Technological Progress and Chinese Residents’ Willingness to Pay for Cleaner Air

Abstract

This study examines whether China’s rapid spread of internet and mobile information technologies has translated into greater household support for government air-quality programs. Using nationally representative data from the Chinese General Social Survey (2018), this study estimates the causal impact of digital media use on residents’ willing to pay (WTP) each month for one additional “good-air” day. Ordinary least squares shows that individuals who rely primarily on the internet or mobile push services are willing to contribute CNY 1.9–2.7 more—about 43 percent above the sample mean of CNY 4.41. To address potential endogeneity, we instrumented digital media adoption using provincial computer penetration; two-stage least squares yielded roughly CNY 10.5, confirming a causal effect. Mechanism tests showed that digital access lowers complacency about local air quality, strengthens anthropogenic attribution of pollution, and heightens the moral norm that economic sacrifice is legitimate, jointly mediating the rise in WTP. Heterogeneity analyses revealed stronger effects among high-income households and renters, while extended tests showed that (i) the impact intensifies when the promised environmental gain rises from one to three or five clean-air days, (ii) attention to international news can crowd out local WTP, and (iii) digital media raise not only the likelihood of paying but also the amount paid among existing contributors. The findings suggest that targeted digital outreach—especially messages with concrete, locally salient goals—can substantially enlarge the fiscal base for air-quality initiatives, helping China advance its ecological-civilization and dual-carbon objectives.

1. Introduction

In recent decades, the intensification of air pollution across the globe has posed a severe threat to public health and sustainable socioeconomic development. According to the World Health Organization [1], nearly the entire world’s population (99%) breathes air that exceeds recommended quality limits, which places them at elevated risk of respiratory diseases and other chronic conditions. A record number of over 6000 cities in 117 countries and regions are monitoring air quality, yet their residents still inhale hazardous levels of fine particulate matter and nitrogen dioxide, with low-income populations disproportionately affected. Prolonged exposure to harmful particles or toxic gases not only increases the likelihood of various long-term illnesses but also exerts profound impacts on healthcare systems and economic structures. Against this backdrop, governments worldwide have escalated their efforts to curb air pollution through enhanced regulations, market-based measures, and broad social engagement.

Despite such endeavors, the process of improving air quality and bolstering environmental infrastructure often encounters formidable resistance. First, some heavily polluting enterprises are reluctant to comply with stricter emission standards or additional expenditures on pollution abatement, deeming these measures costly and detrimental to short-term competitiveness [2]. In regions whose energy mix relies heavily on coal or petroleum, the challenges of transitioning toward greener technologies are especially daunting, frequently resulting in passive resistance or evasive tactics. Second, large-scale environmental projects and advanced pollution-control technologies typically demand considerable fiscal outlays; relying solely on government budgets may be insufficient to meet the escalating demands of environmental governance in the short run. At the same time, public support for environmental policies is not guaranteed. In the absence of collective buy-in or effective outreach, some groups may question the transparency and efficiency of governmental spending on environmental programs, thereby weakening support for tax increases or user fees [3]. Consequently, if governments rely exclusively on regulation and public finances—without mobilizing contributions and cooperation from diverse societal actors, including ordinary households—resistance from polluting enterprises and other dissenting stakeholders will intensify, leading to suboptimal environmental outcomes.

In China, rapid economic expansion and ongoing industrial restructuring have elevated environmental protection to a core national priority, making air-quality improvement central to achieving the nation’s “ecological civilization” agenda and “dual-carbon” targets. Although official statistics report a downward trend in PM2.5 and PM10 concentrations in key urban centers, the overall challenge remains significant. One complication lies in the relocation of high-pollution industries to remote areas, which may shift pollution burdens rather than eliminate them. Another arises from the complexities of imposing stringent regulations on major polluters, where local protectionism and rent-seeking behaviors can thwart decisive policy implementation. Moreover, air pollution abatement exhibits the characteristics of a public good, necessitating not only government-led interventions but also substantial financial backing and broad-based public endorsement to achieve lasting impact [4,5]. In this context, residents’ willingness to pay (WTP) for environmental initiatives assumes critical importance: although individual contributions may seem modest relative to the overall governance costs, they can collectively alleviate government fiscal pressures and strengthen external oversight of high-pollution industries.

Notably, the rise of the internet and information technology opens new avenues for elevating public environmental awareness and willingness to pay. Historically, most people relied on traditional news outlets to learn about pollution, preventing them from easily accessing up-to-date, location-specific data on air quality. However, with the advent of social media, real-time monitoring platforms, and environmental apps, individuals can now promptly obtain granular information on pollution levels and corporate violations, thereby enhancing their understanding of environmental policies [6]. On the one hand, the widespread accessibility and transparency of these digital channels alert citizens to the health risks linked with environmental degradation; on the other hand, they help forge social consensus, encouraging more individuals to view environmental improvements as a collective interest worth supporting [7,8,9].

Moreover, rapid digital communication reshapes how the public perceives and engages in environmental issues. Various nonprofit organizations, environmental advocates, and green startups leverage online platforms for fundraising, social mobilization, and citizen-led watchdog efforts, lowering the barriers to environmental participation [10,11]. Such community-driven interaction not only bolsters public attention to ecological concerns but also fosters a sense of shared responsibility, prompting more proactive responses when governments propose new environmental taxes or fees [12]. Existing scholarship, however, has primarily examined the macro-level influences of government financing and market mechanisms on environmental outcomes [8,13], with relatively limited inquiry into how “technological progress in information dissemination” might shape individual willingness to pay for environmental protection. Even studies that do mention internet-based advocacy typically provide only preliminary evidence that online media may enhance environmental awareness, neglecting to offer rigorous tests of whether, and how, digital access concretely raises individuals’ willingness to finance public efforts to combat air pollution.

