The Bilateral Effects of Population Aging on Regional Carbon Emissions in China: Promotion or Inhibition Effect?

Abstract

To achieve the high-quality model of green, low-carbon, and sustainable development in China, it is necessary to clarify the relationship between population aging and carbon emissions in regions. Based on the panel data of 30 provinces and cities in China from 2011 to 2020, this article employs a bilateral stochastic frontier model to estimate the promotion, inhibition, and net effects of population aging on regional carbon emissions. The results show that regional carbon emissions are decreased by 15.77% due to the inhibition effect, while they are increased by 10.63% due to the promotion effect. As a result, the net effect is that regional carbon emissions are decreased by 5.14% overall due to the composite action of the above effects. In addition, population aging in eastern, western, and central regions significantly reduces regional carbon emissions. And the inhibition effect of population aging on carbon emissions increases continuously and gradually holds the dominant position during the study period. Moreover, the inhibition effect in the eastern region is stronger than that in the central and western regions, which can be strengthened by improving the level of population aging and human capital, as well as urbanization. The conclusions are conducive to providing new perspectives and empirical evidence for understanding the connection between population aging and carbon emissions, as well as policy recommendations for tackling population aging, carbon emission reduction, carbon peaking, and carbon-neutral strategic goals.

1. Introduction

Global warming is a major challenge for contemporary humanity [1]. Extreme phenomena, such as the accelerated melting of global glaciers, rise in the sea level, and polarization of droughts and floods, caused by the continuous increase in CO₂-based greenhouse gases have become more frequent in recent years [2,3], which have caused serious impacts on agricultural production, human life, and socioeconomic activities, and ultimately hindered the process to achieve sustainable development on a global scale. In this context, the promotion of low-carbon development has become an international consensus [4]. The Chinese government has conducted many attempts to undertake the responsibility of reducing carbon emissions, making emission-reduction commitments at the United Nations General Assembly, in which strategic goals were proposed that China would realize carbon peaking by 2030 and carbon neutrality by 2060. However, for a long time, China’s extensive development model characterized by factor-driven factors has led to a significant amount of energy consumption and carbon emissions, along with the rising energy consumption caused by industrialization and urbanization [5], which also leads to the long-term pressure on China to reduce carbon emissions [6]. According to the statistical data of World Energy Statistics Yearbook 2021, China’s carbon emissions increased from 8.83 billion tons to 9.90 billion tons from 2011 to 2020, indicating how severe the situation of the country’s emission reduction is. The issue of carbon emissions has become an important topic in current academic and social circles.

Meanwhile, with the continuous development of human society, the phenomenon of aging has attracted particular attention from China and even the whole world. As a populous country, China has been aging at an astonishing rate since entering aging society in 2000 [7]. By 2025, the proportion of older adults or people aged 65+ or similar is expected to increase to 14% of the total population, which means that China has gone through the aging period that has taken decades or even centuries in developed Western countries in just 25 years, and will be confronted with the dual challenges of “getting old before getting rich” and “getting old without preparation”. Population aging has become a major demographic characteristic in many countries, which brings changes in social production and consumption. And carbon emissions are mainly achieved through economic and social production and consumption, which implies that there may be a correlation between population aging and carbon emissions. Thus, clarifying the impact of population aging on CO₂ emissions is of great significance for the high-quality development of China’s economy and society. Therefore, this study will focus on China, which has the biggest carbon emissions as well as a serious aging issue. However, in view of the differences in the development of various regions in China, the provincial areas of China are divided into three regions—east, west, and center—and the effects of the relationship between aging and carbon emissions are thoroughly examined. This can serve as an important theoretical guide for regions to understand and cope with the relationship between population aging and regional carbon emissions.

There have been many insightful discussions on the relationship between population aging and carbon emissions. However, the previous research has not yet achieved a consensus conclusion, and three different kinds of viewpoints have been formed. The first kind of view confirms that population aging can increase carbon emissions. From the perspective of the national level, Tonn and Eisenberg [8] showed that population aging would lead to an increase in the use of conventional fossil fuels, which could increase carbon emissions for the entire society. Li et al. [9] established a three-sector equilibrium analysis framework based on the PET model for production, consumption, and government. On this basis, the population, economy, and carbon emissions data from 1995–2016 in China are adopted, and the empirical results showed that the relationship between population and consumption factors, especially population aging and carbon emissions, is increasingly positively correlated. In addition, from the perspective of the regional level, Zhang et al. [10] investigated the impact of population aging on national and regional carbon emissions and discovered a positive correlation between population aging and carbon emissions in China, based on the STIRPAT model and Chinese provincial panel data. From the perspective of the household level, Fan et al. [11] investigated the impacts of population aging on household carbon emissions in both rural and urban China and discovered that both urban and rural aging have a significant positive impact on household carbon emissions. The second kind of view shows us that population aging will inhibit carbon emissions. Hu et al. [12] and Kim et al. [13] pointed out that households with several generations or with a high percentage of middle-aged and older members were conducive to reducing household energy consumption. However, the third type of view argues that the importance of population aging for carbon emissions is uncertain in different circumstances. Yang and Wang [14] and Li [15] pointed out that population aging has a threshold effect on reducing carbon emissions. And Liu and Zhang [16] investigated the carbon footprint and inequality of household consumption in 25 Chinese provinces and found that there are significant variances in per capita of carbon emissions among different household groups, with older adults tending to produce fewer emissions.

Overall, the existing literature has conducted many beneficial explorations on the relationship between population aging and carbon emissions. However, these research conclusions remain controversial. And they are mainly focused on either the promotion effect or inhibition effect, separating the comprehensive impact of the two effects on carbon emissions. Moreover, there is no proper answer to whether there is heterogeneity of these various effects and the mechanisms behind it. According to the above analysis, the marginal contributions of this study might be as follows. Firstly, based on the dual attributes of population aging, this article proposes the analysis framework of the bilateral effects of population aging on regional carbon emissions, which theoretically enriches the research on the relationship between population aging and carbon emissions. Secondly, a bilateral stochastic frontier model is adopted while considering the bilateral effects of population aging; this paper specifically measures the promotion, inhibition, and net effects of population aging on carbon emissions. It also further analyzes the spatial-temporal distribution characteristics and variation patterns of the bilateral effects, which provides literature references for understanding the intrinsic mechanism of population aging on regional carbon emissions. Thirdly, this paper investigates the heterogeneous characteristics of the net effect of population aging on carbon emissions under various aging, human capital, and urbanization levels, which is conducive to objectively and comprehensively understanding the general principles of population aging on carbon emissions, and policy recommendations on how to achieve carbon neutrality in the context of population aging are, finally, provided.

