences, narratives, and potential disconnects that shape citizen engagement, this perceptual data is precisely what is required to answer the research questions.

Second, the use of a Likert scale, while a standard and effective tool for collecting standardized data across a large sample, can simplify complex opinions. Furthermore, the scale was designed to capture the perceived frequency of environmental problems, not necessarily the level of concern or the willingness to act on them. Future research could build upon these findings by employing scales that explicitly measure these other key attitudinal dimensions. We sought to mitigate this limitation by complementing the quantitative scale with an open-ended question. This mixed-methods approach allows for a richer interpretation of the results, with qualitative data helping to explain and contextualize the patterns identified in the quantitative analysis.

Third, the sampling method affects the generalizability of our findings. The purposive selection of neighborhoods as clusters means the results are representative of the residents in these specific areas but should not be statistically extrapolated to the entire middle-class population of Jakarta or Phnom Penh. The study should therefore be understood as an in-depth, comparative exploration of representative urban enclaves rather than a comprehensive city-wide survey. Finally, the different data collection periods (pre-pandemic in Jakarta, mid-pandemic in Phnom Penh) present a challenge for a direct, one-to-one comparison. The COVID-19 pandemic undoubtedly altered urban dynamics, potentially affecting traffic levels, air quality, and residents’ general sense of risk and well-being in Phnom Penh. While we acknowledge this as a limitation, the study’s core focus is on understanding perceptions as they relate to long-standing, structural issues of infrastructure and governance, which remained largely consistent in both cities despite the pandemic.

4. Results and Discussion

4.1. Respondent Profiles and a Tale of Two Urban Middle Classes

The demographic profiles of the respondents (

Table 1. Summary statistics.

Jakarta		Phnom Penh
summary statistic/frequency
Age			Age
min	18	18	min
mean	44.31	43.65	mean
max	83	96	max
sd	12.74	13.88	sd
Gender			Gender
Male	33.65%	46.26%	Male
Female	66.35%	53.74%	Female
Household monthly income			Household monthly income
Prefer not to disclose	20.38%	9.85%	Prefer not to disclose
0–268 (USD) *	50.87%	2.19%	No income
269–534 USD (USD) *	25.77%	5.20%	1–249 (USD)
535–800 (USD) *	2.40%	26.19%	250–499 (USD) *
>800 (USD) *	0.58%	28.74%	500–749 (USD) *
		11.77%	750–999 (USD) *
		8.03%	1000–1500 (USD) *
		8.03%	>1500 (USD) *
Education level			Education level
No education	4.42%	4.65%	No education *
Elementary school	16.73%	29.29%	Elementary school *
Junior High school *	21.06%	30.93%	Lower Secondary *
Senior High school *	42.69%	23.45%	Upper Secondary *
Technical school/Diploma *	5%	2.10%	Others (associate degree, technical school, etc.) *
University *	9.04%	7.76%	University *
Postgraduate *	1.06%	1.82%	Postgraduate *
Vehicle Ownership			Vehicle Ownership
No vehicle	18.14%	4.47%	No vehicle
Car *	1.44%	0.64%	Car *
Car and motorbike *	9.88%	33.57%	Car and motorbike *
Motorbike *	70.54%	49.45%	Motorbike *
		11.87%	Others (tricycle, remork motor, etc.) *
Total respondents	1040	1096	Total respondents
Middle class	385 (37.02%)	907 (82.76%)	Middle class

Note: * Inclusion criteria for middle-income households; 1 USD = 14,000 Indonesian Rupiah (IDR).

) reveal a striking initial contrast between the two cities. While the mean age was comparable (around 44 years), the composition of the samples differed significantly. The Jakarta sample contained a higher proportion of female respondents (66.4%), likely an artifact of the pre-pandemic, daytime survey schedule. More importantly, only 37% of respondents in Jakarta were identified as middle class, whereas in Phnom Penh, this figure was a commanding 83%. This disparity suggests more than just a sampling difference; it points to the distinct character of the middle class in each context. The lower proportion in Jakarta may reflect the city’s deep socio-economic stratification, while the higher figure in Phnom Penh could indicate both the success of sampling in targeted middle-income areas and a broader, more aspirational self-identification with middle-class status in Cambodia’s rapidly growing economy. This initial demographic divergence already hints at the deeply fragmented urban contexts we seek to explore, where the very definition and composition of a “middle class” is not uniform.

