Modelling Urban Expansion, Energy Consumption, and Environmental Sustainability: The Moderating Role of Environmental Taxes in Developing Countries

Abstract

Rapid expansion in urbanisation, along with the rising demand for energy consumption, has deepened environmental apprehensions among developing economies and intensified their concerns about long-run environmental sustainability. This article examines how urban expansion and rising energy consumption impact environmental sustainability, and whether environmental taxes moderate this relationship, by using a panel of 110 developing countries over the period of 2010 to 2024. To capture both static and dynamic relationships among the variables, we have applied complementary econometric methodologies that allow for cross-country heterogeneity and persistence in emissions. The estimated outcomes show that urban expansion and energy consumption are significantly increasing gas emissions, and this outcome is consistent with the idea that environmental costs of urban-led growth and energy-intensive development. But as we have added environmental taxes as a moderating policy instrument, the positive impact of energy consumption and urbanisation on emissions becomes negative in most specifications. The significant impact of both interaction terms, i.e., environmental taxes and urbanisation, and environmental taxes and energy consumption, across different estimation strategies, suggests that environmental taxation weakens emissions and encourages structural change with rising energy use. Renewable energy consumption and foreign direct investment have significant influences on emissions, emphasising the role of energy structure and investment composition in shaping environmental outcomes, whereas the income effect varies across models. The outcomes of dynamic models also confirm emissions persistence, but over time, environmental taxes reduce the degree of emissions persistence. The estimated outcomes imply that environmental taxes can support a decoupling of urbanisation and energy-driven growth from environmental degradation. Thus, developing countries should balance urban development, energy demand, and environmental sustainability through credible market-based regulations.

1. Introduction

Among the issues of the contemporary world, environmental sustainability has received serious concerns, especially when climate change threatens the Earth’s ecological balance and impacts human well-being through extreme weather conditions, productivity deteriorations, and health issues [1]. The current trends of greenhouse gas emissions are showing that global warming will not remain less than 2 degrees Celsius in the coming decades, urging world leaders to take swift and large-scale action to control irreversible damage to the ecological system [2]. To these concerns, the global development agenda working under the umbrella of UN Sustainable Development Goals monitors big cities that are inclusive, safe, resilient, and sustainable, but also have deep-rooted effects on climate change and ecological imbalance [3,4,5]. For developing nations, achieving these goals is especially challenging, as they need higher economic growth to meet the demands of a rising urban population with limited institutional and financial resources [6,7,8,9,10]. Environmental sustainability, being a multidimensional factor, presents the capacity of ecological systems to absorb anthropogenic pressures without impacting the structure and resilience over time. Environmental sustainability is inherently multidimensional, encompassing ecological footprints, biodiversity, and broader environmental performance indicators. This study operationalises environmental sustainability using greenhouse gas emissions as an inverse proxy. This approach is consistent with a large body of the empirical literature [11,12,13,14], where emissions are used as a primary indicator of environmental pressure, particularly in cross-country analyses where data availability and comparability are critical.

Urban expansion constitutes a central structural driver of environmental pressure in developing economies [15]. Concentrated population in cities, economic activities, and investment in infrastructure raise energy consumption with environmental externalities through construction, transportation, industrial agglomeration, and consumption-intensive lifestyles. Empirics show that among East Asian economies, urban expansion, with trade openness and economic growth, increases CO₂ emissions, supporting the environmental costs of urban-led development strategies [15,16]. In the absence of green urban planning and ineffective regulations, expanding urban footprints are associated with higher fossil fuel consumption, construction-related emissions, traffic congestion, and greater pollution intensity [17,18,19]. These challenges are more complicated in developing countries, where urban expansion is spatially unplanned and there are ill-functioning governing bodies, no formal settlement expansion, infrastructure gaps, and weaker energy efficiency. This raises a key policy-relevant question: Does rapid urban expansion inevitably intensify emissions and environmental degradation?

Energy consumption, especially the greater use of non-renewable energy resources with a rising population, is strongly integrated with more pollutants and environmental degradation, particularly in developing countries. The empirical literature on disaggregate energy resources [20,21] shows that non-renewable energy growth patterns press the economies toward environmentally unsustainable paths, until true efforts are made to improve the efficiency, technological upgradation, and overall structural transformation. It is the pressure of a rising population that makes most developing countries heavily rely on non-renewable fossil fuels, with lesser investment in low-carbon technologies due to their limited financial, infrastructural, and policy capacity [22,23]. Thus, rising consumption of non-renewable energy is intensified by urban expansion and results in higher emissions with weakening environmental sustainability. The composition of energy consumption and diversity of investment inflow play central roles in mediating the urbanisation energy environmental linkage, as both shape the technology adoption, composition of production, and emission-led growth. On one hand, foreign direct investment promotes updated technologies and advanced managerial practices with capital deepening improving efficient energy uses [23,24]; but if weak environmental regulations prevail, it can expend pollution-intensive production [25,26]. On the other hand, renewable energy consumption is continuous, leading to lower emissions and improving the quality of the environment, particularly as it replaces high-carbon energy sources on a large scale [26].

From a theoretical point of view, the relationship among urbanisation, energy consumption, economic growth, and environmental quality is often framed by using the Environmental Kuznets Curve hypothesis (EKC). This hypothesis explains that an inverted U-shaped relationship exists between environmental degradation and level of income; e.g., pollution rises in the early stages of development, but after reaching a threshold level of income, it starts declining, which may be due to the adoption of cleaner technologies and implementation of proper environmental regulation [17,27]. Although some empirical studies support this pattern [28,29,30], others are against this ideology [31,32]. But the improvements in environmental outcomes at later stages are not automatic, as they depend on the quality of institutions, regulatory credibility, level of technology adoption, and policy instruments that shift relative prices and incentives [33,34]. Following this ideology, environmental tax is the most prominent policy instrument used to fix externalities attached to the environment and encourage environmental sustainability. The main aim of this tax is to internalise the social costs of pollution by raising the price level to those activities which are harmful to the environment, incentivising cleaner production, energy efficiency, and substitution toward low-carbon energy resources [17]. Although different studies have shown how environmental taxes work to improve the quality of environmental sustainability [35,36,37], this study aims to examine the impact of urban expansion and energy consumption on environmental sustainability, with a particular focus on the moderating role of environmental taxes in developing countries. Specifically, this study investigates whether environmental taxation can alter the relationship between structural drivers of growth and emission outcomes, thereby contributing to sustainable development policy design. This type of study is hardly available in the existing literature.