The remainder of this paper is organized as follows: Section 2 describes the theoretical mechanism of the relationship between population aging and reginal carbon emission; Section 3 introduces the models, variables, and data sources adopted in this paper; Section 4 presents the empirical results and the relevant analysis; Section 5 summarizes the conclusions and proposes policy recommendations.

2. Mechanism Analysis and Research Hypothesis

The important role of population aging on local industrial structure, production modes of enterprises, and residents’ consumption patterns leads to changes in the utilization efficiency and consumption of local resources and energy, thereby affecting reginal carbon emissions [17,18]. This study explores the mechanism of population aging on regional emissions from the following two aspects.

2.1. The Inhibition Effect of Population Aging on Carbon Emissions

Population aging can bring changes in resources and energy utilization, thereby playing important roles in carbon emissions. Firstly, population aging has brought changes to the social consumption structure, thereby affecting the level of carbon emissions [19]. According to the life-cycle theory, elements such as current and future income, as well as age, can play an important role in individual consumption. It indicates that the personal consumption of older adults usually comes from their accumulated savings. Moreover, they exhibit a proclivity towards conservative spending patterns. They tend to live a low-carbon lifestyle, such as reducing their demand for transportation services, using public transportation, and saving resources [20,21], which will possibly have an inhibition impact on carbon emissions [14]. Secondly, the aging population encourages the formation of the aging industry, which is primarily the tertiary industry such as older adults’ financial products and services. It is conducive to the tertiary industry as well as energy conservation and emission reduction. Thirdly, with the development of the aging population, the demographic dividend is inevitably decreasing, and the labor cost advantage of low-end manufacturing will also gradually disappear [22]. It will promote enterprises to enhance their competitiveness through more R&D investment and technological innovation, thereby releasing the inhibition effect on carbon emissions [23]. Overall, resources and energy utilization efficiency can be raised by population aging, and regional carbon emissions will be reduced as a result. Based on the above analysis, the following hypothesis can be proposed.

H1. 

Population aging can significantly inhibit regional carbon emissions.

2.2. The Promotion Effect of Population Aging on Carbon Emissions

Population aging can bring changes in resources and energy consumption, thereby playing important roles in carbon emissions. Firstly, as the working population is aging, their difficulties in accepting new knowledge and learning new skills become higher than before [24]. Therefore, it will be difficult for older adults to adapt to the requirements of industrial development towards higher levels. Then, structural unemployment will become prevalent, ultimately lowering the labor productivity level of the entire society, hindering industrial transformation and upgrading, and promoting an increase in carbon emissions [14,25]. Secondly, when aging reaches a certain stage, older adults will encounter more medical and nursing expenses [26,27]. And it requires more economic activities to support this expenditure, consequently contributing to a further increase in carbon emissions [28]. Thirdly, the intensification of population aging will also lead to a significant increase in expenditure on social security for older adults, bringing a heavy financial burden to the government. In order to alleviate this financial pressure, the government will increase the tax burden on enterprises, which will occupy a portion of R&D investment of enterprises. Thus, the technological progress and innovation capability improvement of enterprises would be limited, which can affect the upgrading of industrial structure, and ultimately promote regional carbon emissions. Overall, resources and energy consumption can be raised by population aging, and regional carbon emissions will be increased as a result. Based on the above analysis, the following hypothesis can be proposed.

H2. 

Population aging can significantly promote regional carbon emissions.

According to the above analysis, the mechanism analysis for the impact of population aging on carbon emissions is shown in Figure 1.

3. Methodology and Variables

3.1. Model Settings

To analyze the mechanism of population aging and regional carbon emissions, this study draws on the idea of Kumbhakar et al. [29] and constructs a bilateral stochastic frontier model as follows.

$${lnCo2}_{it}=i\left(x_{it}\right)+{\omega }_{it}-u_{it}+{\epsilon }_{it}=i\left(x_{it}\right)+{\xi }_{it}=x_{it}\delta +{\xi }_{it}$$

(1)

where ${lnCO2}_{it}$ is carbon emissions; $x_{it}$ is a series of control variables affecting carbon emissions. Specifically, it includes factors such as population density, government fiscal expenditure, industrialization, GDP per capita, etc. δ is a vector of parameters to be estimated; $i\left(x_{it}\right)$ is the frontier carbon emissions; ${\xi }_{it}$ is the composite residual disturbance term; and ${\xi }_{it}={\omega }_{it}-u_{it}+{\epsilon }_{it}$, where ${\epsilon }_{it}$ is a random error term that reflects the deviation of carbon emissions from the frontier caused by unobservable factors. Since the conditional expectation of the composite residual term ${\epsilon }_{it}$ may not be equal to 0, it will lead to biased OLS estimation results. Therefore, ${\omega }_{it}, u_{it}$ are decomposed by Equation (1) to indicate the upward and downward bias effects in the ideal situation, respectively, based on the estimation by the method of the maximum likelihood estimation (MLE). In Equation (1), when ${\omega }_{it}$ ⩾ 0, it indicates that population aging can promote regional carbon emissions; when $u_{it}$ ⩽ 0, it indicates that population aging can inhibit regional carbon emissions. And when $u_{it}⩽0, {\omega }_{it}=0$ or ${\omega }_{it}⩾0, u_{it}=0$, the model is a one-sided stochastic frontier model, meaning that the impact of population aging on carbon emissions is only a one-sided effect. If neither of $u_{it}$ and ${\omega }_{it}$ is zero, it signifies that a bilateral effect of population aging on carbon emissions exists.