Against this backdrop, this study aims to delve deeper into the impact of technological advancement on residents’ environmental willingness to pay, focusing on the transformative role of digital information flows in the internet era. Drawing on data from the 2018 Chinese General Social Survey (CGSS), we analyze how reliance on internet or mobile-based news affects individuals’ willingness to pay for government-led environmental interventions aimed at adding one more day of “good” air quality each month. Our findings indicate that, relative to those using traditional media, residents relying primarily on online or mobile news exhibit an average increase of CNY 1.911 in their monthly willingness to pay. Given that the overall mean WTP is CNY 4.41, this translates into a 43% rise in one’s predisposition to fund cleaner air—a substantively large effect. We further verify these results by employing an instrumental variable approach and conduct robustness checks that reaffirm our core conclusions. In addition, our heterogeneity analyses suggest that the positive effect of information technology is pronounced among higher-income households and renters. Mechanism tests reveal that internet usage fosters a stronger pro-environment stance, such as the conviction that economic sacrifices are acceptable to protect the environment. Extended analyses also highlight that residents anticipate paying even more if the government pledges to secure three or five additional days of satisfactory air quality each month. Moreover, among individuals who already exhibit a positive WTP, digital information access significantly raises the magnitude of their contribution. Interestingly, however, we find that those who rely on online or social media channels for international news tend to lower their local environmental WTP, possibly due to attention shifts toward global issues at the expense of local concerns.

As digital platforms, social media, and smartphone applications become ever more ingrained in daily life, the public increasingly acquires environmental information with greater speed and diversity, thereby assuming a more active role in monitoring and understanding governmental environmental projects. By capitalizing on this “information dividend”—for instance, by disseminating compelling data on pollution severity and stressing the importance of public contributions—policymakers and civil society organizations may significantly enhance social engagement. Such efforts can help mitigate fiscal burdens on governments, curtail resistance from polluting enterprises or local vested interests, and foster more effective implementation of environmental policies on a broader scale. In this sense, our study underscores the critical role of digitalization in shaping collective environmental action, offering timely insights for ongoing debates on how to align ecological transformation with the rapid expansion of the information era.

Figure 1 depicts the causal logic that guides our empirical tests. At the top, technological progress in information access—captured by residents’ reliance on internet or mobile-push news—constitutes the exogenous stimulus. Drawing on risk-perception theory and the value–belief–norm (VBN) model, we argue that this stimulus operates through two latent psychological blocks: heightened risk perception and an activated value–belief–norm chain. Concretely, digital media (i) make respondents more critical of past local air quality (Statement 1), (ii) strengthen their belief that energy use causes major ecological problems (Statements 2 and 3), and (iii) legitimize a moral norm that sacrificing private income for environmental protection is acceptable (Statement 4). These three micro-level cognitions form a sequential pathway that elevates the household’s monthly willingness to pay (WTP) for one additional “good-air” day, shown by the solid downward arrow. Accordingly, the framework generates two testable propositions: digital information raises WTP on average, and this effect is mediated by the three numbered cognitive shifts. Subsequent sections estimate the direct effect and probe the mediators.

The remainder of this paper is structured as follows. Section 2 reviews the relevant literature. Section 3 presents the corresponding data, variable settings and estimation methods. Section 4 reports the empirical results. Finally, the conclusions and implications are given in Section 5. Section 6 discusses the limitations and future directions.

2. Literature Review

2.1. Research on Government Environmental Governance

Government-led environmental governance is a focal topic in both academic and public discourse. On the one hand, governments—acting as the primary providers of environmental public goods—shoulder multiple responsibilities, such as formulating regulatory policies, supervising corporate emissions, and investing in environmental infrastructure. On the other hand, governance outcomes are often shaped by institutional mechanisms, the distribution of interests, and the degree of public participation. Existing studies commonly contend that mere reliance on top-down administrative mandates or direct control may fail to mobilize sufficient societal resources and public engagement [14,15,16]. Hence, striking a balance between strong governmental leadership and broad-based societal participation has emerged as a key theme in contemporary research on environmental governance.

When examining the effectiveness of government environmental governance and its driving factors, scholars typically focus on the design of policy instruments and institutional arrangements. Several studies indicate that stringent regulations and emission standards can reduce high-pollution outputs in the short term. Nevertheless, when firms encounter substantial transition costs or intense market competition, resistance and evasive strategies often undermine policy enforcement [17,18]. Moreover, local protectionism—or conflicts between local fiscal objectives and environmental goals—can result in lax enforcement, causing “layer-by-layer attenuation” or implementation biases [19]. These challenges underscore that policymakers must account for regional economic structures, firm-level development stages, and local political interests to achieve a more balanced synergy between environmental protection and economic growth.

A complementary body of literature emphasizes the “fiscal and resource constraints” governments face in delivering environmental governance. Effectively managing air pollution and similar environmental concerns typically requires large-scale funding—from deploying advanced emission control technologies to building renewable energy infrastructures and monitoring systems—costs that may exceed government budgets or earmarked public funds [20]. Under such constraints, authorities must not only secure alternative financing but also rally public participation to share the burden of environmental investments. Compared with compulsory levies, broader publicity campaigns and policy incentives that encourage households to assume part of the cost (e.g., purchasing green electricity, paying environmental fees, or donating to conservation projects) can alleviate fiscal pressures while enhancing the legitimacy and sustainability of governance measures.

Consequently, a number of scholars stress that the success of government environmental governance hinges on public understanding and cooperation. Viewing citizens as crucial stakeholders in the provision of environmental public goods, these studies assert that enhancing residents’ environmental awareness and willingness to pay can directly bolster the scope and efficacy of local governments’ policy execution [21,22,23]. Under this perspective, citizens are not merely passive recipients of environmental measures but may also serve as key overseers and facilitators in the policy process. Should the public lack environmental responsibility or fail to recognize the long-term benefits of ecological initiatives, resistance to environmental taxes or green spending may intensify, thereby undermining the policy’s societal foundation. In contrast, government authorities that skillfully leverage internet platforms and digital media—through transparent information sharing, online engagement, and collective mobilization—stand a greater chance of forging public consensus and mitigating resistance to subsequent environmental measures [24,25].

Empirical evidence suggests that surmounting obstacles to environmental governance and achieving sustainable improvements in air quality and other ecological outcomes calls for robust incentive mechanisms and effective collaboration across governments, enterprises, and the public. For future research, amid the rapid expansion of the digital economy and information-driven society, there remains ample scope to investigate how internet-based tools can further heighten public awareness and policy compliance, thereby reducing governance challenges during policy implementation. In practical terms, the integration of refined, data-oriented, and cooperative approaches in government-led environmental initiatives will likely serve as a cornerstone for advancing green development and ecological civilization.