2. Literature Review

Over the years, the literature on environmental sustainability in developing countries has expanded, showing increasing concern over the ecological consequences of growth, structural transformation, and urbanisation. The previous literature [5,6,7,15,18,21,38,39,40,41] uses greenhouse gas emissions, ecological footprint, or composite measures of environmental performance as indicators of environmental sustainability. Among these, greenhouse gas emissions are the most widely used proxy in cross-country panel studies due to their strong linkage with climate change, global comparability, and consistent data availability. Although this measure does not fully capture all dimensions of environmental sustainability, it provides a reliable and policy-relevant indicator of environmental pressure, particularly in the context of energy use and urban expansion. But recent studies are going beyond the traditional Environmental Kuznets Curve by including variables like urbanisation, trade openness, energy mix, policy instruments, and institutional quality in their modelling of environmental outcomes [36,37,42,43,44]. Among these indicators, urban expansion and energy consumption are considered the most core determinants of environmental outcomes in developing economies, whereas dynamic demography and industrialisation are still energy-intensive and carbon-intensive.

The environmental impact of urbanisation is not homogeneous and depends significantly on the spatial structure of urban growth. Planned, compact urban development characterised by efficient land use, high-density settlement, and integrated public transport systems tends to reduce per capita emissions by lowering transport demand and improving energy efficiency. In contrast, unplanned and low-density urban sprawl is associated with higher emissions due to increased reliance on private transport, inefficient infrastructure, and greater land-use change. Therefore, distinguishing between different forms of urban expansion is essential for accurately assessing its environmental consequences. The empirical literature on cross-country analysis shows that a rising population with dense urbanisation tends to raise emissions, as it raises energy demand, due to industrial agglomeration and mass consumption activities [45,46]. Recent studies on emerging economies confirm that urbanisation is linked to deterioration in the quality of the environment when it coincides with a poor transport system based on fossil fuel, ill and outdated industrial agglomeration, and weak land-use planning [47,48,49]. Kunwal et al. [50] mention that urban expansion and population density increase the level of CO₂ emissions, whereas renewable energy use mitigates these effects, emphasising the role of the energy structure in setting the environmental outcomes of urban expansion. Bashir et al. [51] find that urbanisation and energy use together raise long-run CO₂ emissions in Indonesia, while Shaheen et al. [52] show that energy consumption and level of income are the main long-run drivers of emissions in Pakistan. Oteng-Abayie et al. [53] find that resource-dependent economies that have higher energy use to attain the desired level of economic growth actually intensify environmental degradation.

Within the energy–environment literature, consumption of renewable energy has gotten special attention as a practical mechanism to maintain equilibrium between economic growth and environmental sustainability. Dogan et al. [54] find that renewable energy consumption decreased the environmental degradation in Denmark over 1990–2020, by controlling for financial development, economic growth, and the value added of agriculture. On the other hand, Ahmed et al. [55] and Marc & Ali [56] find that renewable energy reduces ecological footprints for the panel of seven economies when effective regulations and democratic institutions are supporting variables. Azimi & Rahman [57] and Hou & Yuan [58] find that the threshold level of environmental sustainability and renewable energy reduces greenhouse gas emissions in the presence of fiscal soundness and better institutional quality. Empirical evidence from the Organization for Economic Cooperation and Development countries shows that a long-run relationship exists between renewable energy consumption, urbanisation, economic growth, and CO₂ emissions [59]. However, some studies find that renewable consumption faces trade-offs with land use, intermittency, and system integration costs, particularly where governance and technology capacity are limited.

Several empirical studies investigate foreign direct investment, and economic growth is the main driver of environmental degradation, especially when we are testing the pollution haven and pollution halo hypotheses [60,61,62,63]. Furtuna and Atis [64] find that by using nonlinear modelling, carbon-intensive economies show a U-shaped relationship between foreign direct investment and environmental degradation, but they keep technology transfer and efficiency gains intact as independent variables. This reflects that integration of renewable energy and foreign direct investment within a unified framework of environmental sustainability is important to examine. The multi-country literature indicates that renewable energy is linked to improved environmental quality, whereas non-renewable energy has the opposite effect [65,66,67,68].

Recently, as a market-based instrument to control CO₂ emissions and improve environmental governance, environmental taxes have received much attention. Leitão [69] finds that environmental taxes, along with renewable energy consumption, reduce CO₂ emissions. Ali & Kirikkaleli [70] find that environmental taxes reduce consumption-based emissions, whereas negative shocks weaken this relationship in Italy. Gafsi & Bakari [71] find that environmental taxes increase renewable energy consumption in Organization for Economic Co-operation and Development countries, whereas technological changes and environmental governance play a mediator role. This outcome shows that environmental taxes may impact environmental sustainability both directly, by discouraging fossil fuel use, and indirectly, by raising renewable energy consumption and technology upgradation.

Few studies examine the interaction effects between environmental taxes, urbanisation, and energy consumption. However, the effectiveness of environmental taxes in reducing emissions is not uniform across countries and largely depends on institutional quality, including regulatory enforcement, governance effectiveness, and corruption control. In developing economies, weak administrative capacity and limited enforcement mechanisms may reduce the effectiveness of environmental taxes, as firms and households may not fully comply with environmental regulations. Therefore, institutional quality plays a complementary role in determining whether environmental taxation can effectively internalise environmental externalities and promote sustainable outcomes. Xu et al. [72] find that environmental tax reduces urban energy consumption, mainly by inducing industrial upgrading, greater openness, and technological innovation, with stronger effects in small and medium cities in China. Few panel studies are available that check the relationship between environmental taxes and CO₂ emissions [73,74]. Nevertheless, most of the existing literature is concentrated on single-country analysis or the panel of developed countries [75,76,77,78,79,80,81,82], and relatively very few studies explicitly model environmental taxes as a moderating variable that can change the strength and expected sign of the urbanisation, energy consumption, and environmental sustainability relationship in a large panel of developing countries. This gap is significant because environmental taxation may influence CO₂ emissions and has the potential to reshape environmental outcomes when interacting with urban expansion and rising energy demand, although empirical evidence on this relationship remains limited and inconclusive.

3. Theoretical Model

To provide a coherent theoretical foundation, this study integrates three complementary frameworks. The IPAT (environmental impact (I) as the product of population (P), affluence (A), and technology (T)) identity offers a structural perspective, where urban expansion reflects demographic and spatial concentration, and energy consumption captures technological and behavioural intensity. The Environmental Kuznets Curve hypothesis introduces a nonlinear development perspective, suggesting that the environmental impact of these factors evolves with income and structural transformation. Pigouvian taxation complements these frameworks by providing a policy mechanism that internalises environmental externalities and alters the behavioural responses embedded in the IPAT components [83,84,85,86,87]. Thus, while IPAT explains the structural drivers of emissions, the Environmental Kuznets Curve [88] provides the development trajectory, and Pigouvian taxation acts as the policy lever that modifies these relationships [89,90]. This integrated framework underpins both the direct and moderating effects specified in the empirical model. The IPAT identity shows that environmental impact has a multiplicative relationship with demographic scale, economic activities, and technological changes [90]. Following this background, in our model, urban expansion is considered as a spatial and structural manifestation of population and economic agglomeration, and energy consumption has been used as key technological and behavioural changes that impact the emissions through production and consumption. Urban expansion in this study is treated as an aggregate indicator; however, its environmental impact may vary depending on the nature of urban development. Compact and well-planned urban growth may improve energy efficiency and reduce emissions intensity, whereas dispersed and unplanned urban sprawl may increase environmental pressure. Thus, the estimated coefficient on urban expansion should be interpreted as capturing the average effect across different urban forms, which may conceal important heterogeneity. Foreign direct investment shows cross-border technology and capital flows that can impact greenhouse gas emissions intensity, and the renewable energy consumption proxies as a compositional shift in the energy technology set toward lower-carbon sources. The environmental taxation works as an institutional and price-based mechanism that changes the benefits of firms and households, by changing their behavioural responses to environmental pressure [17,91,92,93].