According to Equation (1), the actual carbon emissions are ultimately the result of the bilateral effects of population aging. The promotion effect of population aging makes carbon emissions higher than frontier carbon emissions, while the inhibition effect of population aging makes carbon emissions lower than frontier carbon emissions. And the deviation of actual carbon emissions is measured by the net effect, which is the combined function of the promotion and inhibition effect. Moreover, the method of the maximum likelihood estimation (MLE) is utilized to provide valid estimation results while considering the biased results produced via OLS estimation. Furthermore, the following assumptions about the residual distribution are made as follows. The random error term ${\epsilon }_{it}$ obeys a normal distribution with a mean of zero and a variance of ${\sigma }_{\epsilon }^2$, i.e., ${\epsilon }_{it}~iddN(0,{\sigma }_{\epsilon }^2)$. Both of ${\omega }_{it}, u_{it}$ follow the exponential distributions, i.e., ${\omega }_{it}~iddEXP\left({\sigma }_{\omega },{\sigma }_{\omega }^2\right),u_{it}~iddEXP({\sigma }_u,{\sigma }_u^2)$. These error terms should meet the assumption of independence among themselves and exhibit no correlation with the interprovincial characteristic variables. Based on the above assumptions, the probability density function of ${\mathsf{\xi }}_{\mathrm{i}\mathrm{t}}$ is further derived as follows.

$$f\left({\xi }_{it}\right)=\frac{\mathit{exp}\left({\alpha }_{it}\right)}{{\sigma }_u+{\sigma }_{\omega }}\Phi \left({\gamma }_{it}\right)+\frac{\mathit{exp}\left({\beta }_{it}\right)}{{\sigma }_u+{\sigma }_{\omega }}{\int }_{-{\eta }_{it}}^{\infty } \phi \left(x\right)dx=\frac{\mathit{exp}\left({\alpha }_{it}\right)}{{\sigma }_u+{\sigma }_{\omega }}\Phi \left({\gamma }_{it}\right)+\frac{\mathit{exp}\left({\beta }_{it}\right)}{{\sigma }_u+{\sigma }_{\omega }}\phi \left({\eta }_{it}\right)$$

(2)

In Equation (2),$\Phi \left(·\right)$ and $\phi \left(·\right)$ represent the cumulative distribution function (CDF) and the probability density function (PDF) for standard normal distribution, respectively. And the remaining parameters are specified as follows.

$$\begin{array}{ll}{\alpha }_{it}=\frac{{\sigma }_v^2}{2{\sigma }_{\omega }^2}+\frac{{\xi }_{it}}{{\sigma }_{\omega }}& {\beta }_{it}=\frac{{\sigma }_v^2}{2{\sigma }_u^2}-\frac{{\xi }_{it}}{{\sigma }_u}\\ {\gamma }_{it}=-\frac{{\xi }_{it}}{{\sigma }_v}-\frac{{\sigma }_v}{{\sigma }_u}& {\eta }_{it}=\frac{{\xi }_{it}}{{\sigma }_v}-\frac{{\sigma }_v}{{\sigma }_{\omega }}\end{array}$$

(3)

Then, the expression of the maximum likelihood function (MLE), which is based on the parameter estimation of Equation (3), is constructed as follows.

$$lnL\left(X;\pi \right)=-n\mathit{ln}\left({\sigma }_{\omega }+{\sigma }_u\right)+{\sum }_{i=1}^n ln\left[e^{{\alpha }_{it}}\Phi \left({\gamma }_{it}\right)+\left.e^{{\beta }_{it}}\Phi \left({\eta }_{it}\right)\right]\right.$$

(4)

where $\pi =\left[\beta ,{\sigma }_v,{\sigma }_{\omega },{\sigma }_u\right]$. All parameter values for the extreme likelihood estimation are further obtained based on the likelihood function (4). Furthermore, the conditional density functions for ${\omega }_{it}$ and $u_{it}$ are obtained as follows.

$$f\left({\omega }_{it}\mid {\xi }_{it}\right)=\frac{\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)\mathit{exp}\left[-\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right){\omega }_{it}\right]\Phi \left(\frac{{\omega }_{it}}{{\sigma }_v}+{\eta }_{it}\right)}{\mathit{exp}\left({\beta }_{it}-{\alpha }_{it}\right)\left[\Phi \left({\eta }_{it}\right)+\mathit{exp}\left({\alpha }_{it}-{\beta }_{it}\right)\Phi \left({\gamma }_{it}\right)\right]}$$

(5)

$$f\left(u_{it}\mid {\xi }_{it}\right)=\frac{\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)\mathit{exp}\left[-\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)u_{it}\right]\Phi \left(\frac{u_{it}}{{\sigma }_v}+{\eta }_{it}\right)}{\Phi \left({\eta }_{it}\right)+\mathit{exp}\left({\alpha }_{it}-{\beta }_{it}\right)\Phi \left({\gamma }_{it}\right)}$$

(6)

Then, the conditional expectations of ${\omega }_{it}$ and $u_{it}$ can be calculated.

$$E\left({\omega }_{it}\mid {\xi }_{it}\right)=\frac{1}{\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)}+\frac{{\sigma }_v\left[\Phi \left(-{\eta }_{it}\right)+{\eta }_{it}\Phi \left({\eta }_{ii}\right)\right]}{exp\left({\beta }_{it}-{\alpha }_{it}\right)\left[\Phi \left({\eta }_{it}\right)+exp\left({\alpha }_{it}-{\beta }_{it}\right)\Phi \left({\gamma }_{it}\right)\right]}$$

(7)

$$E\left(u_{it}\mid {\xi }_{it}\right)=\frac{1}{\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)}+\frac{exp\left({\alpha }_{it}-{\beta }_{it}\right){\sigma }_v\left[\Phi \left(-{\gamma }_{it}\right)+{\eta }_{it}\Phi \left({\gamma }_{ii}\right)\right]}{\Phi \left({\eta }_{it}\right)+exp\left({\alpha }_{it}-{\beta }_{it}\right)\Phi \left({\gamma }_{it}\right)}$$

(8)

With Equations (7) and (8), the absolute degree of deviation from the frontier level of carbon emissions can be estimated. For a better comparison, it is necessary to further convert the absolute degree value of the deviation of population aging’s effect on carbon emissions into the percentage above or below the frontier level. The conversion formula is then constructed as follows.

$$\begin{array}{c}E\left(1-e^{-{\omega }_{it}}\mid {\xi }_{it}\right)=1-\\ \frac{\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)\left[\Phi \left({\gamma }_{it}\right)+exp\left({\beta }_{it}-{\alpha }_{it}\right)exp\left(\frac{{\sigma }_v^2}{2}-{\sigma }_v{\eta }_{it}\right)\Phi \left({\eta }_{it}-{\sigma }_v\right)\right]}{\left[1+\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)\right]exp\left({\beta }_{it}-{\alpha }_{it}\right)\left[\Phi \left({\eta }_{it}\right)+exp\left({\alpha }_{it}-{\beta }_{it}\right)\Phi \left({\gamma }_{it}\right)\right]}\end{array}$$