2.2. Technological Progress and Residents’ Awareness of Environmental Protection

With the rapid proliferation of advanced technologies in households, a growing body of literature has begun to explore how the internet shapes residents’ daily lives. In particular, the swift expansion of digitalization and information technology has made it easier for the public to access environmental knowledge and real-time ecological data, thus strengthening pro-environment attitudes and behaviors to a certain degree [26,27]. Social media platforms, online news outlets, and digital communities have dramatically lowered the barrier to disseminating information about pollution incidents, extreme weather, and global climate trends. Through sharing, reposting, and commenting, individuals can swiftly broadcast their perspectives, thereby deepening collective engagement with environmental issues [28,29]. Additionally, some scholars highlight that digitalization has propelled the growth of eco-focused nonprofit and science-communication platforms. Online donations, crowdfunding, and virtual community events have significantly reduced the cost of participation, making it easier for the public to engage in environmental practices, whether through small financial contributions or viral content sharing [30,31].

Nevertheless, other studies caution that technology adoption does not necessarily lead to a broad-based rise in environmental consciousness; in some contexts, it can even diminish individuals’ concern or confine it to surface-level displays. Certain researchers point to the fragmented nature of online environments and the entertainment-oriented culture of social media, which can give rise to “rapid attention and rapid forgetting”. In other words, while an ecological crisis may briefly attract intense focus online, it can be swiftly overshadowed by newer, trending topics, resulting in relatively brief windows of public concern [32,33]. At the same time, “slacktivism”—symbolic support for causes through simple acts such as likes, shares, or short comments—has become widespread. Although users may signal environmental advocacy in a virtual context, this engagement seldom translates into substantial offline actions or monetary support [34]. Furthermore, the fragmented and entertainment-driven nature of social media can overemphasize individualism and immediate gratification at the expense of collective environmental responsibility and long-range benefits.

Therefore, scholarly research on the nexus between technological progress and residents’ environmental awareness has shifted toward more nuanced, context-specific analyses. On the one hand, it is increasingly necessary to factor in broader socioeconomic conditions, governmental policies, and cultural values to understand how different regions and demographic groups adopt technology and develop ecological consciousness. On the other hand, scholars are increasingly employing mixed-methods approaches that combine quantitative and qualitative techniques—such as big-data analyses of social media discourse, behavioral tracking, surveys, and experimental designs—to more thoroughly evaluate how technologies shape environmental attitudes and behaviors [35,36]. Overall, existing studies converge on the view that the rise of digital platforms and internet technologies has indeed facilitated the public’s exposure to, and exchange of, environmental information. Nevertheless, how to overcome content overload and the lure of entertainment media so that environmental awareness is both internalized and transformed into concrete behavioral changes remains an open question requiring further in-depth investigation.

2.3. Research on Residents’ Willingness to Pay for Environmental Protection

As environmental challenges become more acute, and given that government-led solutions often rely on citizens’ financial contributions and consumption choices, scholars have increasingly turned their attention to residents’ willingness to pay (WTP) for environmental protection. In most empirical studies, researchers employ stated-preference methods such as the Contingent Valuation Method (CVM) or Choice Experiments to estimate individuals’ WTP under specific environmental scenarios, and then explore the drivers and underlying mechanisms of these decisions [37].

A foundational strand of this literature examines how personal characteristics—including income, educational attainment, and environmental literacy—shape people’s willingness to pay [38]. It is frequently observed that individuals with higher incomes and more years of education exhibit a stronger sense of environmental responsibility and a correspondingly elevated WTP [39]. In parallel, scholars have highlighted the pivotal role of environmental knowledge and pro-environment attitudes; those with a deeper understanding of ecological preservation tend to be more willing to bear the additional costs required to improve environmental conditions [40]. Moreover, Ajzen’s Theory of Planned Behavior [41] lends a psychological perspective, positing that attitude, subjective norms, and perceived behavioral control collectively determine individuals’ inclination to pay for green products or services. Some studies further emphasize social norms and reciprocity as key mechanisms shaping WTP, suggesting that people become more inclined to contribute once they believe “others are also paying for environmental protection”, thereby reinforcing group-wide support [42,43].

As a variety of green products and services—such as renewable energy, waste sorting programs, and ecotourism offerings—have gained traction, researchers have increasingly focused on how different environmental contexts can modulate WTP [44,45]. For instance, in the context of renewable electricity, pricing schemes, power-supply reliability, and brand recognition can all significantly influence consumer preferences in actual decision-making situations [46]. Some empirical evidence suggests that even individuals with strong pro-environment attitudes may not necessarily make real-world purchases of green products if adequate subsidies or pricing incentives are absent [47,48]. This discrepancy between attitudes and actual spending behavior indicates that policy and market instruments are still necessary to bridge the gap between environmental concern and tangible action.

Overall, existing scholarship has generated rich insights on (i) institutional hurdles in government eco-governance, (ii) the ambivalent role of new media in shaping green awareness, and (iii) the socio-economic determinants of residents’ willingness to pay for environmental improvement. Yet, two important blind spots remain. First, the three strands of literature have largely evolved in parallel: studies of digital information rarely link their findings to concrete monetary outcomes, while WTP research seldom asks how modern communication technologies modify individuals’ cost–benefit calculus. Second, most empirical work is descriptive or correlational; very few papers identify a causal channel running from technological progress to household-level environmental payments, and none—to our knowledge—does so with nationally representative micro-data for China.

The present study stitches these strands together in two ways. Conceptually, we view digital media adoption as a catalyst that transmits risk information, reshapes causal beliefs, and activates pro-environment norms, thereby connecting the “awareness” literature to the “pay-for-quality” literature. Empirically, we link the media-use module and the contingent valuation question in the 2018 CGSS and use provincial computer density as an external instrument to obtain a causal estimate of how relying on internet or mobile news affects households’ monthly willingness to pay for one additional “blue-sky” day. After establishing a large and statistically robust baseline effect, we explore heterogeneity by income tier and housing tenure and trace the mechanism through shifts in risk perception and pro-environment norms. Finally, a set of extension tests—larger air-quality targets, alternative online content, and the intensive margin among respondents who already pay—confirms the breadth and policy relevance of the result. In doing so, this study unifies three previously separate strands of literature and offers a concrete lever for expanding citizen co-financing of urban air-quality programs.

