New Interpretation of Human–Land Relation: Differentiated Impacts of Global Demographic Transition on Carbon Emissions

Abstract

Demographic transition and environmental governance are the most prominent focal points of global concern in the 21st century. We quantitatively evaluate the distinct carbon emission impacts of the global demographic transition by regression models based on C–D production function. Our study highlights that while demographic transition boosts per capita economic growth, it adversely affects overall economic output and aggregate economic growth, highlighting significant reductions in carbon emissions as a result of these demographic changes. However, it has a negative contribution to overall economic output and economic growth. Meanwhile, demographic transition eventually reduces carbon emissions to some extent. The relationship between population urbanization and carbon emissions mainly conforms to an inverted U-shaped curve, while some of it does to a linear growth pattern. However, the relationship between population aging and carbon emissions mainly conforms to an inverted U-shaped pattern. The impacts of demographic transition on carbon emissions confirm the universality of the EKC law in the particular production function.

1. Introduction

Environmental governance and demographic transition are two of the major changes being faced globally in today’s world. The two are interconnected, with demographic transition being essential for the long-term balanced development of human societies, while energy conservation and emission reduction are not directly related to it. The threats to society and humanity posed by climate change are deepening [1]. Among them, carbon dioxide emissions account for 70% of greenhouse gas emissions (GHGs) [2], directly threatening global ecosystems. Otherwise, population urbanization and population aging, as two major demographic transitions in modern societies [3], pose unprecedented challenges to the world [4].

According to World Energy Statistical Yearbook (70th Edition), there has been a significant rise in global carbon emissions since the beginning of the 21st century. In 2019, the total global carbon emissions hit a new peak, at 34.36 billion tons. During 2000–2019, there was a significant increase, of 40.00%, in global CO₂ emissions. To this end, various world organizations have concluded many conventions or declarations. The United Nations Framework Convention on Climate Change, adopted in 1992, is an international convention. In 1997, the Kyoto Protocol established a legal requirement for high-income countries to reduce their greenhouse gas emissions and urged industrialized countries to reduce their greenhouse gas emissions by 5.20% from the 1990 levels by the end of 2012. The Delhi Declaration, adopted in 2002, stressed that the suppression of climate change must be carried out within the framework of sustainable development. This shows that greenhouse gas emission reduction and sustainable development are still important tasks for the parties to implement this convention in the future. In 2009, the Copenhagen Accord, comprising mainly signed agreements, was adopted to reduce CO₂ emissions according to the GDP of countries. It is essential for global greenhouse gas emissions to peak as soon as possible and for net greenhouse gas emissions to reach zero in the latter half of the 21st century in order to mitigate the ecological risks posed by climate change and alleviate the existential threat climate change presents to humanity. The latest data released by the National Climate Centre show that 2023 was the world’s hottest year since meteorological records began, breaking the record for the warmest year. The Intergovernmental Panel on Climate Change has projected a potential increase in global temperatures of 1.40–5.80℃ between the years 1990 and 2100. Affected by the global COVID-19 pandemic, global carbon emissions generally decreased in 2020 and dropped to 32.28 billion tons, a year-on-year decrease of 6.30%.

Due to global warming, there is a growing consensus on the necessity of reducing carbon emissions to achieve sustainable and environmentally friendly development worldwide. However, there remains a significant gap in research regarding the impact of demographic transition on carbon emissions, with few comprehensive studies addressing this issue. Therefore, our study seeks to quantitatively investigate the impacts of demographic transition on carbon emissions within the context of population aging and urbanization. Our approach represents a novel contribution by integrating two critical global concerns and conducting interdisciplinary research that spans demography, environmental economics, and other relevant disciplines. By examining the carbon emission impacts of demographic transition around the globe, we aim to provide valuable insights into global emission trends. Furthermore, our findings are intended to offer theoretical support for policymakers as they design and implement strategies related to demographic transition in pursuit of sustainable development at a global scale.

2. Literature Review

The impacts of demographic transition on carbon emissions have been acknowledged and appreciated, but there is a lack of comprehensive research findings. While there have been various accomplishments in examining specific aspects, such as the impact of population aging or urbanization on carbon emissions, literature reviews can be categorized into two main areas.