The Environmental Kuznets Curve hypothesis posits that environmental degradation increases in the early stages of development, as income, urbanisation, and energy use rise, but it starts declining after a threshold level as economies adopt cleaner technologies and more stringent environmental policies [17,94]. But the turning point of the U-shape is not a guarantee, as developing countries vary in policy mix, institutional and regulatory structure, and urban development [95]. Urban expansion through construction activities, transport intensity, higher energy demand, weak environmental regulations, and no or weak land-use regulations raise the level of greenhouse gases [96]. Energy consumption, especially generated through traditional fossil fuel and ill technological processes, raises greenhouse gas emissions [97]. The Pigouvian taxation [83] concept explains that environmental taxes directly and indirectly operate to mitigate the negative impacts of urban expansion and energy consumption by raising the individual and firm cost of pollution-led consumption and production. Thus, the baseline model can be written as

ES_it = α_i + β₁URB_it + β₂ENC_it + β₃FDI_it + β₄REN_it + γ Z_it + ε_it

where α_i captures unobserved country-specific effects, Z_it is a vector of additional controls (income, etc.), and ε_it is an idiosyncratic error term. In this equation, β₁ and β₂ represent the elasticities of environmental sustainability with respect to urban expansion and energy consumption. When environmental sustainability is measured inversely through emissions, theoretical priors and existing empirical evidence suggest β₁ < 0 and β₂ < 0, and higher urbanisation and higher energy consumption reduce environmental sustainability by increasing emissions [98].

In this study, environmental sustainability is proxied inversely by greenhouse gas emissions. This implies that higher emissions correspond to lower environmental sustainability. While this operationalisation focuses primarily on the climate-related dimension of sustainability, it is appropriate given the study’s emphasis on energy consumption, urban expansion, and environmental taxation, all of which are closely linked to emission dynamics.

The inclusion of foreign direct investment and renewable energy consumption reflects two further theoretical channels. On the one hand, foreign direct investment may increase environmental pressure if it brings pollution-intensive capital to countries with less regulation (pollution haven effect), implying β₃ < 0, when environmental sustainability decreases with higher emissions. On the other hand, it may transfer cleaner technology and management practices, implying a potentially positive contribution to sustainability in more regulated or higher-income settings [99,100]. The sign of β₃ is therefore theoretically ambiguous and must be determined empirically. On the other hand, higher renewable energy consumption is expected to enhance environmental sustainability, as it displaces fossil fuel use and lowers emissions, β₄ > 0, when environmental sustainability rises with cleaner energy, consistent with the findings of Dogan and Seker [54] and Zoundi [101]. The impact of renewable energy on environmental sustainability may also exhibit nonlinear characteristics. At early stages of adoption, renewable energy may not significantly displace fossil fuel consumption, resulting in a scale effect that may temporarily increase emissions. However, beyond a certain threshold level of renewable energy penetration, the substitution effect becomes dominant, leading to a decline in emissions. This implies that the relationship between renewable energy and environmental sustainability may follow a nonlinear or threshold-based pattern, even though the current empirical specification captures the average linear effect.

To capture persistence in environmental outcomes and the dynamic adjustment process, a dynamic extension of the baseline model incorporates a lagged dependent variable:

ES_it = δ ES_it−1 + α_i + β₁URB_it + β₂ENC_it + β₃FDI_it + β₄REN_it + γ Z_it + ε_it

where 0 < δ < 1 captures the inertia of environmental outcomes. A positive and significant δ reflects the tendency for emissions (and thus environmental sustainability) to exhibit substantial persistence over time, as documented in many dynamic panel studies [102,103]. The long-run effects of the structural variables are obtained by dividing the short-run coefficients β_j(1 − δ), which is consistent with environmental adjustment processes whereby current emissions depend on past emissions and accumulated capital stocks.

The inclusion of environmental taxation as a moderating variable systematically changes the impact of urban expansion and energy consumption on environmental sustainability, but the effectiveness of environmental taxation is conditional on institutional quality. Strong regulatory frameworks, effective monitoring, and low levels of corruption enhance the transmission mechanism of environmental taxes by ensuring compliance and reducing tax evasion. In contrast, weak institutional structures may dilute the impact of environmental taxes, limiting their ability to alter production and consumption behaviour. Thus, although environmental taxation is modelled as a direct and moderating policy instrument, its realised impact may vary across countries depending on governance quality and administrative capacity. Let ETAX_it denote environmental taxes, based on Pigouvian tax theory. A higher environmental tax raises the private cost of environmentally harmful activities, and as a result, emissions are reduced [17,54,104,105]. In the case of European economies, environmental taxes reduce emissions directly, and their effectiveness increases beyond certain tax intensity thresholds [74].

In a moderated specification, environmental taxation enters both as a direct regressor and through interaction terms with urbanisation and energy consumption, and here NI is the level of per capita income. The static moderated model can be written as

ES_it = α_i + β₁URB_it + β₂ENC_it + β₃FDI_it + β₄REN_it + β₅ETAX_it + β₆NI_it + θ₁(URB_it × ETAX_it) + θ₂(ENC_it × ETAX_it)
+ γ Z_it + ε_it

In this formulation, β₅ captures the direct effect of environmental taxation on environmental sustainability. When environmental sustainability is inversely related to emissions, theory and the empirical literature suggest β₅ > 0, because higher environmental taxes should reduce emissions and therefore increase environmental sustainability [104]. The parameters θ₁ and θ₂ represent the moderating effects. They indicate whether environmental taxes weaken or strengthen the marginal impact of urbanisation and energy consumption on environmental sustainability.