3. Data and Methodology

3.1. Data Source

We used the 2018 CGSS data. Since 2003, CGSS has been surveying individuals in 125 counties (districts), 500 streets (townships, towns), 1000 neighborhood (village) committees and 10,000 families across the country. By regularly and systematically collecting data on all aspects of Chinese people and Chinese society, researchers can summarize the long-term trends of social change and explore social issues of major theoretical and practical significance. CGSS promotes the openness and sharing of domestic social scientific research and provides data for international comparative research.

The CGSS, conducted across 28 provinces in both urban and rural areas, consists of various modules, including Core, Review of Ten Years and Energy, the East Asian Social Survey (EASS), and the International Social Survey Programme (ISSP). Each module uses different sample sizes to systematically study the changes in social structures and quality of life. For this paper, we focused on the core module and the energy module, which capture individual characteristics and residents’ WTP for environmental protection.

Although a newer CGSS wave from 2021 is now available and has been used to analyze social trust or broad sustainability perceptions, we deliberately retained the 2018 survey for three substantive and methodological reasons. First, 2018 is the most recent CGSS wave that includes a fully fledged contingent valuation question on residents’ monthly willingness to pay for additional “good-air” days; the 2021 questionnaire no longer contains this item, so the core dependent variable required for our study can only be sourced from 2018. Second, by 2018, national internet penetration reached 59.6% [49], and smartphone coverage had already reached a plateau in both urban and rural areas, ensuring that the “internet/mobile push” category in our media module captured a mature stage of technological diffusion rather than the early-adopter phase observed in pre-2013 data. Third, using the pre-pandemic wave avoids confounding effects arising from COVID-19 lockdowns and the unprecedented surge in home-bound screen time, which could blur the normal relationship between media choice and environmental payments. Taken together, the 2018 wave is the only CGSS round that simultaneously (i) measures the outcome of interest, (ii) observes the technology under conditions of near-universal availability, and (iii) precedes major exogenous shocks that might distort behavior. These features make the 2018 data uniquely suited—and still fully relevant—for identifying the causal link between digital information access and residents’ willingness to finance local air-quality improvements.

3.2. Model Setting

In this paper, we used an ordinary least squares (OLS) model to examine the impact of technological progress in information media on residents’ WTP for environmental protection. The specific model setup is as follows:

$${WTP}_i=\alpha +{\beta }_1{\mathrm{TP}}_i+{\beta }_2{X}_i+{\lambda }_i+{ϵ}_i$$

(1)

where ${WTP}_i$ represents the WTP for environmental protection of the i-th household. Specifically, we use the residents’ stated willingness to pay (in yuan per month) for a government program that would raise the number of “good-air-quality” days in every month of 2018 by one through measures such as factory shutdowns. ${\mathrm{TP}}_{\mathrm{i}}$ represents whether a respondent’s primary source of news is either the internet or customized mobile phone messages. The questionnaire lists six mutually exclusive media categories—newspapers, magazines, radio, television, internet, and mobile push services. Because reliance on internet-based channels reflects the diffusion of modern information technology, we treat the internet/mobile category as the operational measure of technological progress. ${\mathrm{X}}_{\mathrm{i}}$ is a vector of control variables, including other factors that may influence WTP for environmental protection, such as number of family members (NFMs), whether they rent their home, family economic status, and floor space. ${\mathsf{\lambda }}_{\mathrm{i}}$ represents the province fixed effects, controlling for heterogeneity across different provinces. ${\mathsf{ϵ}}_{\mathrm{i}}$ is the error term, accounting for random factors that are not explained by the model. Through this model, we aim to quantify the specific impact of technological progress in information media on residents’ WTP for environmental protection, while controlling for other potential confounding factors.

3.3. Variable Selection and Descriptive Statistics

In this paper, we used the households’ stated willingness to pay (in yuan per month) for a government program that would raise the number of “good-air-quality” days in every month of 2018 by one through measures such as factory shutdowns as a proxy for residents’ WTP for environmental protection. Respondents report a non-negative amount in yuan. We label this variable Monthly Clean-Air WTP. The question is narrow enough to yield a concrete valuation—avoiding the hypothetical bias often associated with vague environmental benefits—yet broad enough to reflect a public-goods context in which individual contributions are both plausible and policy-relevant. Because the payment scenario specifies a marginal improvement (one extra “good-air” day) and a recurring monthly fee, the responses can be interpreted as a marginal willingness to pay, which aligns with standard practice in environmental economics research. Figure 2 plots a kernel-density estimate of the variable Monthly Clean-Air WTP.

The distribution is sharply right-skewed: the density peaks are just above zero, indicating that most respondents are prepared to contribute only a very small amount for one additional “good-air” day. The mass then falls away quickly, and although responses extend to roughly CNY 1000, values beyond CNY 50 are exceedingly rare. The long, flat upper tail reflects a handful of households that attach unusually high monetary value to cleaner air, but these cases are exceptional.

We used whether a respondent’s primary source of news was either the internet or customized mobile phone messages as the independent variable. The CGSS asks respondents to identify the single most important channel through which they obtain news, offering six mutually exclusive categories: newspapers, magazines, radio, television, internet, and mobile push notifications. We recoded this information into a dichotomous variable, technological progress in information media, which equals 1 if the respondent selects internet or mobile push and 0 otherwise. We treated reliance on online channels as an observable manifestation of technological progress in information dissemination for three reasons.

• Diffusion of ICT: internet and mobile push services are direct outcomes of advancements in information and communication technology (ICT), differentiating them from legacy media, which rely on one-way broadcast or print distribution.
• Speed and interactivity: digital platforms deliver real-time air-quality indices, pollution alerts, and user-generated content, potentially heightening risk perception and pro-environment engagement.
• Empirical contrast: collapsing the six media categories into a digital versus traditional split maximizes statistical power while preserving a theoretically meaningful contrast between technologically advanced and conventional information sources.

By linking Monthly Clean-Air WTP to technological progress in information media, the analysis isolates whether—and to what extent—the adoption of modern information technology translates into higher private support for collective environmental action.

Figure 3 depicts the share of respondents who name each medium as their single most important source of news. Television dominates the landscape: just over half of all adults rely primarily on TV broadcasts. Online channels constitute the second-largest block—about 45 percent obtain most of their information from the internet, while another 1 percent use mobile push alerts. Together, these digital outlets account for roughly one resident in two, highlighting the rapid diffusion of new media by 2018. In contrast, traditional print and audio platforms play only a residual role: newspapers and radio are each chosen by little more than 2 percent of the sample, and magazines are virtually absent.