2.1. The Relationship between Carbon Emissions and Population Aging

The relationship between population aging and carbon emissions has been studied in various countries. Although there are inconsistencies in the findings, all show that population aging is an important factor in carbon emissions. Compared to technology, population aging has a negative impact on carbon emissions in the United States and to a much greater extent than technology in some cases [5]. The elderly consume more energy-intensive products in OECD countries, leading to an increase in carbon emissions [6]. When examining macroeconomic emissions by incorporating population aging and group composition into indicators, it is found that changes in the demographic age structure increase carbon emissions in OECD countries [7]. With the aging of the population happening at a fast pace, the rise in carbon emissions is attributed to changes in the industrial structure and urbanization. Additionally, the connection between income levels and carbon emissions differs among countries in various developmental stages [8]. High per capita consumption and an increase in older age groups are the main reasons for the higher share of the total carbon footprint of older age groups in most countries [9]. Population aging promotes CO₂ emissions at the national level, with provincial differences [10]. The impact of population aging on household carbon emissions varies significantly between rural and urban areas [11]. The evolving consumption patterns, particularly the growing prevalence of small-size and elderly households, have been identified as significant contributors to the escalating energy consumption and carbon emissions in China [12].

2.2. The Relationship between Carbon Emissions and Population Urbanization

Urbanization currently creates extra global environmental pressure [13]. It is a universal truth that rapid urbanization increases carbon emissions [14]. Urbanization and carbon emissions, which have attracted widespread attention, have been two major characteristics [15] of modern global socio-economy. In the modern era, rapid urbanization has a considerable impact on CO₂ emissions, whose one main source is urbanized areas [16]. Cities are the key areas for implementing carbon reduction policies [17]. Exploring the impacts of urbanization on CO₂ emissions is of great significance [18] in striving toward a low-carbon society. Nevertheless, there have been limited investigations into the variations in carbon intensity and the driving factors at the city level [19]. The primary techniques for measurement involve the use of the ideal point cross-efficiency model [20], the stochastic frontier model [21], the hierarchical spatial autoregressive model [16], the logarithmic mean Divisia index decomposition method [22], the Panel Autoregressive Distributed Lag approach and the heterogeneous causality test [23], the tree-based regression model [19], the carbon intensity of human well-being [24], urban building carbon emissions [25,26,27], energy rebound impact [28], energy service companies [29], passenger transport of cities [19], etc. Although the improvement of methods creates extra uncertainties, some research has identified a direct link between urbanization and the release of carbon emissions [30], with urbanization’s regionally differentiated impacts on CO₂ emissions [22]. Previous studies have revealed heterogeneity in the impacts of urbanization on carbon emissions, such as new-type urbanization [31]. Moreover, the speed of the barycenter of carbon emissions is faster than that of urbanization and industrialization [32]. The inverted U-shaped relationship between urbanization and carbon emissions has been identified [21,33,34]. However, the impacts of urbanization on CO₂ emissions vary with different urbanization subsystems [18,35]. Among them, land urbanization and economic urbanization [30] display a Kuznets curve relationship with carbon emissions, which can be verified in the lower–middle-income group [36], such as India [37]. CO₂ emissions are primarily affected by urbanization in lower–middle-income countries [23]. Additionally, the EKC hypothesis is also confirmed in upper-middle- and high-income groups [38].

In conclusion, there are emerging research findings and deepening theoretical studies related to the impact of demographic transition on carbon emissions. However, it is still necessary to further quantify these impacts. Therefore, we aim to investigate the relationship between population transformation and economic, technological, and environmental factors using the Cobb–Douglas (C–D) production function. Additionally, we will explore the international-scale environmental impacts of demographic transition through regression modeling. Finally, we will construct an environmental Kuznets curve (EKC) for demographic transition and carbon emissions based on economic development indicators. Our key innovation lies in integrating population and environmental issues to demonstrate their coordinated mechanism and promote interdisciplinary research in regional governance encompassing demography, environmental economics, and other disciplines. We hope that our study can establish a new research paradigm and effectively guide policies addressing the carbon emission impacts of demographic transition across various economies.

3. Methodology and Data Sources

3.1. Data

The primary data for carbon emissions and demographic transition were sourced from the World Development Indicators database (https://data.worldbank.org/, accessed on 19 October 2023). In our study, we identified carbon emissions as the dependent variable group, with demographic transition, economic growth, industrial upgrading, and technological input serving as independent variables. Additionally, economic development was included as a control variable in constructing the regression model (

Table 1. Descriptive statistics of main data.

Variables	Specific Indicators	Mean	Std. Deviation	Max.	Min.
Energy-saving and emission reduction efficiency (EE)	Per capita CO2 emissions (t/person) (EE1)	4.40	5.44	47.70	0.02
Energy consumption structure (ES)	Proportion of nuclear energy and other clean energy consumption in total energy consumption (%) (ES1)	8.26	10.61	55.58	0.00
Carbon emissions (CE)	CO2 emissions (104 t) (CE1)	15,573	71,671	1,030,000	3.00
Demographic transition (DT)	Population aged 65 and above (% of total population) (DT1)	7.81	5.49	28.40	0.69
	Urban population (% of total population) (%) (DT2)	58.18	24.04	100.00	8.25
Growth rate (EG)	Annual GDP growth rate (%/year) (EG1)	3.33	5.63	123.14	−62.08
	Annual CO2 emission growth rate (%/year) (EG2)	3.43	15.08	350.00	−74.75
Industrial upgrading (IU)	Industry (including construction), value added (% of GDP) (IU1)	26.32	12.47	94.70	3.15
	Services, value added (% of GDP) (IU2)	54.21	12.83	96.20	10.86
Technical input (TI)	R&D expenditure (% of GDP) (TI1)	1.12	1.01	4.95	0.01
	R&D investment (104 USD) (TI2)	1,630,000	5,570,000	58,200,000	75.68
	Per capita R&D investment (104 USD/person) (TI3)	0.04	0.05	0.21	0.00
Economic development (ED)	Per capita GDP (USD/person) (ED1)	15,216.43	23,643.47	189,487.10	111.93