4.2. The Perception Gap: Jakarta’s Insulation vs. Phnom Penh’s Pervasive Reality

The core quantitative finding (Figure 2) is a clear manifestation of socio-spatial fragmentation, where the lived reality of the middle class becomes detached from the city’s documented environmental crises. In Jakarta, a city infamous for its severe pollution and flooding, over half of middle-class residents reported they “Never” or “Seldom” experienced such problems. This stands in stark opposition to Phnom Penh, where a majority reported facing environmental challenges “Often” or “Always”. We argue this perception gap is not a misreading of reality but a direct consequence of the insulated nature of middle-class life in Jakarta. This finding aligns with broader studies of urbanization in Southeast Asia, which document a turn towards private, enclosed, and consumer-focused lifestyles as a rational response to the precarity and infrastructural deficits of the megacity [14]. The proliferation of self-contained, often gated, communities (komunitas berpagar), high-rise apartments, and a heavy reliance on private vehicles creates an urban bubble that shields residents from the city’s harshest effects [39]. Within this bubble, chronic issues like traffic and pollution become normalized—an accepted, ambient feature of a fragmented city of cities [17] rather than an acute problem requiring action.

This narrative of fragmentation is further reinforced by the Mann–Whitney U test results (

Table 2. Comparison of environmental problems by middle class significance.

	Jakarta	Phnom Penh
	p-value	p-value
Air Pollution	0.049	0.573
Flooding	0.001	0.003
Foul smells	0.020	0.697
Noise pollution	0.001	0.471
Water pollution	0.009	0.159

), which statistically illustrate the different social fabrics of the two cities. In Jakarta, the statistically significant difference in perception between middle-class and non-middle-class residents across all environmental problems indicates that class is a powerful dividing line. This is socio-economic fragmentation made visible; the city is experienced in profoundly different ways depending on one’s financial ability to buy insulation from its environmental disamenities. In Phnom Penh, the lack of a class divide on most issues (except flooding) suggests that environmental problems are so pervasive they create a more shared, and arguably more accurately perceived, urban reality that cuts across class lines.

4.3. Agency and Silence: Interpreting Residents’ Voices

The contrast between the two cities deepens when examining residents’ perceived agency (Figure 3). Around 70% of Jakartan middle-class respondents offered specific ideas for what they would change, suggesting a sense of localized agency. Critically, the overwhelming silence from Phnom Penh’s middle class (85% gave no answer) is best interpreted as a symptom of perceived powerlessness rooted in a political context where citizen participation is historically limited [12,40].

A qualitative analysis of the open-ended responses, in Figure 4, reveals the distinct character of resident priorities in each city. To statistically investigate whether this localized agency was fragmented, we tested the hypothesis that the priorities residents identify are dependent on their neighborhood. In other words, we tested whether the distribution of themes was random or if certain issues were significantly more likely to be mentioned in specific locations.

Intra-City Priorities in Jakarta. Fisher’s Exact Test confirmed a significant association between a resident’s location and the primary concerns they raised (p-value < 0.001). As shown in

Table 3. Resident-Identified Priorities by Neighborhood in Jakarta.

Neighborhood	Waste	Flooding	Traffic	Total Issues Mentioned
1. Benhil	11 (26.2%)	11 (26.2%)	20 (47.6%)	42
2. Cengkareng	5 (12.2%)	28 (68.3%)	8 (19.5%)	41
3. Tanah Merah	5 (62.5%)	1 (12.5%)	2 (25%)	8
Total	21	40	30	91

Note: Table shows the frequency and row percentage of key themes from open-ended responses.

, the priorities differed starkly: traffic-related issues constituted nearly half (47.6%) of the concerns mentioned in the high-rise enclave of Benhil, whereas flooding was the dominant theme in the planned community of Cengkareng (68.3%), and waste was the top issue for Tanah Merah residents (62.5%).

Intra-City Priorities in Phnom Penh. Even among the few who answered in Phnom Penh, a similar pattern of fragmentation emerged. Fisher’s Exact Test again confirmed a statistically significant variation in themes across the different neighborhoods (p-value < 0.01). The concerns raised were markedly different from those in Jakarta, focusing primarily on basic infrastructure and public order rather than environmental issues. As detailed in

Table 4. Resident-Identified Priorities by Neighborhood in Phnom Penh.

Neighborhood	Roads	Drainage/Sewer	Security/Law	Environment/Waste	Total Issues Mentioned
1. Boeng Trabaek	1 (12.5%)	2 (25%)	1 (12.5%)	0 (0%)	8
2. Tuol Sangkae	18 (30%)	0 (0%)	12 (20%)	0 (0%)	60
3. Krang Thnong	35 (35.7%)	0 (0%)	13 (13.3%)	1 (1%)	98
4. Dangkao	2 (20%)	0 (0%)	3 (30%)	0 (0%)	10
Total	56	2	29	1	176

Note: Table shows the frequency and row percentage of key themes from open-ended responses.

, improving roads was a dominant theme, particularly in the middle-income areas of Tuol Sangkae (30.0%) and Krang Thnong (35.7%), while security and law was also a frequently cited concern.