Although our key explanatory variable—whether the respondent’s primary news source is the internet or mobile push messages—is recorded at the individual level, we deliberately retain a household-level control set for two methodological reasons that are consistent with contingent valuation practice.

The CGSS question explicitly asks how much the respondent’s household would pay each month for an extra clean-air day. Previous work shows that such stated amounts are normally determined in light of joint resources and shared living conditions rather than purely personal tastes [43]. Household attributes—family size, rental status, family economic class, floor space (lnarea)—therefore capture the budgetary and incentive constraints that are most relevant for the payment decision.

To isolate the effect of information technology on willingness to pay (WTP), we followed earlier contingent valuation work and condition on household-level attributes that theory and prior evidence link to environmental payments.

Number of family members. Larger households face tighter per-capita budget constraints but also enjoy economies of scale in shared expenses; empirical studies therefore report ambiguous signs for the WTP coefficient and stress the need to control for family size when the payment unit is “per household” rather than “per person” [50].

Rent. Tenants cannot directly capitalize environmental improvements into property values, yet they are often more exposed to neighborhood-level externalities and less able to install private mitigation equipment. Several Chinese surveys found that renters display higher stated WTP for municipal pollution-control fees than homeowners once income is held constant [24]. Controlling for rent therefore adjusts for tenure-related incentives that might otherwise confound the media-use coefficient.

Family economic class. Disposable income remains the single most robust predictor of environmental payments across CVM studies in developing countries [24]. Rather than rely on raw earnings—which are noisy in self-reports—we adopted the CGSS class ranking, which captures permanent income and wealth perceptions better than annual cash flow.

Ln (floor space). Dwelling size serves as a proxy for household wealth and living-standard aspirations [51]. Bigger homes may imply higher capacity to pay but also greater indoor mitigation possibilities; controlling for floor area ensures that the media coefficient is not conflated with wealth or substitution effects.

Together, these controls align with the contingent valuation literature’s recommendation to account for budget capacity, exposure/risk heterogeneity, and tenure-specific incentives when modelling household-level environmental payments, thereby lending econometric credibility to our baseline estimate. The detailed descriptive statistics are provided in

Table 1. Statistical description of variables.

Variable	Definition	N	Mean	Std	Min	Max
Monthly Clean-Air WTP	Willingness to pay (in yuan per month) for a government program that would raise the number of “good-air-quality” days each month	10,384	4.409	26.836	0	999
Technological progress	1 = Internet or mobile push 0 = Otherwise	10,384	0.462	0.499	0	1
NFM	number of family members	10,384	2.807	1.438	1	21
Rent	0 = Do not rent 1 = Rent	10,384	0.129	0.335	0	1
Ln (floor space)	m2	10,384	4.507	0.634	1.609	7.601
Family economic class	1 = Far below average 2 = Below average 3 = Average 4 = Above average 5 = Far above average	10,384	2.588	0.723	1	5

.

4. Empirical Results and Analysis

4.1. Benchmark Model Results

Table 2. The benchmark model results.

Variables	Monthly Clean-Air WTP
	(1)	(2)	(3)
Technological progress	2.708 ***	2.181 ***	1.911 ***
	(0.508)	(0.551)	(0.561)
NFM		0.223	0.229
		(0.192)	(0.195)
Rent		1.382	1.260
		(0.845)	(0.857)
Ln (floor space)		−0.355	0.0146
		(0.454)	(0.497)
Family economic class		1.204 ***	1.059 ***
		(0.370)	(0.377)
_cons	3.060 ***	1.083	1.106
	(0.345)	(2.170)	(2.380)
Prov Fixed	No	No	Yes
N	10,825	10,384	10,384
R2	0.003	0.004	0.008
adj. R2	0.003	0.003	0.005

Note: The standard error is reported in parentheses; *** denote significance levels of 1%.

reports the baseline estimates. The dependent variable, Monthly Clean-Air WTP, measures the amount (in yuan) a respondent is willing to pay each month if the government were to secure one additional “good-air” day through measures such as factory shutdowns. The key explanatory variable, technological progress in information media, equals 1 when the respondent’s main news outlet is the internet or mobile push notifications and 0 when it is a legacy medium (newspapers, magazines, radio, or television).

Column (1) includes only the core regressor. Column (2) adds household-level controls—NFM, rent, ln (floor space), and family economic class. Column (3) further introduces province fixed effects to net out regional heterogeneity. Across all specifications, the coefficient on technological progress remains positive and highly significant, indicating a robust association between technological progress in information media and a higher willingness to pay for cleaner air.

In Column (1), the coefficient on technological progress is 2.708 (p < 0.01), implying that, ceteris paribus, respondents who rely on internet-based sources are willing to pay roughly CNY 2.71 more per month than their counterparts who rely on traditional media. After adding covariates and province dummies, the effect attenuates but remains sizable at CNY 1.911 (p < 0.01). Notably, family economic class is also positive and significant in Columns (2) and (3), suggesting that households in higher housing tiers possess both greater capacity and stronger inclination to finance environmental improvement.

For context, the sample mean of Monthly Clean-Air WTP is CNY 4.41. Consequently, the estimated digital media premium of CNY 1.9–2.7 represents a 43–61 percent increase over the average—an economically meaningful magnitude by the standards of the WTP literature.

Taken together, the baseline results corroborate the hypothesis that technological progress in information dissemination—proxied by primary reliance on internet or mobile news—raises residents’ willingness to pay for government air-quality initiatives. These findings imply that, in an information-rich society, amplifying environmental messaging through digital channels could materially enhance public support and financing for state-led pollution-control programs.

4.2. Robustness Check

4.2.1. Instrumental Variable (IV)

A key empirical concern is that the propensity to rely on digital media (technological progress) may be endogenously correlated with unobserved factors that also influence residents’ willingness to pay for cleaner air. Individuals who are more environmentally conscious, better informed, or technologically savvy could both (i) prefer internet-based news and (ii) express a higher willingness to finance public environmental programs. In addition, reverse causality cannot be ruled out: households that already care more about air quality might actively seek real-time pollution information on the internet or via mobile push services. These channels would bias the OLS estimates upward.