).

3.2. The Carbon Emission Impact Model Based on Demographic Transition

Based on previous research [3] and the Cobb–Douglas production function (Y = A^γK^αL^β + ε, α + β + γ = 1), we presented a mechanistic model to evaluate the carbon emission impacts of demographic transition and subsequently decomposed the production function into four specific categories, as delineated here.

$$\left\{\begin{array}{l}①\ln E{D}_1+\ln C{E}_2=\frac{\alpha }{\gamma }\ln T{I}_2+\frac{\beta }{\gamma }\ln D{T}_1+\frac{\beta }{\gamma }\ln D{T}_2+{\epsilon }_0\hfill \\ ②\ln E{G}_1+\ln E{G}_2=\frac{\alpha }{\lambda }\ln I{U}_2+\frac{\beta }{\gamma }\ln D{T}_1+\frac{\beta }{\gamma }\ln D{T}_2+{\epsilon }_1\hfill \\ ③\ln E{D}_1+\frac{1}{\gamma }\ln E{E}_1=\frac{\alpha }{\gamma }\ln T{I}_3+\frac{\beta }{\gamma }\ln D{T}_1+\frac{\beta }{\gamma }\ln D{T}_2+{\epsilon }_2\hfill \\ ④\ln E{D}_1=\frac{1}{\gamma }\ln T{I}_1+\frac{\alpha }{\gamma }\ln I{U}_1+\frac{\alpha }{\gamma }\ln E{S}_1+\frac{\beta }{\gamma }\ln D{T}_1+\frac{\beta }{\gamma }\ln D{T}_2+{\epsilon }_3\hfill \end{array}\right.$$

(1)

The particular production function (①) represents the impacts of demographic transition on economic output and CO₂ emissions. The industrial upgrading function (②) illustrates the impact of demographic transition on economic growth and CO₂ emission growth in the context of industrial upgrading and energy efficiency improvement. The Individual function (③) demonstrates how demographic transition affects per capita economy under the condition of increasing per capita R&D investment. The clean energy function (④) shows the influence of demographic transition on per capita economy and per capita CO₂ emissions with adjustments to the energy structure and upgrades to the industrial structure.

3.3. Operation Mechanism

Due to shifts in the age distribution of the population, there are significant changes in the sizes of different age groups, leading to carbon emission impacts from demographic transition. Some of these impacts are directly related to household energy consumption and living habits. Additionally, population aging indirectly impacts carbon emissions through production and ecology. Furthermore, rapid urbanization has direct or indirect impacts on population aging in terms of lifestyle, production, and ecology. This results in complex trade-offs that make quantitative research more challenging. In conclusion, there is a causal relationship between demographic transition and carbon emissions, which we illustrate in Figure 1 by mapping the carbon emission impacts of demographic transition. Therefore, we conducted a quantitative analysis of the carbon emission impacts of demographic transition around the globe based on C–D production function.

4. Results

Before initiating the regression analysis, we conducted unit root tests and co-integration analysis to ensure stable data through Stata17.0. To minimize the impact of varying dimensions of different variables, we initially applied a logarithmic transformation to each variable. The results passed significance and heteroscedasticity tests using SPSS26.0. We then validated the influence of demographic transition on carbon emissions through theoretical analysis combined with relevant data and quantitatively determined the contribution rates of different core variables on the dependent variables using partition models to assess their correlation impacts. Finally, by calculating regression equations for each partition model, we obtained the linkage mechanism between demographic transition and other factors. Our partition models were constructed based on the annual CO₂ emission growth rate, namely EG₂ ≤ 0.0%; 0.0% < EG₂ ≤ 5.0%; and EG₂ > 5.0%. Ultimately, based on this classification (

Table 2. Data included within every partition function.

Function	Period	Samples
		EG2 ≤ 0.0%	0.0% < EG2 ≤ 5.0%	EG2 > 5.0%	Subtotal
The particular production function	2000–2018	513	460	281	1254
The industrial upgrading function	2000–2018	1119	915	1063	3097
The individual function	2000–2018	513	460	281	1254
The clean energy function	2000–2014	384	332	199	915

), we investigated the impact of demographic transition on carbon emissions using both overall models and partition models.