This variation in resident-identified priorities lends support to the fragmentation thesis, showing that perceptions of the urban commons are shaped by localized challenges—ranging from consumerist concerns such as traffic in Jakarta to developmental issues like road quality in Phnom Penh. These findings suggest that residents’ awareness often remains neighborhood-bound, limiting the prospects for a unified, city-wide environmental politics.

4.4. Intra-City Fragmentation: The Neighborhood as the World

Beyond the fragmentation observed between cities and social classes, the ANOVA and subsequent post hoc tests provide the strongest empirical evidence for our thesis, revealing significant perceptual differences within the middle class of each city. This detailed analysis, summarized in

Table 5. Summary of Tukey HSD Pairwise Comparisons for Environmental Perceptions in Jakarta.

	Cengkareng—Benhil	Tanah Merah—Benhil	Tanah Merah—Cengkareng
Air Poll.	C > B (***)	TM > B (***)	ns
Flooding	B > C (**)	TM > B (***)	TM > C (***)
Noise poll.	B > C (***)	ns	TM > C (*)
Traff. cong.	B > C (***)	B > TM (**)	ns
Water poll.	B > C (**)	TM > B (***)	TM > C (***)
Foul smells	B > C (***)	ns	TM > C (***)

Note: Table entries show the direction and statistical significance of the difference in mean perception scores. B > C indicates that Benhil reported a significantly higher frequency than Cengkareng. Significance levels (in parentheses) are denoted as: ns = not significant; * = p-value < 0.05; ** = p-value < 0.01; *** = p-value < 0.001.

and

Table 6. Summary of Tukey HSD Pairwise Comparisons for Environmental Perceptions in Phnom Penh.

	Tuol Sangkae—Boeng Trabaek	Krang Thnong—Boeng Trabaek	Dangkao—Boeng Trabaek	Krang Thnong—Tuol Sangkae	Dangkao—Tuol Sangkae	Dangkao—Krang Thnong
Air Poll.	ns	KT > BT (***)	ns	KT > TS (*)	ns	KT > D (*)
Flooding	ns	ns	D > BT (***)	ns	D > TS (***)	D > KT (***)
Noise poll.	ns	ns	BT > D (**)	ns	TS > D (*)	KT > D (**)
Traff. cong.	ns	BT > KT (***)	BT > D (*)	TS > KT (***)	TS > D (*)	D > KT (*)
Water poll.	ns	ns	BT > D (***)	ns	TS > D (***)	KT > D (*)
Foul smells	ns	ns	D > BT (***)	ns	D > TS (***)	D > KT (***)

Note: Table entries show the direction and statistical significance of the difference in mean perception scores. BT = Boeng Trabaek, TS = Tuol Sangkae, KT = Krang Thnong, D = Dangkao. Significance levels (in parentheses) are denoted as: ns = not significant; * = p-value < 0.05; ** = p-value < 0.01; *** = p-value < 0.001.

, demonstrates that the neighborhood unit often serves as the primary shaper of environmental reality.

Fragmentation Patterns in Jakarta. In Jakarta, the results highlight a layered fragmentation, distinguished both by class and by the specific forms of middle-class enclaves. Two distinct profiles of heightened environmental concern emerge from the data. For lifestyle impediments, the high-rise, car-dependent community of Benhil is a clear hotspot. As shown in Table 5, perceptions of Traffic Congestion in Benhil were significantly higher than in both the planned neighborhood of Cengkareng (p-value < 0.001) and the informal settlement of Tanah Merah (p-value < 0.01). Conversely, for issues related to public health and basic infrastructure, the legally informal settlement of Tanah Merah stands out. Residents there reported significantly higher frequencies of Air and Water Pollution compared to both other neighborhoods (p-value < 0.001). Cengkareng, the more conventionally planned area, generally serves as a perceptual baseline, though its residents report significantly more frequent Flooding than those in Benhil (p-value < 0.001). This pattern clearly illustrates how the specific form of the built environment structures perception, creating divergent realities even among households of similar economic standing.

Fragmentation Patterns in Phnom Penh. In Phnom Penh, the results indicate a clear urban–suburban divide, with the suburban fringe bearing the brunt of unregulated development. The results (Table 6) are dominated by the unique experience of Dangkao (see Figure 1), a commune on the developing periphery. Dangkao residents reported significantly higher perceived frequencies of infrastructure-related problems like Flooding and Foul Smells compared to all three other surveyed areas (p-value < 0.001 in all pairwise comparisons). This highlights the severe environmental costs borne by those living on the city’s expanding edge. In contrast, for problems associated with density and pollution, the urban core is not a uniform experience. Perceptions of Noise and Water Pollution were significantly higher in the central neighborhoods of Boeng Trabaek and Tuol Sangkae than on the fringe. This pattern of core-periphery fragmentation suggests that environmental concern is primarily structured by a neighborhood’s position within the city’s uneven development trajectory.