To disentangle causal effects, we instrumented technological progress with the number of personal computers per one hundred residents at the provincial level (computer/100 persons). This variable satisfies two requirements:

• Relevance. Provinces with greater computer penetration provide easier access to internet infrastructure and digital services, making it more likely that residents designate online sources as their primary medium. The first-stage results (Column 2 in

  Table 3. The results of instrumental variable model.

  Variables	Monthly Clean-Air WTP	First Stage
  	(1)	(2)
  Technological progress	10.525 *
  	(6.241)
  Computer/100 person		0.003 ***
  		(0.000)
  Control variables	Yes	Yes
  Cragg–Donald Wald F statistic	87.761
  Kleibergen–Paap rk Wald F statistic	87.187
  liml	Yes
  N	10,384
  R2	−0.018
  adj. R2	−0.019

  Note: The robust standard error is reported in parentheses; * and *** denote significance levels of 10% and 1%, respectively.

  ) confirm this intuition: the instrument’s coefficient is 0.003177 (p < 0.01), while the Cragg–Donald (87.761) and Kleibergen–Paap (87.187) F-statistics far exceed the conventional threshold of 10, eliminating concerns about weak identification.
• Exogeneity. Conditional on household covariates and province fixed effects, provincial computer density should affect an individual’s willingness to pay for clean air only through its impact on digital media adoption. Computer ownership is largely driven by historical infrastructure rollout and regional economic development—factors already captured by the control set—rather than by any direct preference for paying environmental levies.

The second stage estimate underscores the economic significance of technological progress (Column 1 in Table 3): the coefficient on technological progress rises to 10.52 (p < 0.10), markedly larger than the baseline OLS range of CNY 1.9–2.7. Economically, a household that switches from traditional to internet or mobile news would be willing to contribute roughly CNY 10.5 more per month for one additional “good-air” day—an order-of-magnitude increase that highlights the potent role of information technology in mobilizing private funding for environmental improvement.

4.2.2. Alternative Measures of Technological Progress

To verify that our main result was not an artifact of how technological exposure is coded, we re-specified the key regressor along two theoretically motivated dimensions. A binary indicator may mask meaningful variation among digital users. The CGSS provides a five-point frequency scale (“never” to “daily”) for internet access. Using this continuous measure allowed us to test whether greater immersion in online content, rather than a simple yes/no distinction, drives willingness to pay (WTP). In addition, information channels differ not only in digital versus traditional format but also in their technological sophistication. We therefore recoded the respondents’ primary medium into an ordinal six-level variable—newspapers (1), magazines (2), radio (3), television (4), internet (5), and mobile push notifications (6). This ranking captures a monotone progression from low-tech broadcast media to highly interactive, real-time digital platforms. If our argument hinges on “technological progress”, the estimated effect should increase with each step up the scale.

Table 4. The results of the ordered probit model.

Variables	Monthly Clean-Air WTP
	(1)	(2)	(3)	(4)
Internet usage intensity	0.912 ***	0.622 ***
	(0.184)	(0.210)
Technological gradient			1.481 ***	1.090 ***
			(0.358)	(0.380)
Control variables	No	Yes	No	Yes
Prov fixed	No	Yes	No	Yes
N	11,133	10,665	10,825	10,384
R2	0.002	0.007	0.002	0.008
adj. R2	0.002	0.004	0.001	0.005

Note: The standard error is reported in parentheses; *** denotes significance levels of 1%.

shows the robustness test results of measuring “Internet usage Intensity” or “Technological Gradient” in different ways to verify the reliability of the core conclusions.

When internet usage intensity replaces the binary indicator, its coefficient remains positive and statistically significant (0.912 without and 0.622 with province fixed effects). Hence, residents who log on more frequently are systematically more willing to finance air-quality improvements, corroborating the intensity channel.

Using the six-level Technological Gradient, the coefficients are likewise positive and significant at the 1 percent level (1.481 and 1.090, respectively). Movement toward more advanced media types—especially internet and mobile push—translates into a higher WTP, consistent with the notion that greater technological sophistication enhances environmental engagement.

4.3. Mechanism Analysis

Reliance on internet-based news or mobile push services exposes respondents to a denser stream of pollution maps, expert commentary, user-generated photos, and normative appeals compared to traditional media, and this constant, interactive information flow shapes three intertwined cognitive layers that the risk-perception and value–belief–norm frameworks identify as prerequisites for pro-environmental behavior: first, by repeatedly highlighting smog episodes and real-time AQI alerts, it dampens optimism bias and cultivates a more critical appraisal of local air quality, so those who obtain news online are less likely to agree that “air quality in my area was good in 2017” (Statement 1); second, articles, infographics, and short videos that trace greenhouse gases (Statement 2) and acid rain (Statement 3) back to coal-dominated energy systems strengthen causal attribution, making internet and mobile users more inclined to endorse the statements that energy use is the main cause of the greenhouse effect and acid rain; third, digital platforms circulate stories of crowd-funded lawsuits, low-carbon lifestyle challenges, and peers’ willingness to pay for green products (Statement 4), thereby normalizing personal economic sacrifice for collective environmental benefits and increasing agreement that people should be willing to give up some economic gains to protect the environment. Taken together, these sequential shifts in risk appraisal, causal belief, and normative commitment constitute the micro-level psychological mechanism through which our operationalized measure of technological progress—using internet or mobile channels as the primary news source—ultimately raises residents’ willingness to pay for better air quality.

Respondents rate their agreement on a five-point scale (1 = “strongly disagree”, 5 = “strongly agree”) with the following statements:

• “Air quality in my area was good in 2017.”
• “Energy use is the main cause of the greenhouse effect.”
• “Energy use is the main cause of acid rain.”
• “People should be willing to sacrifice some economic benefits to protect the environment.”

Higher scores indicate stronger endorsement. The pathway described above matters precisely to the extent that heightened risk appraisal, clearer causal attribution, and stronger pro-environmental norms—each fostered by reliance on internet or mobile news—jointly translate into a greater monetary commitment; in other words, digital information technology shapes the way residents perceive local pollution, understand its energy-related causes, and internalize the social expectation of bearing economic costs. These cognitive shifts converge to elevate their stated willingness to pay for cleaner air, as captured by WTP.

Table 5. The results of the mechanism analysis.