(9)

$$E\left(1-e^{-u_{it}}\mid {\xi }_{it}\right)=1-\frac{\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)\left[\Phi \left({\eta }_{it}\right)+exp\left({\alpha }_{it}-{\beta }_{it}\right)exp\left(\frac{{\sigma }_v^2}{2}-{\sigma }_v{\gamma }_{it}\right)\Phi \left({\gamma }_{it}-{\sigma }_v\right)\right]}{\left[1+\left(\frac{1}{{\sigma }_u}+\frac{1}{{\sigma }_{\omega }}\right)\right]\left[\Phi \left({\eta }_{it}\right)+exp\left({\alpha }_{it}-{\beta }_{it}\right)\Phi \left({\gamma }_{it}\right)\right]}$$

(10)

Furthermore, the net effect of population aging on carbon emissions is derived based on Equations (9) and (10), which are calculated as follows:

$$NE=E\left(1-e^{-{\omega }_{it}}\mid {\xi }_{it}\right)-E\left(1-e^{-u_{it}}\mid {\xi }_{it}\right)=E\left(e^{-u_{it}}-e^{-{\omega }_{it}}\mid {\xi }_{it}\right)$$

(11)

In this equation, $NE$ represents the difference between the promotion effect and the inhibition effect. If $NE$ > 0, it indicates that the promotion effect is stronger than the inhibition effect, i.e., the promotion effect plays a dominant role in this case. Conversely, if $NE$ < 0, it indicates that the promotion effect is weaker than the inhibition effect, i.e., the inhibition effect plays a dominant role in this condition.

3.2. Description of Data and Variables

3.2.1. Variable Selection

According to the above theoretical and empirical model settings, this study selects the provincial-level panel data from China from 2011 to 2020 to analyze the impact of population aging on provincial carbon emissions. The variables are selected as follows.

(1) Carbon emissions (lnCO₂). The formula established by the United Nations Intergovernmental Panel on Climate Change (IPCC) is widely utilized for generating GHG emission data. Moreover, according to the research of Kuang et al. [30], considering the greenhouse gas emissions generated by energy consumption on the production side, carbon emissions are estimated based on the consumption of eight types of energy sources, including raw coal, coke, crude oil, gasoline, kerosene, and diesel. And the calculation formula is as follows.

$$CE={\sum }_{i=1}^8 C{O}_{i_{i,jjt}}={\sum }_{i=1}^{\epsilon } M_{ijt}\times K_j$$

(12)

where $t$ is the year; $i$ is the region; $j$ is the energy resource; $M$ is the energy consumption; $CE$ is ${\mathrm{c}\mathrm{o}}_2$ emissions; $K_j$ is the carbon emission factor. Then, the logarithm of the obtained carbon emissions is denoted as lnCO₂.

(2) The level of population aging. Drawing on the research of Shen Ke [31], population aging is measured by the percentage of residents who are people aged 65+. It is denoted as Old in the subsequent analysis.

(3) Inter-provincial characteristics variables. According to the study of Kuang et al. [30], human capital is represented by population density, which is logarithmically transformed and denoted as lnPM; government fiscal expenditure is characterized by general fiscal budget expenditure as a percentage of GDP, which is also logarithmically transformed and denoted as lnGOF; the degree of industrialization is represented by the value added by the secondary sector as a share of GDP, which is logarithmically transformed and denoted as lnInd; the level of economic development is expressed by GDP per capita [32], which is logarithmically transformed and denoted as lnPGdp; drawing on the study by Xu Weixiang et al. [33], urbanization is expressed by using the share of the non-farm population, which is logarithmically transformed and denoted as lnCity; environmental regulations are characterized by the amount of investment in environmental pollution to GDP, which is denoted as EG [34]; the energy structure is measured by the coal consumption to total energy consumption, which is denoted as EnS [35]; energy intensity is represented by energy consumption per unit of GDP [36], which is denoted as EQ. The results of the descriptive statistics for the main variables are shown in

Table 1. Descriptive statistics of main variables.

Variable	Symbol	Sample Size	Mean	Covariance	Minimum	Maximum
Carbon emissions	lnCO2	300	10.430	0.741	8.494	12.184
Population aging	Old	300	10.519	2.391	5.473	17.736
Population density	lnPM	300	7.892	0.410	6.639	8.710
Government financial spending	lnGOF	300	3.147	0.376	2.400	4.160
Degree of industrialization	lnInd	300	3.064	0.409	1.678	4.019
GDP per capita	lnPGdp	300	10.841	0.436	9.706	12.013
Urbanization	lnCity	300	4.046	0.199	3.555	4.495
Environmental regulation	EG	300	0.049	0.090	0.016	0.767
Energy structure	EnS	300	0.387	0.148	0.008	0.687
Energy intensity	EQ	300	0.825	0.485	0.207	2.327

.

3.2.2. Data Source

Considering the availability of data, Tibet, Hong Kong, Macao, and Taiwan are not included. The data used in this research are sourced from the China Statistical Yearbook (2012–2021), China Science and Technology Statistical Yearbook (2012–2021), as well as the ESP Global Database. In addition, variables involving price factors are deflated using 2011 as the base period. And according to the geographical location, all provinces are classified into three regions, which is the east, the middle, and the west. The eastern region includes 11 provinces or cities, which are Hebei, Liaoning, Fujian, Shandong, Jiangsu, Zhejiang, Guangdong, Hainan, Beijing, Tianjin, and Shanghai. The central region includes 8 provinces or cities, which are Heilongjiang, Jilin, Shanxi, Hubei, Hunan, Anhui, Jiangxi, and Henan. In addition, the western region includes 11 provinces or cities, which are Sichuan, Yunnan, Inner Mongolia, Ningxia, Guangxi, Xinjiang, Gansu, Guizhou, Chongqing Shaanxi, and Qinghai.

4. Empirical Analysis

4.1. Estimation of Bilateral Stochastic Frontier Model

4.1.1. Estimation of Baseline Regression Models

Based on the MLE, the bilateral effects of population aging on carbon emissions are decomposed. The results of the baseline regression models are shown in

Table 2. Basic estimation results of bilateral stochastic frontier model for population aging.