4.1. The Particular Production Function

Through the particular production function, the contribution rate of each variable was obtained (Figure 2). Next, regression models were used to derive the regression equations and coefficients for each variable (

Table 3. Regression statistics of the particular production function.

Dependent Variable	Classification	Regression Model	R2	Adjusted R2	F	p-Value
ED1	Overall	$E{D}_1=\frac{D{T}_2^{2.139}\cdot D{T}_1^{0.575}\cdot T{I}_2^{0.158}}{e^{4.370}}$	0.726	0.725	1102.545	0.000
	EG2 ≤ 0.0%	$E{D}_1=\frac{D{T}_2^{1.770}\cdot D{T}_1^{0.651}\cdot T{I}_2^{0.190}}{e^{3.625}}$	0.715	0.713	425.829	0.000
	0.0% < EG2 ≤ 5.0%	$E{D}_1=\frac{D{T}_2^{2.358}\cdot D{T}_1^{0.590}\cdot T{I}_2^{0.141}}{e^{5.005}}$	0.711	0.709	374.578	0.000
	EG2 > 5.0%	$E{D}_1=\frac{D{T}_2^{2.208}\cdot D{T}_1^{0.393}\cdot T{I}_2^{0.129}}{e^{3.756}}$	0.702	0.699	217.363	0.000
CE1	Overall	$C{E}_1=e^{3.860}\cdot \frac{T{I}_2^{0.626}}{D{T}_1^{0.939}\cdot D{T}_2^{0.822}}$	0.701	0.701	977.955	0.000
	EG2 ≤ 0.0%	$C{E}_1=e^{4.433}\cdot \frac{T{I}_2^{0.624}}{D{T}_1^{0.981}\cdot D{T}_2^{0.941}}$	0.687	0.686	373.251	0.000
	0.0% < EG2 ≤ 5.0%	$C{E}_1=e^{4.274}\cdot \frac{T{I}_2^{0.670}}{D{T}_2^{1.137}\cdot D{T}_1^{0.933}}$	0.734	0.733	420.359	0.000
	EG2 > 5.0%	$C{E}_1=e^{2.718}\cdot \frac{T{I}_2^{0.585}}{D{T}_1^{0.812}\cdot D{T}_2^{0.390}}$	0.665	0.661	183.369	0.000

).

In terms of economic output, DT₂ and DT₁ are the primary drivers of ED₁. Among them, ED₁ increases by 2.14% for every 1% increase in DT₂. ED₁ increases by 0.58% for every 1% increase in DT₁. ED₁ increases by 0.16% for every 1% increase in TI₂. Specifically, DT₁ and DT₂ encourage variously the ED₁ of 66 countries, as shown in Table 3 and Figure 2, while clear classifications appear. There is significant heterogeneity in the contribution rate of DT₂ to ED₁. This means that DT₁, DT₂, and TI₃ all have positive impacts on ED₁. In these countries, the contribution rate is at its lowest when EG₂ ≤ 0.0%, while it is the highest when EG₂ > 5.0%. Similarly, the contribution rate of TI₂ to ED₁ gradually increases as EG₂ increases. However, there is an inverted U-shaped curve, that is, compared to EG₂ ≤ 0.0% (Figure 2b), and the contribution of DT₁ decreases when 0.0% < EG₂ ≤ 5.0% (Figure 2c) but increases when EG₂ > 5.0% (Figure 2d).

In terms of pollution emissions, DT₁ is the primary driver of CE₁ (Table 3). CE₁ decreases by 0.94% for every 1% increase in DT₁. CE₁ decreases by 0.82% for every 1% increase in DT₁. However, CE₁ increases by 0.63% for every 1% increase in TI₂. Specifically, TI₂ encourages variously the CE₁ of 66 countries (Table 3 and Figure 2). Regardless of the annual CO₂ emission growth rate, the contribution rate of TI₂ to CE₁ is absolute, resulting in minimal contributions from other variables. Among them, TI₂ has an obviously positive impact on CE₁. Conversely, DT₁ and DT₂ have little negative impact on CE₁.

4.2. The Industrial Upgrading Function

Through the industrial upgrading function, the contribution rate of each variable was obtained (Figure 3). Next, regression models were used to derive the regression equations and coefficients for each variable (

Table 4. Regression statistics of the industrial upgrading function.