Variables	Statement 1	Statement 2	Statement 3	Statement 4
	(1)	(2)	(3)	(4)
Technological Progress	−0.241 ***	0.091 ***	0.103 ***	0.067 **
	(0.035)	(0.027)	(0.028)	(0.028)
Control Variable	Yes	Yes	Yes	Yes
Prov Fixed	Yes	Yes	Yes	Yes
N	3845	3545	3402	3747
R2	0.185	0.037	0.033	0.043
adj. R2	0.178	0.028	0.024	0.035

Note: The standard error is reported in parentheses; ** and *** denote significance levels of 5%, and 1%, respectively.

reports OLS estimates with household covariates and province fixed effects. Consistent with the expectations, primary reliance on technological progress is associated with a significantly lower assessment of past air quality (coefficient = –0.241, p < 0.01), while its effects on the two causal-attribution items and on the sacrifice norm are positive and statistically significant (0.091, 0.103, and 0.067, respectively; p < 0.05). These magnitudes are robust to alternative specifications and suggest a coherent narrative: digital platforms expose users to real-time pollution data and critical commentary, making them more skeptical about the state of their local environment; simultaneously, the same channels disseminate scientific explanations of climate and acid-rain mechanisms and circulate discourses that legitimize bearing private costs for public environmental goods. The resulting combination of heightened risk perception, clearer causal understanding, and strengthened moral conviction provides a plausible pathway through which technological progress in information access translates into a higher monetary commitment to air-quality improvement.

4.4. Heterogeneity Analysis

The effect of digital information on environmental willingness to pay is unlikely to be uniform across the population. We focus on household economic class and housing tenure because these two characteristics directly affect both a family’s exposure to online information and its practical ability to pay for environmental improvements. Higher-status households typically have more resources to act on new information, while renters and owners face different cost constraints and incentives when supporting collective air-quality measures.

We therefore stratified the sample along two dimensions that theory suggests should mediate both information uptake and the capacity to act on it.

Households in higher economic class typically enjoy higher incomes and better living conditions. They possess a larger discretionary budget and may place a higher marginal value on incremental improvements in environmental quality. If the internet primarily raises awareness, wealthier households should be better positioned to translate that awareness into actual payments.

Housing tenure (rent). Renters and owners face different exposure horizons and financial constraints. Renters have fewer sunk costs in the dwelling, are more mobile, and often lack long-term control over indoor mitigation measures. They may therefore rely more heavily on collective interventions—such as government air-quality programs—and, once informed, be willing to pay more for them. By contrast, owners, tied down by mortgage obligations and property investments, might weigh additional spending more cautiously. These hypotheses motivate the subgroup regressions reported in Figure 4.

Columns (1) and (2) (Figure 4) contrast households with high (family economic class ≥ 3) and low (family economic class < 3) family economic class. Digital media usage (technological progress) remains positive and significant in both groups, but the magnitude is larger for affluent families (β = 2.007, p < 0.01) than for less-affluent ones (β = 1.571, p < 0.10). Thus, wealthier households convert the informational advantage of the internet into a stronger readiness to finance air-quality improvements—consistent with greater ability to pay and higher environmental demand.

Columns (3) and (4) (Figure 4) split the sample by tenure. Among renters, the coefficient on technological progress reaches 4.392 (p < 0.05), far exceeding that for homeowners (1.554, p < 0.01). Because renters exercise limited control over structural mitigation and are less encumbered by mortgage payments, they appear more willing to exchange disposable income for government-provided environmental amenities once informed of local pollution risks.

All specifications include province fixed effects to net out unobserved regional heterogeneity. The pattern confirms that technological progress in information media raises environmental willingness to pay across the board, but the size of the effect varies systematically with household resources and tenure. Digital outreach is therefore especially potent among economically advantaged families and, even more so, among renter households—two constituencies that can be strategically targeted in future environmental-finance campaigns to broaden the fiscal base of air-quality policy.

4.5. Further Study

To probe the scope and boundaries of our core finding, we conducted three sets of additional tests.

First, we asked whether the effect of digital media becomes stronger when the environmental benefit is larger.

Second, we examined whether what residents read on digital platforms—specifically, domestic versus international news—matters for their willingness to pay (WTP).

Third, we restricted the sample to individuals who already exhibit a positive WTP in order to assess whether information technology primarily affects the extensive margin (whether to pay) or the intensive margin (how much to pay). The detailed results of the further study are presented in

Table 6. The results of further analysis.

Variables	Three Good-Air Days	Five Good-Air Days	Monthly Clean-Air WTP		Intensive Margin
	(1)	(2)	(3)	(4)	(5)
Technological progress	3.819 ***	5.602 ***			6.987 **
	(0.729)	(1.152)			(3.127)
Online portals			−5.727 ***
			(0.944)
Social media				−5.640 ***
				(1.184)
Control Variable	Yes	Yes	Yes	Yes	Yes
Prov Fixed	Yes	Yes	Yes	Yes	Yes
N	10,371	10,367	10,467	10,467	1697
R2	0.013	0.010	0.009	0.008	0.041
adj. R2	0.010	0.007	0.006	0.005	0.023

Note: The standard error is reported in parentheses; ** and *** denote significance levels of 5%, and 1%, respectively.

.

Columns (1) and (2) of Table 6 replace the baseline dependent variable (one extra “good-air” day) with scenarios promising three and five additional good-air days per month.

The coefficients on technological progress rise to 3.819 and 5.602, respectively (both p < 0.01). We considered these outcomes for two reasons. First, marginal valuation theory predicts that WTP should increase with the size of the benefit; confirming that pattern guards against the possibility that our main result is tied to an arbitrarily small improvement. Second, policy makers often contemplate packages that yield larger, discrete air-quality gains; demonstrating that digital media foster even stronger support for such packages enhances the practical relevance of our study. The sizable jump in coefficients suggests that internet and mobile channels amplify residents’ enthusiasm for more ambitious clean-air targets.

Columns (3) and (4) introduce two new indicators: reading international news primarily through online portals and through social media such as WeChat or Weibo. Both coefficients are negative (−5.727 and −5.640, p < 0.01). We analyzed these variables because digital platforms are heterogeneous; some users focus on local environmental updates, whereas others concentrate on global events. If attention is diverted to foreign affairs or climate stories abroad, the perceived urgency of contributing to local air-quality programs may wane. The negative signs do not contradict the overall positive role of digital media; rather, they reveal that content orientation moderates its effect. Future outreach efforts should therefore localize environmental messaging to avoid diluting citizens’ sense of responsibility for nearby pollution.