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)	Model (7)
	lnCo2	lnCo2	lnCo2	lnCo2	lnCo2	lnCo2	lnCo2
lnPM	−0.392 **	−0.059 ***	0.024 ***	−0.062 ***	−0.256 ***	−0.050 ***	−0.240 ***
	(−2.33)	(−7.96)	(34.19)	(−131.47)	(−411.96)	(−187.45)	(−634.67)
lnGOF	−0.112	0.389 ***	0.029 ***	0.015 ***	0.077 ***	0.005 ***	0.073 ***
	(−0.82)	(63.65)	(58.42)	(43.26)	(103.12)	(21.99)	(244.02)
lnInd	0.670 **	0.760 ***	0.041 ***	0.392 ***	0.423 ***	0.232 ***	0.323 ***
	(2.59)	(298.20)	(25.16)	(335.06)	(142.91)	(464.49)	(535.53)
lnPGdp	−0.072	0.053 ***	0.017 ***	0.010 ***	0.026 ***	0.011 ***	0.037 ***
	(−0.74)	(6.31)	(55.61)	(27.58)	(38.41)	(72.69)	(344.05)
lnCity	1.592 *	1.019 ***	1.094 ***	1.170 ***	−0.013 *	0.946 ***	0.217 ***
	(1.84)	(103.70)	(248.85)	(579.18)	(−1.94)	(697.32)	(60.98)
EG	−0.395	−1.418 ***	−0.117 ***	−0.459 ***	−0.497 ***	−0.279 ***	−0.419 ***
	(−0.82)	(−116.56)	(−88.62)	(−755.16)	(−510.72)	(−257.95)	(−949.21)
EnS	1.835 **	1.830 ***	0.500 ***	0.748 ***	0.237 ***	0.911 ***	0.336 ***
	(2.34)	(69.88)	(57.22)	(333.28)	(104.08)	(306.64)	(93.98)
EQ	−0.358	0.051 ***	0.153 ***	0.538 ***	0.389 ***	0.254 ***	0.433 ***
	(−1.08)	(7.00)	(82.97)	(356.96)	(493.75)	(267.75)	(939.11)
_cons	5.172	1.945 ***	4.425 ***	2.708 ***	8.807 ***	4.612 ***	7.871 ***
	(1.09)	(15.40)	(260.33)	(261.78)	(308.51)	(754.86)	(453.74)
_cons		−32.030	−17.090	−17.325	−17.100	−17.413	−17.655
		(−0.01)	(−0.03)	(−0.04)	(−0.06)	(−0.04)	(−0.05)
Old					−0.340 ***		−0.363 ***
					(−12.41)		(−12.73)
_cons			−1.713 ***	−2.001 ***	−5.369 ***	−1.690 ***	−5.667 ***
			(−23.93)	(−25.40)	(−17.42)	(−24.19)	(−17.98)
Old						0.131 ***	0.183 ***
						(4.11)	(5.18)
_cons			−1.641 ***	−1.456 ***	−2.261 ***	−3.160 ***	−4.064 ***
			(−23.48)	(−22.21)	(−29.10)	(−8.99)	(−11.13)
Pro fixed	No	No	Yes	Yes	Yes	Yes	Yes
Year fxed	No	No	No	Yes	Yes	Yes	Yes
$R^2$	0.757
N	300	300	300	300	300	300	300

Note: ***, **, and * are tabulated as passing the test at 1%, 5%, and 10% significance levels, respectively; the corresponding Z statistics are given in parentheses.

. In model (1), the OLS estimation is adopted without considering the bias effect; in model (2), neither the time-fixed effect nor the area-fixed effect is controlled; in model (3), only the area-fixed effect is controlled; while in model (4), both the area-fixed effect and the time-fixed effect are controlled. In addition, model (5) only focuses on the one-sided estimation of the inhibition effect of population aging, while model (6) concentrates on the one-sided estimation of the promotion effect; and in model (7), both the inhibition and promotion effect of population aging are considered, judging by the model residual terms ${\mathsf{\omega }}_{\mathrm{i}\mathrm{t}}$ and ${\mathrm{u}}_{\mathrm{i}\mathrm{t}}$. According to the model likelihood ratio test (LR), model (7) is more reasonable compared with the others. Thereby, model (7) is chosen as the basis for the subsequent analysis of the bilateral effect of population aging.

From the estimated results of model (7), it is obvious that the estimated coefficient for the promotion effect of population aging is 0.183, which is significantly positive, indicating that the promotion effect of population aging leads to an increase in carbon emissions; while the estimated coefficient of the inhibition effect of population aging is 0.363, which is significantly negative, indicating that the inhibition effect of population aging significantly suppresses the increase in carbon emissions. Thus, the above results verify both hypothesis 1 and 2 of this study, namely, the existence of bilateral effects of population aging on regional carbon emissions. Specifically, the promotion and inhibition effects.

4.1.2. Covariance Decomposition: Measuring the Bilateral Effect of Population Aging on Carbon Emissions

In order to fully analyze which of the two effects of population aging on carbon emissions dominates, it is required to decompose the promotion and inhibition effects of population aging on carbon emissions on the basis of model (7). And the results of the decomposition are presented in

Table 3. Variance decomposition of the effects of population aging on carbon emissions.

	Variable Meaning	Symbols	Measurement Coefficient
Population aging impact	Random error term	${\sigma }_v$	0.0000
	Promotion effect	${\sigma }_{\omega }$	0.1301
	Inhibition effect	${\sigma }_u$	0.2410
	Total random error term	${\sigma }_u^2$+${\sigma }_w^2$+${\sigma }_v^2$	0.0750
Variance decomposition	The weight of the two effects	$\frac{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2}{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2+{\mathsf{\sigma }}_v^2}$	1.0000
	Weight of promotion effect	$\frac{{\mathsf{\sigma }}_w^2}{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2}$	0.2258
	Weight of inhibition effect	$\frac{{\mathsf{\sigma }}_u^2}{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2}$	0.7742

. It can be seen that the positive and negative effects of population aging on carbon emissions are quantified as 0.1301 and 0.2410, respectively. And the degree of the net effect of population aging on carbon emissions is E ($\omega -u$) = −${\mathsf{\sigma }}_{\mathrm{u}}$ = 0.1109. The decomposition results demonstrate that the net effect of population aging on carbon emissions inhibits the generation of carbon emissions. In summary, population aging has both positive and negative effects on carbon emissions. However, the inhibition effect of population aging outweighs the promotion effect, which eventually leads to the actual regional carbon emissions being lower than the optimal level, i.e., population aging inhibits the generation of regional carbon emissions.