The marginal effect of urban expansion on environmental sustainability, conditional on environmental taxation, is given by

$${\displaystyle \frac{\partial ESit}{\partial URBit}={\mathit{\beta }}_1+{\mathit{\theta }}_1{ETAX}_{it}}$$

and the marginal effect of energy consumption is

$${\displaystyle \frac{\partial ESit}{\partial ENCit}={\mathit{\beta }}_2+{\mathit{\theta }}_2{ETAX}_{it}}$$

If, as expected, urban expansion and energy consumption are harmful for environmental sustainability in the absence of strong policy (that is, β₁ < 0 and β₂ < 0), and if environmental taxes mitigate these adverse effects, then the interaction coefficients should be positive (θ₁ > 0 and θ₂ > 0). This shows that, with higher environmental taxation, the negative impact of urbanisation and energy consumption on environmental sustainability becomes weaker and, after a certain level, becomes positive [74].

A dynamic version of the moderated model can be written as

ES_it = δ ES_it−1 + α_i + β₁URB_it + β₂ENC_it + β₃FDI_it + β₄REN_it + β₅ETAX_it + β₆NI_it + θ₁(URB_it × ETAX_it) + θ₂(ENC_it ×
ETAX_it) + γ Z_it + v_it

Here, δ again captures persistence in environmental outcomes, while the long-run moderated effects are computed as (β₁ + θ₁ETAX_it)/(1 − δ) for urbanisation and (β₂ + θ₂ETAX_it)/(1 − δ) for energy consumption. Thus, in the long run, effective environmental taxation encourages urbanisation, and energy consumption becomes less harmful to environmental sustainability or even positive at higher levels of ETAX_it.

For the empirical analysis, a panel of 110 developing countries was selected for the period 2010 to 2024. The list of selected countries is provided in Appendix A. The variables used in this study are obtained from the World Development Indicators database. Urban expansion is measured as the annual growth rate of the urban population, which captures the pace of urbanisation and spatial expansion in developing economies. Therefore, negative values may occur in cases where urban population growth slows down or declines due to migration patterns, demographic transitions, or reclassification of urban areas. Such observations are consistent with cross-country panel data and do not imply a reduction in the level of urbanisation, but rather a deceleration in its growth. Environmental taxes are proxied using total environmental tax revenue (in current US dollars), as reported in the World Development Indicators (WDIs). It reflects the fiscal magnitude of environmental policy rather than its exact stringency or effective tax rates. While this measure captures the overall scale of environmental taxation, it may not fully reflect differences in policy design, enforcement, or coverage across countries. The relatively large values observed in the descriptive statistics reflect cross-country differences in economic size and tax capacity, as the variable is expressed in absolute terms rather than as a percentage of gross domestic product or per capita measure. To ensure robustness and comparability, logarithmic transformations are also applied in selected model specifications (

Table 1. Variable definitions and data sources.

Variables	Definitions	Measurements	Sources
Environmental Sustainability	Proxy using greenhouse gas emissions	Metric tons per capita (or total)	WDI
Urban Expansion	Growth of urban population	Annual percentage growth	WDI
Energy Consumption	Energy use	kg of oil equivalent per capita	WDI
Foreign Direct Investment	Net inflows	% of GDP	WDI
Renewable Energy	Renewable energy consumption	% of total energy use	WDI
Environmental Taxes	Environmental tax revenue	Current US dollars	WDI
Income	Gross national income	Current US dollars per capita	WDI

).

Empirical Strategy

To ensure a coherent identification strategy, this study adopts a multi-step empirical framework. First, panel ordinary least squares estimation is used as a baseline to establish the average relationship between urban expansion, energy consumption, and environmental sustainability. Second, robust least squares estimation is employed to address the influence of outliers and non-normality in the data. Third, quantile regression is applied to capture heterogeneity across the conditional distribution of emissions, allowing the effects to vary across low- and high-emission countries. Finally, the generalised method of moments estimator is employed to address potential endogeneity arising from simultaneity, omitted variables, and the inclusion of lagged dependent variables. The estimation uses internal instruments, where lagged levels and differences in the explanatory variables serve as instruments, following the Arellano–Bond framework. The validity of the instruments and model specification is evaluated using standard diagnostic tests, including the Hansen test of overidentifying restrictions and the Arellano–Bond tests for serial correlation.

4. Results and Discussion

This section presents the empirical results following a structured estimation strategy. We begin with baseline panel ordinary least squares estimates to establish the average relationships among the variables. These results are then validated using robust least squares to account for outliers and distributional irregularities. Next, quantile regression is employed to explore heterogeneity across different emission levels, providing insights into how the relationships vary across countries. Finally, the dynamic panel generalised method of moments estimation is used to control for endogeneity and persistence in emissions. Across all models, the moderating role of environmental taxation is consistently examined through interaction terms, allowing for a unified interpretation of results. It is important to note that the results should be interpreted in terms of emission-based environmental sustainability, rather than a fully comprehensive measure of environmental performance.

Table 2. Descriptive statistics and correlation matrix.

	ES	URB	ENC	FDI	RNE	ETAX	Mean	Median	Maximum	Minimum	Std. Dev.	Skewness	Kurtosis	Jarque-Bera	Observations
ES	1						219.5	60.50	3426.3	1.618	538.4	3.89	17.49	7114.5	1650
URB	−0.1 ***	1					2.36	2.02	31.2	−4.196	2.19	5.28	63.61	99,382	1650
ENC	−0.1 ***	−0.2 ***	1				46.05	43.14	99.9	0.00	33.53	0.13	1.70	45.6	1650
FDI	0.62 ***	−0.3 ***	0.30 ***	1			0.19	0.14	0.66	0.03	0.13	1.49	4.86	324.3	1650
RNE	0.13 ***	0.38 ***	−0.8 ***	−0.3 ***	1		45.78	44.25	100	0.30	32.2	0.15	1.77	41.67	1650
ETAX	0.78 ***	−0.1 ***	−0.0236	0.72 ***	−0.01	1	1285.4	187.1	27,466	0.003	3199	4.29	24.42	13,985	1650

ES represents greenhouse gas emissions (proxy for environmental sustainability). URB is measured as the urban population growth rate. ENC denotes energy consumption. FDI represents foreign direct investment inflows. RNE is renewable energy consumption. ETAX represents total environmental tax revenue in current US dollars. The large variation in ETAX reflects differences in country size and fiscal capacity across the sample of developing economies. Note: *** indicate significance at 1 percent level.

presents descriptive statistics and the correlation matrix of the selected variables of the study. The outcomes indicate that there is substantial cross-country dispersion in greenhouse gas emissions, as the mean is greater than the median and there is sufficient distance between the minimum and maximum values. This pattern shows that at the upper tail, a small subset of economies have high emissions, a feature that is frequently observed in panel analysis on emissions across developing countries. Urban expansion exhibits modest central tendencies, but its spread has some negative observations. These negative observations explain that developing countries have urban stagnation and periodic demographic adjustment [50,106]. Energy consumption and renewable energy display heterogeneity, showing that the selected economies have diversified energy structures and higher renewable penetration. Foreign direct investment and environmental tax have lower central tendencies with large standard deviations. The large dispersion in environmental tax values is primarily driven by cross-country differences in economic size, as the variable is measured in absolute monetary terms. This shows that sizeable foreign capital inflows and mature environmental tax regimes are concentrated among selected developing countries, and most of them are at earlier stages of green fiscal reform and global capital market integration. Moreover, the distributional features of the data show that due to non-normality, traditional methods are inappropriate for the data analysis; thus, we have used panel least squares, robust least squares, quantile regression, and the generalised method of moments, which are robust in the presence of non-normality, heterogeneity, and cross-dependence.