Dependent Variable	Classification	Regression Model	R2	Adjusted R2	F	p-Value
Solution 1: EG1 > 0
EG1	Overall	$E{G}_1=\frac{e^{4.594}}{D{T}_1^{0.121}\cdot D{T}_2^{0.097}\cdot I{U}_2^{0.694}}$	0.092	0.091	92.467	0.000
	EG2 ≤ 0.0%	$E{G}_1=\frac{e^{3.878}}{D{T}_1^{0.148}\cdot I{U}_2^{0.628}}$	0.067	0.065	31.476	0.000
	0.0% < EG2 ≤ 5.0%	$E{G}_1=\frac{e^{5.332}}{D{T}_2^{0.172}\cdot I{U}_2^{0.867}}$	0.082	0.080	38.528	0.000
	EG2 > 5.0%	$E{G}_1=\frac{e^{4.104}}{I{U}_2^{0.661}}$	0.058	0.057	21.084	0.000
EG2	Overall	$E{G}_2=\frac{e^{4.426}}{D{T}_1^{0.326}\cdot D{T}_2^{0.384}\cdot I{U}_2^{0.232}}$	0.116	0.115	116.633	0.000
	EG2 ≤ 0.0%	$E{G}_2=\frac{e^{3.384}}{D{T}_2^{0.574}}$	0.048	0.046	39.427	0.000
	0.0% < EG2 ≤ 5.0%	$E{G}_2=\frac{e^{1.214}}{D{T}_1^{0.304}}$	0.057	0.056	52.583	0.000
	EG2 > 5.0%	$E{G}_2=\frac{e^{3.771}}{D{T}_1^{0.087}\cdot D{T}_2^{0.189}\cdot I{U}_2^{0.148}}$	0.063	0.060	22.411	0.000
Solution 2: EG1 < 0
EG1	Overall	$E{G}_1=\frac{D{T}_2^{0.361}}{e^{6.058}\cdot D{T}_1^{0.323}}$	0.024	0.018	4.247	0.015
	EG2 ≤ 0.0%	$E{G}_1=\frac{e^{4.544}}{I{U}_2^{0.961}}$	0.038	0.034	9.252	0.003
	0.0% < EG2 ≤ 5.0%	$E{G}_1=\frac{e^{1.361}}{D{T}_1^{0.732}}$	0.098	0.079	5.204	0.027
EG2	Overall	$E{G}_2=\frac{e^{2.690}\cdot I{U}_2^{0.441}}{D{T}_2^{0.685}}$	0.088	0.082	15.653	0.000
	EG2 ≤ 0.0%	$E{G}_2=\frac{e^{3.894}}{D{T}_2^{0.530}}$	0.048	0.044	11.003	0.001

).

In terms of economic growth, IU₂ contributes significantly to driving EG₁. In the solution 1 (EG₁ > 0) area, for every 1% increase in IU₂, EG₁ decreases by 0.69%. EG₁ only decreases by 0.12% and 0.09% for every 1% decrease in DT₁ and DT₂, respectively. Specifically, the influence of IU₂ on economic growth varies among the majority of the 163 countries, with diverse impacts on EG₁. Notably, there is considerable heterogeneity in how IU₂ affects EG₁. The negative impacts of IU₂ on EG₁ increase as EG₂ increases (Figure 3).

In terms of pollution growth, DT₁ and DT₂ are the dominant factors in EG₂. In the solution 1 (EG₁ > 0) area, EG₂ decreases by 0.39% for every 1% increase in DT₂. EG₂ decreases by 0.33% for every 1% increase in DT₁. EG₂ decreases by 0.23% for every 1% increase in IU₂. Specifically, DT₁ and DT₂ mitigate pollution emissions in most of the 163 countries, achieving various levels of decrease in EG₂. The contribution rate of DT₂ to EG₂ is positive at EG₂ ≤ 0.0%, the contribution rate transitions from positive to negative as EG₂ rises, and the negative contribution rate is greater than 0.0% < EG₂ ≤ 5.0% at EG₂ > 5.0% in Figure 3. Meanwhile, the contribution rate of DT₁ to EG₂ is negative in both the overall model and the partition model, with 0.0% < EG₂ ≤ 5.0% (Figure 3f).

4.3. The Individual Function

Through the individual function, the contribution rate of each variable was obtained (Figure 4). Next, regression models were used to derive the regression equations and coefficients for each variable (

Table 5. Regression statistics of the individual function.