Furthermore, the weight of both the inhibition and promotion effects of population aging on carbon emissions is estimated, and the results are shown in Table 3. It can be seen that the population aging effects inhibit the generation of carbon emissions by 77.42%, whereas the promotion effect promotes carbon emissions by 22.58%. This result indicates that population aging has a significantly larger inhibition effect than the promotion effect. This finding verifies the accuracy of the aforementioned estimation results, i.e., eventually, population aging inhibits the growth of carbon emissions in regions.

4.1.3. The Bilateral Effects of Population Aging on Carbon Emissions

After analyzing the effect of population aging on carbon emissions, we proceeded to calculate the deviation of regional carbon emissions from the frontier level, based on the Equations (8)–(10). The estimation results are shown in

Table 4. Estimated bilateral effects of population aging on carbon emissions (%).

Variable	Mean	Variance	p25	p50	p75
Promotion effect	10.63	11.19	4.97	7.42	11.73
Inhibition effect	15.77	20.72	4.64	8.27	14.08
Net effect	−5.14	22.21	−5.74	−0.44	1.56

. It can be seen that the promotion effect of population aging makes carbon emissions higher than the frontier level by 10.63%, whereas the inhibition effect of population aging makes carbon emissions lower than the frontier level by 15.77%. With the function of both effects, the actual carbon emissions are eventually 5.14% lower than the frontier level, which indicates that the asymmetry of the bilateral effect of population aging makes a suppressive effect of population aging on carbon emissions. Specifically, under the p25 and p50 quantiles, population aging reduces carbon emissions by 5.74% and 0.44%, indicating that the inhibition effect of population aging on carbon emissions counterbalances its promotion effect. Thus, this finding supports the conclusion that population aging inhibits carbon emissions [37,38]. However, under the p75 quantile, the effect of population aging on carbon emissions reverses. The promotion effect significantly exceeds the inhibition effect, indicating that the promotion effect of population aging on carbon emissions dominates at higher levels of population aging, which means population aging increases carbon emissions [39]. A possible reason for this phenomenon may be that older adults tend to use traditional fossil fuels more than the youth, which means that as the level of aging increases, there is a greater consumption of energy and resources. In addition, factors such as transportation, household electricity consumption, and medical care of the aging population can also lead to an increase in regional carbon emissions.

In order to visualize the distribution of the promotion effect, inhibition effect, and net effect of population aging, Figure 2 is shown to illustrate the frequency distributions of the above effects. As per the trend demonstrated in this figure, the inhibition effect of population aging on carbon emissions shows a right-trailed feature, and it still exists on the right side of 90%, indicating that the carbon emissions of some provinces are more sensitive to negative changes in population aging. While the promotion effect of population aging on carbon emissions ends at approximately 80%, significantly lower than the inhibition effect, indicating that the carbon emissions of some provinces are less influenced by the promotion effect of population aging. The distribution of the net effect also reveals that, due to the composite function of bilateral effects, most of the regions are influenced by the inhibition effect of population aging.

4.2. Regional Characteristics of Population Aging Impacts on CO₂ Emissions

The distribution characteristics of the net effect of population aging on carbon emissions in different provinces and regions are further examined, and the results are shown in

Table 5. Characteristics of the annual distribution of the net effect of population aging on carbon emissions (%).

Province	Net Effect Average	Province	Net Effect Average	Province	Net Effect Average
Hebei	−12.46	Heilongjiang	−7.37	Sichuan	−13.10
Liaoning	−17.29	Jilin	0.42	Yunnan	−1.70
Fujian	1.06	Shanxi	−8.95	Inner Mongolia	−10.46
Shandong	−8.78	Hubei	−8.50	Ningxia	−1.59
Jiangsu	−13.69	Hunan	−7.37	Guangxi	−4.27
Zhejiang	−15.18	Anhui	−12.73	Xinjiang	4.07
Guangdong	−13.19	Jiangxi	0.35	Gansu	11.97
Hainan	8.93	Henan	−3.01	Guizhou	−0.31
Beijing	−16.38			Chongqing	−13.28
Tianjin	9.78			Shaanxi	−16.61
Shanghai	1.27			Qinghai	14.20
Eastern Region	−7.03	Central Region	−5.9	Western Region	−2.83

. In terms of regional distribution, the net effects of population aging on carbon emissions in the eastern, western, and central regions are all negative, with values of −7.03%, −5.90%, and −2.83%, respectively, indicating that population aging in all three regions significantly decreases carbon emissions. Specifically, the inhibition effect of population aging on carbon emissions is the greatest in the eastern region, second in the central region, and the least in the western region. The main reason is that although population aging in the eastern region is greatly serious, the higher levels of local industrial structure, economic development, and science and technology, as well as better social security, result in a stronger inhibition effect of population aging on carbon emissions. Moreover, the central and western regions are continuously affected by the “siphon effect” of human capital and industrial transfer from the eastern region; thereby, the development of labor productivity and industrial transformation as well as regional upgrading can be restricted. Thus, it leads to a lower inhibition effect of population aging on carbon emissions. Overall, the spatial pattern of population aging on regional carbon emissions is characterized as a trend of decreasing spatial distribution in the eastern, central, and western regions.

4.3. Temporal Characteristics of the Impacts of Population Aging on Carbon Emissions

Further estimates of the temporal trend of the population aging effect on carbon emissions are presented in Figure 3, which shows that the inhibition effect of population aging dominates in the majority of the sample years, ranging from −0.17% to −22.13%. Overall, the net effect of population aging on carbon emissions changes from positive to negative. Meanwhile, the inhibition effect of population aging on carbon emissions gradually dominates and enhances over time.

The reason may be that public awareness of energy conservation and environmental protection has gradually increased, and green energy and technology has also been widely accepted, resulting in lower energy consumption and carbon emissions. In addition, the development of population aging has further pushed the enhancement of human capital, upgrading of industrial structure, and level of technological innovation, which indirectly reduce regional carbon emissions eventually.