The results show that greenhouse gas emissions are positively correlated with foreign direct investment and environmental tax; they have a weak correlation with renewable energy consumption, whereas they have a negative correlation with urban expansion and energy consumption. The positive correlation of emissions and foreign direct investment indicates that in developing countries, the inflows are energy-intensive manufacturing, construction, or extractive activities unless regulatory constraints are stringent. The positive correlation between emissions and environmental taxation shows that with larger emission burdens, they tend to be more likely to expand green fiscal instruments. This correlation points to policy endogeneity and motivates modelling the marginal role of taxation through interaction effects and dynamic specifications, where the objective is to evaluate whether taxation weakens the emission response to urbanisation and energy use. The negative correlation between greenhouse gas emissions and urban expansion may appear counterintuitive under a purely mechanical view of urbanisation as emission-increasing. However, similar patterns have been reported in samples that combine rapidly urbanising low-emission economies with slower-growing but highly industrialised, high-emission economies. In such cases, simple correlations can mask heterogeneity in the urbanisation process, compact, infrastructure-efficient urban growth versus dispersed expansion, and differences in the energy mix and industrial composition. The outcomes show that the partial impact of urban expansion becomes positive as income, structure of energy, level of industrial agglomeration, and policy variables are controlled. The positive correlation between renewable energy and emissions explains that renewable energy consumption is in its initial stage, as developing countries’ energy sector is still dominated by traditional methods with higher emissions.

To check the stationarity of the selected variables, we have applied Levin–Lin–Chu, Im–Pesaran–Shin [107], Fisher-Type Augmented Dickey–Fuller [108] and Phillips–Perron tests [109], and the Hadri [110] and HC-Z [111] unit root tests. The estimated results have been provided in

Table 3. Unit root outcomes.

	LLC		IMPSW		ADF-F		PP-F		Hadri		HC-Z
Variables	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)
ES	−2.9 ***	−12.2 ***	−2.43 ***	−13.7 ***	67.26 ***	257.1 ***	76.4 ***	468.5 ***	3.516 ***	1.638 **	7.747 ***	3.749 ***
URB	−4.3 ***	−16.8 ***	−3.47 ***	−16.4 ***	81.1 ***	247.4 ***	75.4 ***	231.6 ***	6.52 ***	8.69 ***	7.94 ***	3.08 ***
ENC	−3.9 ***	−17.1 ***	−6.3 ***	−16.5 ***	58.4 ***	254.7 ***	58.07 ***	447.2 ***	7.74 ***	1.91 **	8.11 ***	2.50 ***
FDI	1.007	−10.2 ***	1.14	−12.9 ***	39.23	238.9 ***	45.39	447.7 ***	14.01 ***	2.10 ***	9.54 ***	1.97 ***
RNE	−2.6 ***	−12.3 ***	−0.98	−14.4 ***	54.4 *	269.4 ***	56.09 *	469.4 ***	10.7 ***	2.97 ***	9.05 ***	2.74 ***
ETAX	−6.6 ***	−9.2 ***	−10.5 ***	−19.3 ***	194.8 ***	369.4 ***	339.5 ***	515.2 ***	0.17	2.91 ***	1.15	4.42 ***

Note: ***, **, and * indicate significance at 1 percent, 5 percent, and 10 percent levels, respectively.

; the outcomes show that greenhouse gas emissions, urban expansion, and energy consumption are stationary at the level and first difference for all the tests. The balanced outcomes of the unit root reveal that emissions, urban expansion, and energy consumption are integrated at the same order, suggesting that traditional, dynamic panel approaches are favourable for the empirical analysis. Foreign direct investment, renewable energy consumption, and environmental taxes have a mixed order of integration. This suggests that dynamic methodologies are most suitable for the empirical analysis. Overall, unit root test outcomes support static and dynamic panel estimators, with and without moderation analysis.

Table 4. Panel ordinary least squares and robust least squares estimates with and without moderation.

	Dependent Variable: GHG Method: Panel OLS		Dependent Variable: LGHG Method: Panel OLS		Dependent Variable: GHG Method: Robust Least Squares		Dependent Variable: LGHG Method: Robust Least Squares
Explanatory Variables	Without Moderation	With Moderation	Without Moderation	With Moderation	Without Moderation	With Moderation	Without Moderation	With Moderation
URB	−5.460875	−50.21269 ***	0.067872 ***	−0.15990 ***	2.134769 ***	−2.458872 **	0.210259 ***	−0.255503 ***
ENC	0.207792 **	−4.811062 ***	0.015299 ***	0.001489 *	0.254206 ***	−0.426840 ***	0.005717 ***	−0.009328 ***
FDI	3255.207 ***	−1762.279 ***	8.516241 ***	4.503902 ***	141.6961 ***	98.45949 ***	8.885693 ***	4.703924 ***
RNE	7.969286 ***	3.256426 ***	0.026668 ***	0.020438 **	0.510764 ***	0.489127 ***	0.017741 ***	0.015225 ***
NI	2.170702	0.839907	0.018002 **	0.003185	0.383243	−0.171335	0.007791	−0.005994
ENT	-	−0.088786 ***	-	0.000121 ***	-	−0.124537 ***	-	−8.53 × 10−5 ***
URB × ETAX	-	−37.82211 ***	-	0.141684 ***	-	−0.063274 ***	-	−0.187691 ***
ENC × ETAX	-	−2.672452 ***	-	0.005122 ***	-	−0.032606 ***	-	−0.006775 ***
C	−799.2497	−403.3469 ***	0.194223	1.049722 ***	−11.18287	−33.27761 **	0.841647 ***	1.650670 ***

Note: ***, **, and * indicate significance at 1 percent, 5 percent, and 10 percent levels, respectively.

provides the baseline static panel estimates for greenhouse gas emissions both without and with logarithmic form, by using panel least squares and robust least squares, without and with the moderating role of environmental taxation. The results without moderation show that urban expansion is insignificant without log specification, but positive and highly significant in the log specifications. These findings are consistent with the theoretical expectation that urbanisation raises energy demand, transport intensity, and industrial activities. Thus, there is a rise in emissions due to the developing countries’ ineffective planning and regulatory structure [112]. The previous literature on Asia and Africa found long-run positive effects of urban expansion on emissions, where rising urban expansion is attached to a traditional fossil fuel-based energy structure and carbon-intensive infrastructure [42,43,44,65,66,113,114]. However, it is important to recognise that the positive impact of urban expansion on emissions reflects an average effect across heterogeneous forms of urban development. In many developing countries, urban expansion is often characterised by unplanned and low-density sprawl, weak land-use regulation, and inadequate infrastructure, which tend to increase energy consumption and emissions. In contrast, planned and compact urban development could potentially mitigate these effects by improving energy efficiency, reducing transport intensity, and promoting sustainable land use. Therefore, the estimated positive relationship should not be interpreted as universal but rather as reflecting the dominant pattern of urban expansion in the sampled developing economies.