By moving beyond city-wide averages, this intra-city analysis demonstrates that the lived experience of the environment is hyper-localized. When a resident’s world of concern is defined by the specific realities of their immediate enclave—be it traffic in a high-rise or flooding on the fringe—it creates a powerful barrier to forming the unified, city-wide consciousness necessary for collective environmental politics.

5. Conclusions

This study has examined the multifaceted and often contradictory nature of environmental perceptions among middle-class residents in Jakarta and Phnom Penh. Our findings reveal a significant perception gap in Jakarta, where a sense of insulation and normalization appears to mute the recognition of severe environmental problems. In contrast, perceptions in Phnom Penh seem to more closely align with the city’s pervasive challenges, yet this awareness is coupled with a profound sense of powerlessness. Across both cities, we find strong evidence that the urban experience is highly fragmented, with residents’ concerns shaped more by the specific conditions of their local enclave than by a sense of a shared urban environment. A key contribution of this paper, therefore, is to contribute a more context-specific understanding of the relationship between lived experience and environmental concern. While many models effectively document differences in concern between, for instance, urban and rural populations (e.g., [33]), our findings show evidence that this relationship is powerfully mediated by structural factors, such as urban form, class stratification, and the nature of governance, that are particularly pronounced in the fragmented cities of the Global South. We show that in these contexts, perceived “indifference” is not a simple matter of apathy, but can be a rational outcome of middle-class insulation, the normalization of chronic dysfunction, or a pragmatic response to perceived powerlessness. This moves the discussion beyond individual attitudes to consider the structures that shape them. This insight extends beyond Southeast Asia, contributing to a broader theory of urban political ecology in the Global South. The mechanisms of fragmentation and insulation identified in Jakarta resonate with middle-class experiences in São Paulo’s gated communities (condomínios fechados) of São Paulo [41] or Johannesburg’s securitized suburbs [42], where private infrastructure often substitutes for state provision. Similarly, the sense of political powerlessness observed in Phnom Penh reflects the widespread cynicism toward unresponsive governance in many rapidly growing cities across Latin America and Sub-Saharan Africa. By showing how local conditions produce systematic environmental disconnects, this study offers a transferable framework for analyzing a key challenge of the 21st-century urban condition.

These findings carry significant policy implications, suggesting that generic environmental awareness campaigns are likely to fail if they do not address the specific form of fragmentation in each city. For a Fragmented and Insulated city like Jakarta, the policy challenge is to make the consequences of collective problems tangible. For instance, municipal environmental agencies, in partnership with local community councils (Rukun Warga), could deploy hyper-local air quality sensors in middle-class neighborhoods and disseminate this real-time data through community-specific apps to make invisible threats visible and immediate. Such uses of digital technology represent a powerful tool for bridging the perception gaps created by physical insulation, a key challenge in contemporary urban governance. Furthermore, fiscal policies can pierce the bubble of insulation; city revenue departments could pilot measures such as localized congestion pricing in affluent business districts or progressive property taxes that penalize low green-space ratios, directly linking lifestyle choices to financial consequences. More broadly, the digitization of urban governance offers a way to overcome perceptual disconnects caused by fragmentation. By aggregating hyper-local, real-time data, digital platforms can create a shared understanding of the urban environment. Yet without equitable access and responsive institutions to act on this information, such technologies risk remaining symbolic rather than transformative.

In contrast, for a context of a Powerless and Cynical citizenry like that of Phnom Penh, the imperative is not simply to raise awareness but to build institutional trust and foster genuine agency. To achieve this, municipal authorities can build trust by creating public, accessible dashboards that track the budgets, timelines, and progress of key infrastructure projects, countering the prevailing narrative of opaque governance. To combat the pervasive sense of powerlessness, national development agencies could fund and pilot small-scale, community-led projects where residents in specific sangkats are given real decision-making power and budgets to address their self-identified local priorities. Such an initiative would require a clear mandate from the municipal authority for the sangkat chief to oversee implementation, with transparent reporting on outcomes, thereby rebuilding a sense of efficacy and trust from the ground up.

Finally, this study’s limitations, particularly its cross-sectional design and the different data collection periods, highlight avenues for future research. Longitudinal studies are needed to track how perceptions evolve over time in response to policy changes or worsening environmental conditions. Furthermore, our survey’s focus on the frequency of environmental problems successfully captured experiential exposure, but future work could employ scales that explicitly measure other crucial dimensions like attitudinal concern, willingness to act, and institutional trust. In-depth ethnographic research could then offer a deeper understanding of the contradictions surfaced by this survey, illuminating the cultural and political frameworks that shape environmental attitudes within this influential urban demographic.