4.4. Analysis of Differences in the Impact of Various Population Aging Levels

The distribution of the bilateral effects under different population aging development levels is analyzed by dividing the population aging development levels into low, medium, and high groups according to 25%, 50%, and 75% quantiles, and the results are shown in

Table 6. Differences in the bilateral effects at various population levels (%).

Old	Effect Decomposition	Mean	SD	p25	p50	p75
Old ≤ 8.789	Promotion effect	6.72	8.26	3.02	3.80	8.28
	Inhibition effect	5.47	8.53	2.74	3.52	5.69
	Net effect	1.25	12.43	−2.50	0.00	5.09
8.789 < Old ≤ 12.051	Promotion effect	9.36	11.16	5.35	6.75	8.16
	Inhibition effect	12.40	15.61	5.22	7.41	11.83
	Net effect	−3.04	19.88	−5.16	0.00	1.08
Old > 12.051	Promotion effect	16.99	11.18	11.11	13.62	17.79
	Inhibition effect	32.91	27.23	12.39	18.99	50.65
	Net effect	−15.92	29.55	−37.29	−5.09	0.00

.

As the population aging development level increases, the mean value of the promotion effect of population aging on carbon emissions rises from 6.72% for population aging (Old) ≤ 8.789 to 16.99% for population aging (Old) > 12.051, respectively. And the mean value of its inhibition effect increases from 5.47% for population aging (Old) ≤ 8.789 to 32.91% for population aging (Old) >12.051. The combined function of these two kinds of effects is that the mean value of the net effect turns from positive to negative, indicating that the inhibition effect of population aging on carbon emissions gradually takes the dominant position as population aging develops, thus confirming again that regional population aging has a significant inhibition effect on regional carbon emissions.

4.5. Analysis of the Impacts at Different Levels of Human Capital

Human capital is one of the important factors for regional economic and social development. When regional human capital is large enough, its industrial structure and population structure will undergo corresponding changes and give birth to the fluctuation of carbon emissions. Herein, the average years of education were selected to describe human capital. And the formula for the average years of education is as follows: average years of education per capita = the percentage of the population with primary school education*6 + the percentage of the population with middle school education × 9 + the percentage of the population with high school education×12 + the percentage of the population with college and above education ×16. And it is denoted as EDU in this part. Human capital is then divided into three groups with quantiles of 25%, 50%, and 75%. The results are shown in

Table 7. Differences in the bilateral effects under various human capital levels (%).

EDU	Effect Decomposition	Mean	SD	p25	p50	p75
EDU ≤ 8.725	Promotion effect	10.45	13.37	4.49	6.01	11.66
	Inhibition effect	10.37	12.04	4.20	7.27	11.13
	Net effect	0.09	17.24	−5.09	0.00	1.37
8.725 < EDU ≤ 9.485	Promotion effect	9.44	9.03	4.83	7.12	10.94
	Inhibition effect	14.45	19.45	4.77	8.17	13.25
	Net effect	−5.01	19.20	−5.07	−0.51	1.19
EDU > 9.485	Promotion effect	13.10	12.32	6.41	8.43	14.44
	Inhibition effect	23.91	26.95	5.47	11.37	26.75
	Net effect	−10.82	29.86	−18.46	−1.06	2.02

. When EDU ≤ 8.725, the net effect of population aging is 0.09%; when 8.725 < EDU ≤ 9.485, the net effect of population aging on carbon emissions is −5.01%; when EDU > 9.485, the net effect of population aging on carbon emissions is −10.82%. In summary, it can be seen that the net effect of population aging on carbon emissions turns from positive to negative as human capital grows, thus verifying that human capital can exacerbate the inhibition effect of population aging on carbon emissions, which is largely in line with the conclusion of Guo Feng et al. [40]. The possible reasons may be, compared to physical investment, human capital is a cleaner production factor that can provide cleaner technological choices for the development of countries, industries, and enterprises. Therefore, the agglomeration effect of human capital can to some extent buffer the negative effects of the energy rebound caused by population aging.

4.6. Analysis of the Impacts at Different Levels of Urbanization

Urbanization is another important factor for regional economic and social development. The improvement in regional urbanization will bring about changes in the population structure and industrial structure, leading to a certain impact on the carbon emission effect of population aging. Accordingly, we choose the proportion of the non-agricultural population to characterize urbanization, which is denoted as city in this part.

And similarly, we divide urbanization into three groups with quantiles of 25%, 50%, and 75%. The results are presented in

Table 8. Differences in the impact of population aging on carbon emissions under different urbanization (%).

	Effect Decomposition	Mean	SD	p25	p50	p75
city ≤ 49.97%	Promotion effect	7.48	9.26	4.19	5.15	8.04
	Inhibition effect	7.58	5.24	4.05	5.78	9.64
	Net effect	−0.10	10.44	−4.69	0.00	0.71
49.97% < city ≤ 64.09%	Promotion effect	11.56	12.04	5.54	7.90	13.11
	Inhibition effect	16.02	20.20	4.83	8.21	14.64
	Net effect	−4.46	22.93	−5.16	−0.21	2.26
city > 64.09%	Promotion effect	11.83	10.63	6.22	9.07	13.93
	Inhibition effect	23.56	27.46	5.69	11.05	24.78
	Net effect	−11.73	27.45	−10.73	−1.83	0.58

. When city ≤ 49.97%, the net effect of population aging on carbon emissions is −0.10%; when 49.97% < city ≤ 64.09%, the net effect of population aging is −4.46%; when city > 64.09%, the net effect of population aging on carbon emissions is −11.73%. It demonstrates that the inhibition effect of population aging on carbon emissions steadily increases as urbanization is advanced, which indicates that urbanization can enhance the inhibition effect of population aging on carbon emissions. The possible reasons for this result may be, with the improvement in regional urbanization, regional economic resources such as labor, science and technology, and capital are accordingly introduced into regions, promoting the transformation of regional production and lifestyle towards an intensive low-carbon direction. At the same time, the gradual establishment of environmental governance regulations and industrial development policies have effectively reduced the carbon emissions contributed by regional aging.

4.7. Robustness Test

To verify the robustness of the results, we choose to replace variables and change the research period, respectively, in this section. Referring to the research of Shen and Li (2021) [31], the old-age dependency ratio is chosen to characterize population aging first to observe the reliability of the above results. And then, considering the influence of COVID-19 around the world, data in 2019 to 2020 are excluded in another measurement in this section. On this basis, the promotion effect, inhibition effect, and net effect of population aging on carbon emissions are evaluated again. And the results are shown in

Table 9. Robustness tests for the effects and variance decomposition.