Energy consumption also behaves as expected in the non-moderated models. The coefficients on energy consumption are positive and statistically significant in both the level and log specifications under panel ordinary least squares and robust least squares. This implies that higher energy use is strongly associated with higher greenhouse gas emissions, which is consistent with the central role of fossil fuels in the energy mix of many developing countries [42,43,44,102]. Numerous panel studies for developing and emerging economies similarly find that energy consumption is the dominant driver of emissions and ecological footprints, even after controlling for income, trade, and structural variables [57,65,98].

The results show that foreign direct investment has a positive and significant impact on emissions for all specifications, with and without moderation. In developing countries, foreign direct investment inflow is attached to higher emissions; this pattern follows the pollution haven and scale effects [61,63,65,115]. There are other studies that found a negative relationship between foreign direct investment and CO₂ emissions [23,25,116,117,118]. Our results support the concept that weaker regulatory structure and emission-intensive activities and foreign direct investment raise emissions in low- and middle-income regions [23,25,119,120]. The inclusion of environmental tax as a moderator changes this relationship to some extent, but the overall positive effect remains, reinforcing the need for integrating environmental tax design with foreign investment and industrial policies.

Renewable energy consumption has a positive and significant impact on emissions, which at first glance appears counterintuitive given its expected role in mitigating environmental degradation. However, this outcome can be explained through the dominance of the scale effect over the substitution effect in the context of developing economies [26,27,28,29,121]. In many of these countries, renewable energy expansion occurs alongside rapid growth in overall energy demand driven by urbanisation, industrialisation, and population increase [57,59,65,66,122]. As a result, although renewable energy capacity is increasing, it remains insufficient to displace fossil fuel consumption at a scale large enough to reduce total emissions. Instead, renewable energy is often added to the existing energy mix rather than replacing carbon-intensive sources [69,70,123,124]. Importantly, this relationship is not linear and is expected to change beyond a certain threshold level of renewable energy penetration. At lower levels of renewable energy adoption, the scale effect dominates, as total energy demand continues to expand and fossil fuels remain the primary energy source. However, once renewable energy reaches a critical threshold—where it begins to significantly substitute for fossil fuel-based energy—the substitution effect becomes dominant, leading to a reduction in emissions. Although the present study does not explicitly estimate this threshold level, the existing empirical literature suggests that such nonlinear dynamics are common, particularly in economies undergoing energy transition. Therefore, the positive coefficient observed in this study reflects an early-stage transition phase, and the environmental benefits of renewable energy are expected to materialise more strongly as its share in the energy mix increases beyond this threshold.

The income variable, measured as gross national income, is insignificant in the non-moderated models, with only a small positive effect in the log ordinary least squares specification. This indicates that, once urban expansion, energy use, foreign direct investment, and renewable energy are controlled for, income per capita does not have strong additional explanatory power for emissions within the sample. These findings support the EKC hypothesis as follows: with nonlinear specification at the early stage, income and emissions move in the same direction in developing countries [25,27,29,32,125]. In many panels, income becomes insignificant once energy, urbanisation, and structural factors are included, explaining that the environmental trajectory is determined more by the composition of growth and the policy environment than by income levels per se [57,65]. Our findings indicate improvements in emission-based environmental sustainability rather than all dimensions of environmental quality.

To provide an intuitive interpretation of the moderating role of environmental taxes, the interaction terms can be understood as shifting the slope of the relationship between urban expansion, energy consumption, and emissions. In the absence of environmental taxes, increases in urban expansion and energy consumption lead to higher emissions. However, environmental taxes have a negative and statistically significant coefficient in all moderated specifications, both in level and log-dependent variables and under ordinary least squares and robust least squares. This indicates that, holding other factors constant, a higher environmental tax is associated with lower greenhouse gas emissions, which is precisely the expected direction if environmental taxes are effectively internalising environmental externalities and incentivising cleaner technologies and energy efficiency [17,74,104]. A growing body of evidence for European and Organisation for Economic Co-operation and Development countries finds that environmental taxes, especially energy and carbon taxes, significantly reduce emissions, particularly when set above certain thresholds and when taxes are recycled into green investment or used to lower distortionary taxes [69,74]. The results confirm that environmental taxes significantly reduce emissions and moderate the impact of urban expansion and energy consumption; these effects should be interpreted in light of institutional heterogeneity across developing countries. The effectiveness of environmental taxation depends critically on enforcement capacity, regulatory transparency, and governance quality. In countries with weaker institutions, the observed effects may be smaller in practice due to compliance gaps, informal economic activity, and limited monitoring mechanisms. Therefore, the estimated coefficients capture an average effect, which may vary depending on institutional strength.

The moderation between urban expansion and environmental taxes and between energy use and environmental taxes shows that environmental taxes diminish the positive impact of urban expansion and energy consumption on CO₂ emissions. In the presence of higher environmental tax intensity, a positive change in urban expansion or in energy consumption increases emissions at a lesser rate or may reduce them. These findings support the concept that environmental taxes force urban expansion and energy consumption to reduce carbon-intensive configurations [4,126]. In the case of developed countries, environmental taxes significantly reduce fossil fuel energy intensity and emissions from urban, industrial, and residential sectors [127]. In log specifications, urban expansion and energy consumption hurting emissions moderation of environmental tax has been incorporated. The marginal effect of urban expansion on emissions depends on the level of environmental taxation and is given by the sum of the coefficient on urban expansion and the interaction term. Therefore, the negative coefficient in the moderated model does not imply a direct reversal effect, but rather indicates that the marginal impact of urban expansion becomes more negative as environmental taxation increases. This suggests that environmental taxes amplify the emission-reducing effect of urban expansion under certain conditions, rather than simply reversing its sign. This outcome is consistent with the EKC hypothesis and green transitions, as due to strong policy instruments, urban expansion and energy consumption reduce emissions [128,129]. The outcomes of the moderation model are robust across ordinary least squares and robust least squares estimators. While the baseline and robust estimates provide consistent evidence on average effects, they do not capture potential heterogeneity across countries or account for dynamic adjustment processes. Therefore, we extend the analysis using quantile regression and dynamic panel techniques.

Table 5. Quantile regression and generalised method of moments estimates with and without moderation.