Dependent Variable	Classification	Regression Model	R2	Adjusted R2	F	p-Value
ED1	Overall	$E{D}_1=e^{3.993}\cdot D{T}_2^{0.808}\cdot T{I}_3^{0.480}$	0.872	0.872	2834.941	0.000
	EG2 ≤ 0.0%	$E{D}_1=e^{5.677}\cdot T{I}_3^{0.535}\cdot D{T}_2^{0.379}$	0.887	0.886	1329.732	0.000
	0.0% < EG2 ≤ 5.0%	$E{D}_1=e^{3.591}\cdot D{T}_2^{0.897}\cdot T{I}_3^{0.482}$	0.870	0.869	1014.588	0.000
	EG2 > 5.0%	$E{D}_1=e^{2.645}\cdot D{T}_2^{1.194}\cdot T{I}_3^{0.390}$	0.822	0.820	427.006	0.000
EE1	Overall	$E{E}_1=\frac{D{T}_2^{1.119}\cdot T{I}_3^{0.172}}{e^{3.881}}$	0.547	0.546	502.753	0.000
	EG2 ≤ 0.0%	$E{E}_1=\frac{D{T}_2^{0.859}\cdot T{I}_3^{0.163}}{e^{2.867}}$	0.495	0.492	166.296	0.000
	0.0% < EG2 ≤ 5.0%	$E{E}_1=\frac{D{T}_2^{0.947}\cdot T{I}_3^{0.209}}{e^{3.305}}$	0.564	0.562	196.973	0.000
	EG2 > 5.0%	$E{E}_1=\frac{D{T}_2^{1.482}\cdot T{I}_3^{0.166}}{e^{5.252}}$	0.561	0.557	118.149	0.000

).

DT₂ has a strong correlation with the increase in ED₁. Specifically, a 1% increase in DT₂ is associated with an 0.81% increase in ED₁. Furthermore, a 1% rise in TI₃ corresponds to a 0.48% increase in ED₁. However, the impact of DT₁ is not considered significant. Notably, the upsurge in TI₃ serves as a stimulus for the economy across most of the examined 66 countries, leading to varying levels of achievement regarding ED₁. However, there exists discernible heterogeneity concerning the contribution of TI₃; this heterogeneity intensifies as ES₁ increases. This relationship exhibits an inverted U-shaped curve, that is, compared to EG₂ ≤ 0.0% (Figure 4b); the contribution of TI₃ increases when 0.0% < EG₂ ≤ 5.0% (Figure 4c), while it decreases when EG₂ > 5.0% (Figure 4d). However, the contribution of DT₂ increases with an increase in EG₂. Conversely, the influence of DT₂ amplifies with increasing EG₂ and also contributes to economic growth across most countries at various levels.

In terms of per capita pollution emissions, DT₂ is more strongly correlated with EE₁. With each 1% rise in DT₂, EE₁ increases by 1.12%. In addition, TI₃ has a lower positive contribution (0.17%). Specifically, the increase in TI₃ affects pollution emissions in most of the 66 countries, achieving an increase in EE₁. The contribution rate of DT₂ varies significantly, with the highest rate observed in countries where EG₂ exceeds 5.0% (Figure 4h), while it is the lowest in these countries when EG₂ ≤ 0.0% (Figure 4f). Similarly, the contribution rate of DT₂ also varies significantly. The countries exhibit the highest contribution rate when EG₂ ≤ 0.0% (Figure 4f), while the contribution rate is the lowest in these countries when 0.0% < EG₂ ≤ 5.0% (Figure 4g). This means that TI₃ and DT₂ have a positive impact on EE₁.

4.4. The Clean Energy Function

Through the clean energy function, the contribution rate of each variable was obtained (Figure 5). Next, regression models were used to derive the regression equations and coefficients for each variable (

Table 6. Regression statistics of the clean energy function.

Dependent Variable	Classification	Regression Model	R2	Adjusted R2	F	p-Value
ED1	Overall	$E{D}_1=\frac{D{T}_1^{0.638}\cdot D{T}_2^{2.088}\cdot T{I}_1^{0.328}\cdot E{S}_1^{0.067}}{e^{0.956}}$	0.720	0.718	484.074	0.000
	EG2 ≤ 0.0%	$E{D}_1=\frac{e^{1.514}\cdot D{T}_1^{0.587}\cdot D{T}_2^{1.756}\cdot T{I}_1^{0.521}\cdot E{S}_1^{0.044}}{I{U}_1^{0.296}}$	0.728	0.724	202.428	0.000
	0.0% < EG2 ≤ 5.0%	$E{D}_1=\frac{D{T}_1^{0.635}\cdot D{T}_2^{2.166}\cdot T{I}_1^{0.305}\cdot E{S}_1^{0.085}}{e^{1.638}}$	0.617	0.607	65.282	0.000
	EG2 > 5.0%	$E{D}_1=\frac{D{T}_2^{2.091}\cdot T{I}_1^{0.184}\cdot D{T}_1^{0.570}}{e^{1.469}}$	0.686	0.678	84.326	0.000

).

In terms of per capita economy, DT₂ and DT₁ play a major role in influencing ED₁. Specifically, for every 1% increase in DT₂, there will be an increase of 2.10%, 0.64%, 0.33%, and 0.07%, respectively, in DT₁, TI₁, ES₁, and ED₁. However, the impact of IU₁ is not considered significant. In particular, TI₁ promotes economic growth in most of the 61 countries studied, leading to varying levels of ED₁. There is noticeable diversity in how much TI₁ contributes to this impact: it has the greatest impact on countries where EG₂ ≤ 0.0%, while its influence is lowest in those with 0.0% < EG₂ ≤ 5.0% (see Figure 5).