	Variance Meaning	Symbols	Measurement Coefficient (1)	Measurement Coefficient (2)
Population aging impact	Random error term	${\sigma }_v$	0.0000	0.0000
	Promotion effect	${\sigma }_{\omega }$	0.1693	0.2658
	Inhibition effect	${\sigma }_u$	0.2049	0.4528
	Random total error term	${\sigma }_u^2$+${\sigma }_w^2$+${\sigma }_v^2$	0.0707	0.2757
	The weight of the two effects	$\frac{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2}{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2+{\mathsf{\sigma }}_v^2}$	1.0000	1.0000
Variance decomposition	Promotion effect weight	$\frac{{\mathsf{\sigma }}_w^2}{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2}$	0.4056	0.2563
	Inhibition effect weight	$\frac{{\mathsf{\sigma }}_u^2}{{\mathsf{\sigma }}_u^2+{\mathsf{\sigma }}_w^2}$	0.5944	0.7437

. It is evident that the promotion effect of population aging on regional carbon emissions is 0.1693 and 0.2658, and the inhibition effect is 0.2049 and 0.4528, respectively. It indicates that there is a bilateral effect of population aging on regional carbon emissions, which is consistent with the previous results. In terms of the effect weight, the promotion effect of population aging accounts for 40.6% and 25.6%, and the inhibition effect accounts for 59.4% and 74.4%, respectively, implying that the inhibition effect of population aging always dominates in the above circumstances.

In addition, the deviations of the inhibition effect, the promotion effect, and the net effect of population aging on regional carbon emissions are further estimated, which are shown in

Table 10. The bilateral effects of population aging on regional carbon emission (%).

Variable	Mean (1)	Covariance (1)	Mean (2)	Covariance (2)
Promotion effect	12.93	13.78	20.96	13.12
Inhibition effect	14.17	18.86	29.92	21.29
Net effect	−1.24	21.52	−8.96	8.17

. The findings show that, under these two kinds of circumstances, the promotion effect of population aging raises regional carbon emissions by 12.93% and 20.96%, while the inhibition effect decreases regional carbon emissions by 14.17% and 29.92, respectively. As a result of the net effect, regional carbon emissions are relatively lower than the frontier level by 1.24% and 8.96%, which is mainly the same as the previous conclusions. Overall, the robustness of the results can be further verified.

5. Conclusions and Policy Implications

5.1. Conclusions

To clarify the relationship between population aging and carbon emissions in regions, this study applies a bilateral stochastic frontier model to measure and analyze the inhibition, promotion, and net effects of population aging on regional carbon emissions, using panel data from 30 Chinese provinces from 2011 to 2020. And the following conclusions are obtained: (1) The promotion effect of population aging increases carbon emissions over the frontier level by 10.63%, while the inhibition effect reduces carbon emissions by 15.77%. And the net effect of population aging on carbon emissions is −5.14%, indicating that regional carbon emissions are greatly decreased by population aging. (2) In terms of regional distribution characteristics, population aging in eastern, western, and central regions significantly reduces the regional carbon emission level, and the inhibition effect of population aging on carbon emissions in the eastern region is stronger than that in the central and western regions. (3) In terms of temporal distribution characteristics, the inhibition effect of population aging on carbon emissions increases continuously, and gradually holds the dominant position during the study period. (4) With the growth of population aging, human capital, and urbanization, the inhibition effect of population aging on regional carbon emissions can be strengthened.

5.2. Policy Implications

Based on the above analysis, the policy recommendations are proposed as follows.

Firstly, the problem of reducing carbon emissions should be considered in terms of population. The human production and consumption behaviors have the determinant function on economic development, the industrial structure, and technological innovation; thus, it makes more sense to focus on the population while reducing carbon emissions. As aging has gradually become a trend in the development of the world’s population in recent years, it is necessary to pay attention to the important role of demographic changes on regional carbon emissions. Appropriate policies should be created to fully exploit the impact of population aging on carbon reduction. To alleviate the pressure brought about by population aging as much as possible, policies can be focused on the features of the arrangements and adjustments of social structures and socioeconomic institutions for older adults. Moreover, promoting cooperation between the young and older adults is a good alternative.

Secondly, heterogeneous governance strategies based on regional development differences should be implemented. Nowadays, there still exists the relative backwardness condition of older adults’ social security coverage in the central and western regions of China. Moreover, older adults usually have much lower income levels, pension insurance coverage rates, and basic medical insurance participation rates. Therefore, governments should strengthen their support for older adults in the central and western regions and give appropriate policy welfare to them in these districts. For instance, establish the minimum income guarantee system and purchase a certain amount of community services to improve the income level and life quality of old adults, thus achieving the goal of reducing carbon emissions.

Thirdly, it is necessary to increase investments in human capital through the way of improving the quality and quantity of labor resources of older adults. In order to increase older adults’ options for education and training, a flexible retirement age can be chosen to realize the growth of life expectancy. And some relevant regulations on the employment of them can be developed to encourage enterprises to eliminate age discrimination and intergenerational conflicts in the employment of older adults, thus reducing the crowding-out effect of population aging on human capital investment.

Moreover, low-carbon consumption concepts and behaviors should be encouraged in the whole society. Therefore, governments should give correct guidance to citizens about the consumption patterns of social groups, including how to lessen their need for high-energy-consuming products and unnecessary daily expenses. At the same time, governments must put forward their efforts to promote energy conservation and environmental protection, encourage more environmentally friendly and low-carbon modes of transportation, and accelerate the formation of green and low-carbon lifestyles and consumption patterns for all the regions in China.

5.3. Research Limitation and Future Work

This study concentrates upon the bilateral effects of population aging, providing a theoretical reference to facilitate regional green and sustainable development. Nevertheless, there still exist certain limitations which demand to be enhanced further. For one thing, on account of data availability, the indicators we chose cannot accurately cover the various aspects of regional population aging, such as aging size, aging structure, and aging quality. Therefore, a more complicated evaluation system is preferred to enhance the practical significance of relative research. For another thing, this study focuses on the different dimensions of the effects and fails to capture various and specific paths to maintain the positive function of population aging, which will be the expanded orientation of the follow-up work.