Column (5) restricts the sample to respondents whose baseline WTP is strictly positive. For this group, the coefficient on technological progress jumps to 6.987 (p < 0.05)—far above the baseline range of CNY 1.9–2.7. We examined this subset because policy adoption requires not only attracting new contributors (extensive margin) but also encouraging existing supporters to pay more (intensive margin). The result indicates that once individuals accept the legitimacy of environmental fees, digital information further raises the amount they are willing to contribute, presumably by reinforcing peer norms and highlighting the marginal effectiveness of additional funds.

Taken together, the extended analyses show that (i) technological progress in digital media exert larger effects when the promised environmental gain is substantial or when the individual already intends to pay; (ii) attention to international news can dampen local environmental contributions, underscoring the importance of message targeting; and (iii) the type and direction of online content are as crucial as the mere presence of information technology.

For practitioners, these findings imply that digital campaigns should articulate concrete, sizable air-quality targets and foreground local relevance to maximize public financing. Among citizens who are already environmentally minded, personalized online engagement can unlock even greater fiscal support, helping government and civil-society actors close funding gaps and strengthen collaborative air-quality governance.

5. Conclusions and Policy Recommendations

5.1. Conclusions

Drawing on nationally representative CGSS-2018 data, this study investigated whether—and through which channels—the technological progress in information media shapes Chinese residents’ willingness to pay (WTP) for government air-quality improvements. Ordinary least squares estimates showed that households whose primary news source is the internet or mobile push services are willing to contribute roughly 45% more each month than households that rely on legacy media. Instrumenting technological progress in information media use with provincial computer penetration confirmed that this relationship is causal. The mechanism analysis revealed that technological progress lowers complacency about local air quality, sharpens understanding of anthropogenic pollution, and strengthens the normative acceptance of economic sacrifice for environmental goals. Heterogeneity analyses demonstrated that the effect is strongest among households from high economic class families and renters, while further study indicated even larger elasticities when the promised environmental gain is substantial or when individuals are already predisposed to pay. A notable caveat is that residents who primarily consume international news online display lower local WTP, underscoring the importance of content orientation.

While our empirical evidence was drawn from the 2018 Chinese General Social Survey, the underlying mechanism—digital access reshaping risk perception, causal beliefs, and moral norms—rests on psychological processes that are not unique to China. The results are therefore most transferable to settings that meet three conditions: (i) widespread internet and smartphone penetration, such that online news competes directly with legacy media; (ii) measurable local air-pollution problems for which government remediation is plausible; and (iii) a fiscal framework in which small household contributions can be channeled into public clean-air programs. Many rapidly industrializing economies in Asia (e.g., India, Vietnam) or Latin America (e.g., Mexico, Brazil) satisfy these criteria, suggesting that targeted digital outreach could likewise unlock additional citizen funding there. By contrast, in regions where pollution is already low or where online platforms are tightly restricted, the marginal informational gain—and thus the effect on willingness to pay—may be smaller. Future cross-country work that replicates our survey design in such varying contexts would help delineate the precise boundaries of the digital media effect documented for China.

5.2. Policy Recommendations

Firstly, government environmental authorities, city bureaus of ecology, and NGO partners should move beyond sporadic website updates and adopt a multi-platform strategy centered on mobile apps, short-video portals, and social-media mini-programs. Push notifications can deliver daily Air Quality Index (AQI) readings, but richer content—interactive maps, personalized exposure dashboards, and infographics that translate AQI scores into health tips—will further raise salience. Pilot evidence from our study suggests that digital users respond especially well to quantified pledges; therefore, publicity campaigns should frame fundraising goals in concrete units rather than vague environmental slogans. Partnerships with dominant “super-apps” (e.g., WeChat, Alipay) can integrate these messages into users’ routine payment interfaces, lowering transaction costs and expanding reach beyond traditional green constituencies.

Secondly, the negative coefficient on international news consumption highlights an “attention displacement” risk: global climate headlines may dilute the perceived urgency of local pollution. To counter this, environmental portals should automatically geolocate users and foreground city-specific PM_2.5 trends, local emissions inventories, and even neighborhood-level micro-sensor data where available. Embedding health calculators—e.g., expected reduction in hospital visits or asthma attacks if PM_2.5 falls by 10 µg/m³—can translate abstract numbers into tangible personal stakes. Municipal authorities might also push short, shareable videos featuring local physicians, school principals, or taxi-driver testimonials to connect clean-air benefits with everyday experience.

Thirdly, behavioral science suggests that contributors need timely feedback to avoid “donor fatigue”. Clean-air funds could post weekly progress meters and automatically notify donors when a new “blue-sky-day” threshold is crossed. Augmented-reality filters—turning skyline images progressively clearer as funding milestones are reached—offer a low-cost way to visualize collective impact. Municipalities might further experiment with conditional-commitment schemes: residents pledge a monthly amount that is charged only if official monitoring confirms the promised number of clean-air days, aligning incentives and credibility.

Lastly, our first-stage estimates reveal that digital engagement hinges on provincial computer and broadband penetration. National and provincial development plans could include joint projects that roll out 5G base stations, public Wi-Fi in township squares, and environmental apps. As connectivity improves, local governments can gradually transition from bulletin-board notices to mobile-first environmental governance, broadening both information diffusion and the fiscal base for air-quality interventions.

6. Limitations and Future Directions

Several caveats temper the generality of our findings. First, the analysis relies on a single cross-section—CGSS 2018—because it is the most recent wave that includes a contingent valuation item on “good-air-day” payments. The six-year gap to our publication date means that the media landscape has evolved (e.g., short-video apps, 5 G streaming) and new environmental levies have been piloted. Replicating this study with a purpose-built 2025 survey or a future CGSS wave that re-introduces a WTP module would reveal whether the digital effect strengthens or plateaus in an era of near-universal smartphone use. Second, WTP is self-reported and hypothetical; observed payment behavior—such as voluntary green-electricity purchases or donations to smog-control funds—would provide a harder test of the mechanism. Third, although we control for key household factors and use an external instrument, unobserved preferences (e.g., underlying pro-social values) could still bias the estimates. Panel data or natural experiments that track media adoption before and after broadband roll-outs would tighten identification.