5. Discussion

We categorized changes in total carbon emissions and changes in efficiency in 187 countries, with complete data according to the economic stage, the urbanization stage, and the population aging stage (Figure 6). In terms of economic stages, the normalized economy has a lower average annual growth rate of carbon emissions compared to the abnormalized economy. In addition, there is obvious spatial heterogeneity among the various types of economies. Among them, carbon emissions are growing at an average annual rate of less than 1% in low-economy countries, among which, some countries, represented by France and Germany, have already achieved a “carbon peak” and low-carbon transition. However, carbon emissions have increased significantly, with an average annual growth rate of nearly 7%, in countries with rapid economies, represented by low–middle-income countries in Asia and Africa.

From the perspective of urbanization stages, with population urbanization, the growth trend of CO₂ emissions shows an upward and then a downward trend. Meanwhile, the average annual growth rate of CO₂ emissions in countries in the late stages of urbanization is about 1.44%, which is only two-fifths of that in intermediate-urbanization-stage countries and one-third of that in the early-urbanization countries. CO₂ emissions have already shown a negative growth trend in European countries in the late urbanization stage, such as Denmark, Finland, and Sweden. From the perspective of population aging stages, with the process of population aging, the growth of CO₂ emissions is on a downward trend, but carbon emissions continue to grow at a high rate in the un-aging countries. CO₂ emissions are on a negative growth trend in countries with an aged society and a super-aged society. Based on the earlier model analysis, it is evident that there are notable variations in the impacts of demographic transition on carbon emissions.

5.1. The Impacts of Demographic Transition on Carbon Emissions

Considering the impacts of additional factors, the findings from regression analysis (Table 3, Table 4, Table 5 and Table 6) and basic statistics on demographic transition and various economic stages (Figure 6 and Figure 7) are presented. The results show that DT₁ and DT₂ have a strong positive correlation with economic development and contribute positively to the growth of ED₁, while DT₁ and DT₂ contribute negatively to CE₁, EG₁, and EG₂. Meanwhile, demographic transition will ultimately reduce carbon emissions to some extent.

The environmental impact of population urbanization is mainly reflected in the changes in CO₂ emissions (CE₁) at different urbanization stages. Specifically, the annual growth rate of per capita CO₂ emissions (EE₁) exceeds 2.0% in countries in the initial stage and the intermediate urbanization stage and even exceeds 5.0% in countries with 30% < PU ≤ 50%, which is much higher than that in countries in the late urbanization stage. However, per capita CO₂ emissions (EE₁) show a negative growth rate in countries with an urbanization rate over 85% in the late urbanization stage.

The impacts of population urbanization (DT₂) on carbon emissions are also reflected in the differences in CO₂ emissions (CE₁) and per capita CO₂ emissions (EE₁) across different economies. In an abnormalized economy with a high growth of DT₂ (Figure 7), the environmental impact of DT₂ is higher than that in a normalized economy, and the increase in DT₂ increases CE₁. Among normalized economies, a highly developed rapid economy has the fastest growth of CE₁, and a medium economy and a low economy have gradually decreasing average annual growth rates of CE₁. It is clear that the higher demand for energy consumption leads to an increase in CE₁ in the abnormalized economy and the rapid economy.

The impacts of population aging (DT₁) on carbon emissions are also primarily manifested in CO₂ emissions (CE₁) in normalized economies, with CO₂ emissions (CE₁) in an aged society and a super-aged society showing a negative trend. This is closely related to the life pattern of the elderly, who are more inclined to a frugal lifestyle, which reduces CO₂ emissions. The total CO₂ emissions show a rising trend in a shallow-population-aging society and a deep-population-aging society. Most of these countries are lower–middle-income countries with a late start, and the current need for industrialization and transformation is still at a high level, resulting in rising CO₂ emissions.

The environmental impact of population aging is also reflected in per capita CO₂ emissions (EE₁). The data show that the EE₁ of an aged society and a super-aged society has entered a negative growth stage, with these countries mainly located in Europe and America. Conversely, the EE₁ in countries with shallow- and deep-population-aging societies are still on an upward trend, such as those in Asia and Africa, unlike in countries with an aged society and a super-aged society. In nations with established economies over the long term, the impact of DT₁ is primarily evident in the increase in CE₁ during periods of rapid economic growth. This is because DT₁ is low and the energy demand is high in the early stage of industrialization.