	Dependent Variable: GHG Method: Quantile Regression		Dependent Variable: LGHG Method: Quantile Regression		Dependent Variable: GHG Method: GMM		Dependent Variable: LGHG Method: GMM
Explanatory Variables	Without Moderation	With Moderation	Without Moderation	With Moderation	Without Moderation	With Moderation	Without Moderation	With Moderation
URB	13.02544 ***	−4.257007 ***	0.194392 ***	−0.172490 ***	0.183340 *	−8.161615 ***	−0.023028 *	−0.047924 ***
ENC	0.563929 *	−1.308609 ***	0.002723 ***	−0.010890 ***	0.156682 ***	−0.318824 ***	−0.001572 **	−0.003953 ***
FDI	1250.398 ***	−100.1352	8.432932 ***	4.868950	190.9020 ***	−2.500836	0.421377	1.101307
RNE	1.645907 ***	0.265333	−0.013407	−0.019920 **	−0.01566 ***	−0.553786 ***	−0.000855 **	0.000902 **
NI	0.154696	−0.352121	0.010541 ***	−0.024634 ***	−0.09908 ***	0.337620 ***	−0.00201 ***	0.003657
ENT	-	−0.124203 ***	-	7.10 × 10−5 **	-	−0.008928 ***	-	−4.50 × 10−5 ***
URB × ETAX	-	−4.766703 ***	-	0.116176 ***	-	−3.607480 ***	-	−0.020962 ***
ENC × ETAX	-	−0.704720 ***	-	0.009037 ***	-	−0.354402 ***	-	−0.001522 ***
C	−251.438 ***	−12.17239	1.340125 ***	1.571377 ***	-	-	-	-
GHG (−1)	-	-	-	-	0.728063 **	0.673539 ***	0.492334 ***	0.471634 ***
AR(1) (p-value)	-	-	-	-	0.001	0.001	0.003	0.003
AR(2) (p-value)	-	-	-	-	0.214	0.214	0.287	0.287
Hansen Test (p-value)	-	-	-	-	0.356	0.356	0.421	0.421

Note: ***, **, and * indicate significance at 1 percent, 5 percent, and 10 percent levels, respectively.

reports the results from quantile regression and dynamic panel generalised method of moments specifications for greenhouse gas emissions, using both level and log dependent variables, with and without the moderating role of environmental taxes. In the quantile regressions without moderation, the urban expansion and energy consumption are positive and significantly impact emissions in both the level and logarithmic specifications. Quantile regression estimates the conditional median (or another quantile) rather than the conditional mean, so the positive coefficients imply that, for countries at typical emission levels, increases in urbanisation and energy use are associated with higher greenhouse gas emissions, even after controlling for foreign direct investment, renewable energy, and income [130]. This shows that a major part of foreign capital inflows in developing countries is still attached to higher emissions; this finding supports pollution haven theory.

Renewable energy is positively linked to the level of emissions without a moderation quantile model. This explains that developing countries are still attached to fossil fuel or non-renewable energy consumption [131], and renewable energy has a very small portion in overall energy consumption [132]. The inclusion of the moderation term as environmental taxation in quantile regression has changed the pattern of the relationship. Individually, environmental taxation hurts emissions, and following the Pigouvian tax framework, higher environmental taxes lower emissions in developing countries [133]. Studies [134,135,136] from the developed countries also have similar findings. The outcomes show that the moderation of environmental taxes with energy consumption and urban expansion have a negative relationship with emissions in developing countries. This evidence explains that in the presence of environmental taxation, urban expansion and energy consumption reduce emissions [137]. The countries having strong environmental tax systems, urban development, and energy consumption are associated with cleaner technologies, energy efficiency, and structural change to reduce typical emissions [138].

The estimated results of the log specification of quantile regression show that without moderation, energy consumption, urban expansion, level of income, and foreign direct investment have a positive and significant impact on emissions. This confirms that energy consumption, urban expansion, level of income, and foreign direct investment are still at the early stages of EKC and raise the level of emissions [139]. But after the inclusion of environmental tax as a moderator in both level and log specifications, urban expansion and energy consumption are reducing the level of emissions in the developing countries. The results show that the quantile framework with moderating effects indicates that vast behavioural shifts are necessary to control the extreme emitters in developing countries.

The results of both level and log specifications of the generalised method of moments show that the dynamic coefficient of the non-moderated model is positive and significant, which explains that the current level of emissions is also affected by its own past values. Without the moderation and log model, urban expansion, foreign direct investment, and energy consumption have a positive and significant impact on emissions. Renewable energy consumption and level of income have a negative and significant impact on emissions, which reveals that these two factors are closely related to the structural conditions of the countries. The developing countries are improving their technologies and structural conditions to reduce the level of emissions [140,141].

In a dynamic setting, the moderation of the environmental taxes model shows that urban expansion and energy consumption are reducing the level of emissions in developing countries. Environmental taxes have a direct and indirect role in cutting emissions and also help the structural drivers, e.g., urban expansion and energy consumption, over time. This explains that strong fiscal policy with environmental regulations can accelerate the transition towards cleaner energy, more efficient urban infrastructure, and less carbon-intensive growth paths [25,27,133]. The lagged emissions outcomes show that although emissions have a decreasing trend, they are slightly smaller than in the non-moderated case.

In the log-based GMM specifications, the moderating effects of environmental taxes are different from those in the level models. Without moderation, urban expansion and energy consumption in the log GMM are small and negative, and renewable energy and income depress emissions. Thus, by controlling the endogeneity and past emissions, the proportional increase in selected factors results in the lower growth rates of emissions being witnessed [64,65,129,131,132]. In the presence of environmental taxes as a moderating factor, both urban expansion and energy consumption are negative and significant. Therefore, in the log specification, environmental taxes strengthen the relationship to reduce emissions [133,134,135,140]. The variation in the sign of renewable energy coefficients across specifications reflects differences in estimation techniques and underlying dynamics. In static models, renewable energy may capture scale effects associated with expanding energy systems, particularly in developing countries where fossil fuels remain dominant. In contrast, dynamic models that control for persistence and endogeneity tend to reveal the long-run substitution effect of renewable energy, leading to emission reductions. This divergence is consistent with the transitional nature of energy systems and highlights the importance of model specification in capturing short-run versus long-run effects.

The validity of the dynamic panel estimates is confirmed by standard diagnostic tests. The Arellano–Bond test for first-order serial correlation is significant, while the second-order serial correlation test is insignificant, indicating no evidence of higher-order autocorrelation in the residuals. Furthermore, the Hansen test of overidentifying restrictions fails to reject the null hypothesis, suggesting that the instruments used in the model are valid. The number of instruments is also kept within acceptable limits relative to the number of cross-sectional units, reducing concerns regarding instrument proliferation and overfitting.

5. Conclusions and Suggestions

This study has examined the impact of urban expansion and energy consumption on environmental sustainability, specifically in the presence of environmental taxes as a moderating factor for the 110 developing countries over the period of 2010 to 2024. We have conducted static and dynamic analyses with the help of panel least squares, robust least squares, quantile regressions, and the generalised method of moments. The results show that urban expansion and energy consumption are, on average, associated with higher greenhouse gas emissions in the absence of strong policy instruments. The developing countries typically increase energy consumption, motorisation, and construction activity in ways that intensify environmental pressure when planning and regulation are weak. Foreign direct investment tends to worsen environmental pressures, while renewable energy consumption has so far taken place largely in high-emitting contexts and therefore does not yet fully offset the upward trend in emissions. However, environmental taxes directly reduce emissions and significantly alter the marginal effects of both urbanisation and energy consumption. Across estimation strategies, the moderation term between environmental taxes and the structural drivers is statistically significant and indicates that environmental taxation weakens, and in some cases reverses, the harmful impact of urban expansion and energy consumption on emissions. The consistently negative and statistically significant coefficients on environmental taxes across static, robust, quantile, and dynamic models are highly informative. They indicate that, even in developing countries where environmental tax systems are relatively young and sometimes narrow, higher environmental tax intensity is associated with lower emissions once other structural factors and dynamics are controlled for. These patterns are consistent with the theoretical expectations from Pigouvian taxation and the Environmental Kuznets Curve, which predict that strong environmental policy can shift the relationship between structural drivers and environmental outcomes, enabling conditional decoupling. In simple and mean-based models, renewable energy consumption is positively associated with emissions, reflecting the fact that countries with higher renewable shares in this sample are often those with larger energy systems and higher overall emissions. However, in dynamic generalised method of moments specifications that better capture long-run adjustment and address endogeneity, renewable energy acquires a negative coefficient, indicating that, over time, higher renewable energy penetration contributes to emission reductions. This pattern suggests that renewable energy deployment in developing countries is still in a transitional phase, where the scale effect dominates in the short run, but the substitution effect becomes more visible when dynamics are accounted for. The implication is not that renewable energy fails to mitigate emissions, but rather that its benefits become more apparent in a medium- to long-term perspective and when accompanied by strong policy instruments such as environmental taxes.

5.1. Policy Implications

The results carry several concrete policy implications for governments in developing countries that aim to balance urban development, energy needs, and environmental sustainability.

First, negative and significant environmental tax coefficients suggest that even relatively modest tax regimes can exert measurable downward pressure on emissions. Policymakers should therefore broaden the base and raise the rates of environmental taxes in a gradual but predictable manner, focusing on energy and carbon taxes that directly price greenhouse gas emissions and local pollutants.

Second, in rapidly urbanising economies, congestion pricing can be implemented in major cities to directly reduce transport-related emissions by discouraging excessive private vehicle use and encouraging public transport adoption. Similarly, land value capture mechanisms can be used to finance sustainable urban infrastructure by taxing the increase in land values generated by public investments, thereby promoting compact and efficient urban development. These instruments are particularly relevant in managing the environmental consequences of urban expansion identified in this study. Moreover, policymakers should promote compact, transit-oriented, and well-planned urban development to reduce the environmental footprint of cities. In contrast, unregulated urban sprawl should be discouraged, as it leads to higher energy consumption, transport emissions, and inefficient infrastructure use. Integrating land-use planning with environmental and energy policies can significantly enhance the sustainability of urban expansion.

Third, the positive emission effect of energy consumption and the moderating role of environmental taxes imply that energy sector reforms are central to any sustainability agenda. Governments should progressively remove fossil fuel consumption subsidies and replace them with targeted support for clean energy and efficiency investments, using environmental tax revenues to ease the transition. Well-designed energy and carbon taxes can shift the relative prices of fuels and technologies, accelerating the penetration of renewable energy and efficient appliances, while long-term power sector planning should prioritise grid upgrades, storage, and flexible resources to accommodate higher shares of variable renewables.

Fourth, the consistently positive relationship between foreign direct investment and emissions calls for a greening of investment and industrial policies. Host countries should integrate environmental performance criteria into investment promotion frameworks, for instance, by granting fiscal incentives and streamlined procedures to foreign projects that bring low-carbon technologies, while tightening environmental standards and enforcement for emission-intensive activities. Environmental taxes can support this agenda by increasing the cost of polluting operations and rewarding cleaner technologies, but they need to be accompanied by clear and credible regulatory signals so that investors internalise environmental costs in their location and technology decisions. Development partners and international financial institutions can also play a role by aligning climate finance and investment guarantees with environmental tax and regulatory reforms in host countries.

Fifth, the importance of institutional quality and data cannot be overstated. The effectiveness of environmental taxes in moderating the emission effects of urbanisation and energy use depends on administrative capacity, transparency, and enforcement, which are often constrained in developing countries. Strengthening tax administration, environmental monitoring, and statistical systems is therefore a prerequisite for effective policy implementation. Regular evaluation of environmental tax performance and its interaction with urban and energy policies, using empirical approaches similar to those in this study, can help calibrate tax rates, broaden coverage, and identify unintended distributional or sectoral impacts. Cross-country cooperation, including exchange of experiences on environmental tax design and administration, can accelerate learning and avoid repeating costly policy mistakes.

Finally, an important prerequisite for the effectiveness of environmental taxes is strong institutional quality. Governments should strengthen regulatory frameworks, improve monitoring systems, and reduce corruption to ensure that environmental taxes are effectively implemented and enforced. Without adequate administrative capacity, environmental taxes may fail to achieve their intended environmental outcomes, as firms may evade compliance or shift activities to informal sectors. Therefore, environmental tax reforms should be accompanied by institutional strengthening to maximise their effectiveness.

5.2. Limitations

Despite providing comprehensive evidence on the moderating role of environmental taxes, this study does not explicitly incorporate institutional quality variables such as governance effectiveness or regulatory stringency into the empirical model. This omission is primarily due to data limitations and the need to maintain consistency across a large panel of developing countries. Future research may extend this framework by integrating institutional indicators to examine how governance conditions influence the effectiveness of environmental taxation in shaping environmental sustainability outcomes. The environmental taxation is measured using aggregate tax revenue, which may not fully capture differences in tax rates, coverage, or enforcement across countries.

This study measures environmental sustainability using greenhouse gas emissions as an inverse proxy, which primarily captures the climate dimension of environmental performance. While this approach is widely used and suitable for cross-country analysis, it does not encompass broader aspects such as biodiversity, ecological footprints, or resource depletion. Future research may incorporate composite indicators to provide a more comprehensive assessment of environmental sustainability.

This study does not distinguish between different forms of urban expansion, such as compact urban development and low-density sprawl, due to data constraints. Since these urban forms have significantly different environmental implications, the estimated results capture an average effect that may mask important heterogeneity. Future research may incorporate spatial and structural indicators of urbanisation to better understand how different patterns of urban growth influence environmental sustainability.