5.2. The EKC Test

Considering the influence of various factors on different economies, DT₂ contributes to the growth of ED₁ and ES₁ to a certain extent, which inevitably increases EE₁. There is an obvious and long-standing spatial heterogeneity in the carbon emission impacts of population urbanization, mainly conforming to an inverted U-shaped pattern (Figure 8). Among them, population urbanization and carbon emissions conform to an inverted U-shaped curve in a normalized economy, with the peak value occurring at around 82.56%. However, in an abnormalized economy, there is a certain linear relationship between population urbanization and CO₂ emissions. Similarly, the relationship between population urbanization and carbon emissions shows a continuous trend in a low economy and a medium economy. However, in a rapid economy, the relationship between population urbanization and carbon emissions shows an inverted U-shaped curve, with an extreme value of 54.16%.

In most countries, CO₂ emissions follow an inverted U-shaped curve with population aging (Figure 8). An extreme value of CO₂ emissions in a normalized economy occurs at an approximate population aging rate of 13.00%. Among them, the relationship between population aging and CO₂ emissions also shows an inverted U-shaped pattern in a rapid economy and a medium economy, with extremes of 10.07% and 12.32%, respectively. Before reaching the extremes, an increase in the rate of population aging raises CO₂ emissions. After reaching the extremes, the population aging rate starts to have a dampening impact on the local economy. This is due to a decrease in the labor force and the potential strain on social welfare systems, which can lead to decreased productivity and an increased financial burden on the working-age population. However, in a low economy, there is a U-shaped curve between population aging and CO₂ emissions. Simultaneously, as a result of the growing demand for medical and nursing services among the elderly population, combined with the rapid expansion of the global pension industry, there has been a rise in economic activities, contributing to an uptick in CO₂ emissions.

6. Conclusions

6.1. Main Conclusions

Population aging and urbanization are two pivotal components of the current demographic transition, exerting a significant influence on both regional total CO₂ emissions and per capita CO₂ emissions. This study aimed to explore the carbon emission implications of demographic transition by examining the impact of population aging and urbanization. Drawing from empirical analysis around the globe, our key findings are as specified next.

Our research findings provide evidence that while demographic transition contributes positively to per capita economic growth, it, adversely, has a negative contribution to overall economic output and economic growth. Meanwhile, demographic transition will eventually reduce CO₂ emissions to some extent. Based on different economies, the relationship between population urbanization and CO₂ emissions mainly conforms to an inverted U-shaped curve, while some of it does to a linear growth pattern. There is obvious spatial heterogeneity in the carbon emission impacts across different urbanization regions, which is always present. Meanwhile, the relationship between population aging and CO₂ emissions mainly conforms to an inverted U-shaped pattern. The environmental impact of population urbanization is significantly stronger than that of population aging. The impacts of the population urbanization rate on carbon emissions in rapid economies, with a much higher growth rate of population urbanization, are less than those in medium and low economies.

6.2. Policy Implications

Based on these conclusions, we propose the following policy implications:

(1)

The optimization of energy consumption structures should be pursued globally, with a focus on developing renewable resources, such as solar, nuclear, and tidal energy. Low-carbon industries should be actively developed and encouraged to achieve carbon-peak and carbon-neutral goals in terms of energy consumption.

(2)

Enhancements to public transportation networks are essential. With an increasing proportion of elderly individuals who tend to adopt frugal travel habits, there exists an opportunity to reduce carbon emissions within the transportation sector.

(3)

Technological innovation should be encouraged. With population urbanization, the agglomeration effect of technological innovation should not be ignored. Enterprises and research institutes can be encouraged to share technologies, such as clean energy and energy efficiency improvement, by setting up economic development zones, research and development funds, and tax incentives.

(4)

As there are obvious differences in the levels of global economic development, interregional cooperation should be strengthened to promote high-tech industries in all countries with a low economic level to achieve global energy conservation and emission reduction.

6.3. Research Prospects

Based on the traditional EKC theory, this paper adds demographic indicators to reinterpret the human–land relationship in the 21st century and explores the different impacts of global urbanization and population aging on carbon emissions. We hope that our findings can provide strong guidance for the carbon emission effects of demographic transition in various economies. However, there are some limitations of this paper. First, it is important to note that socio-economic variables are constrained in this paper, and control variables, such as levels of education, accessibility of health services, energy prices, and policy shifts, can be added in subsequent studies. A more comprehensive explanation of the important topic of strengthening the human–land relationship is required. Second, in the process of model construction, a classification study was conducted based on the economic level, the population urbanization rate, and the population aging rate, but the EKC results obtained after multi-sample classification reflect spatial heterogeneity and also weaken its universality. Third, when analyzing the correlation between demographic factors and carbon emissions, cross-sectional data were used, which resulted in regression outcomes lacking robustness compared to panel data analyses. Subsequent studies could use panel data for testing. How to solve these limitations will be the direction of our future research.