A Study of the Impact of Population Aging on Fiscal Sustainability in China

Abstract

China has entered a deeply aging society, and the aging population poses a significant public risk to fiscal sustainability. In this regard, researchers have conducted a large number of studies, but the fiscal sustainability indicators used in the existing literature are not scientific enough, the sample data are too macro, and the heterogeneity analysis is not comprehensive enough. This paper innovatively constructs fiscal sustainability indicators based on data from 4 municipalities directly under the central government, 8 provincial capitals, and 88 prefecture-level cities in China from 2010–2019, and analyzes the impact of population aging on fiscal sustainability in eastern, central, western, and multi-level cities in China, using methods such as two-way fixed-effects models. The study finds that (1) fiscal sustainability is significantly hampered by population aging; that is, the more aging there is, the less fiscal sustainability there is. (2) The inhibitory effect of population aging on fiscal sustainability is greater in developed regions compared to backward regions. Compared to prefecture-level cities, provincial cities (including municipalities and provincial capitals) are much more negatively impacted by population aging on fiscal sustainability. (3) The paths through which population aging inhibits fiscal sustainability are healthcare expenditures and social security employment expenditures. The policy recommendations put forward in this paper are to raise the fertility rate, protect the fiscal expenditures of developed regions and provincial capitals to deal with population aging, and increase the effectiveness of the use of funds for medical and health expenditures and social security employment expenditures. The conclusions and policy recommendations drawn in this paper have a positive effect on China’s response to the fiscal sustainability problems caused by an aging population.

1. Introduction

On 17 January 2023, the latest statistics released by China’s National Bureau of Statistics showed that China’s total population at the end of the year 2022 decreased by 850,000 compared to the total population at the end of the previous year, China’s annual births in 2022 were 9.56 million with a birth rate of 6.77%, and the annual deaths reached 10.41 million with a population mortality rate of 7.37%. The natural population growth rate was −0.6%, which is the first time in the past 61 years that China’s population has experienced negative growth. China has officially entered a deeply aging population. With the rapid development of the global economy and society, population aging, as a common issue and problem worldwide, is inevitable in any country and region (Zhou J.) [1]. Due to the high economic growth and continuous social progress, the population’s average life expectancy will dramatically rise, which will also cut fertility rates, which will both accelerate population aging and have a greater negative influence on the economy and society. Clements B et al. [2] predicted that the percentage of people aged 65 and older in the world will increase from the present 12% to 38% by the end of this century, and in most countries and regions, the population will also show a declining trend after reaching an inflection point. The continual slowing of the economic growth rate is the most obvious manifestation of population aging’s effects on economic growth. The impact of population aging on finance is primarily manifested in the change in the scale and proportion of government fiscal expenditure and the increase in the level of fiscal burden in old age. In this case, if the birth rate continues to be low and the economic growth rate slows down, the fiscal burden is bound to further increase, and fiscal sustainability faces a major challenge or even becomes unsustainable (X J Wang and W D Ren) [3]. The focus of this paper is on the impact of population aging on fiscal sustainability, and based on the observability of the sample data, this paper draws on the Aristovnik [4] fiscal burden indicator ($\frac{\mathrm{FinExp}-\mathrm{FinRev}}{\mathrm{GDP}}$) and innovatively proposes to express fiscal sustainability as ($1-\frac{\mathrm{FinExp}-\mathrm{FinRev}}{\mathrm{GDP}}$), where the higher the fiscal burden, the lower the level of fiscal sustainability. In China, Liu Z et al. [5] argued that due to the long-standing population policy of family planning and the economic policy of reform and opening up, the birth rate has shown a decreasing trend year by year and the population’s average life expectancy has increased, and China’s population aging process will be further accelerated by such a decrease and increase.

According to the National Bureau of Statistics of China, for the decade from 2010 to 2019, the total population of China grew from 1340.9 million to 1412.1 million, an increase of 5.3%; the elderly population grew from 118.9 million to 190.6 million, an increase of 60.3%; and the elderly dependency ratio increased from 11.9% in 2010 to 17.8% in 2019. Among the many factors affecting population aging, the most direct factor is the double decrease in population mortality and birth rate (Bucher S). [6]. Although China attaches great importance to the issue of population aging and has taken a series of policy measures to protect it, the problem of population aging still shows an aggravating trend, which to a certain extent reduces the financial sustainability of China.

The innovations of this paper are as follows: First, innovative construction of fiscal sustainability indicators based on fiscal burden indicators; second, collecting and using panel data of prefecture-level cities for analysis; and third, verifying the heterogeneity of the impact of population aging on fiscal sustainability.

The following chapters are arranged as follows: The second section presents a literature review and theoretical hypothesis, the third section describes the research design, variable description, and data source, the fourth section presents empirical analysis, and the last section contains the conclusion, policy recommendations, and future research.

2. Literature Review and Theoretical Hypothesis

2.1. The Impact of Population Aging on Fiscal Sustainability

Due to the increasing prominence of population aging, there are many studies by domestic and international scholars on the impact of population aging on fiscal sustainability, but scholars have reached different conclusions due to the different research subjects and research methods. Some academics think that the financial effects of population aging are insignificant. Narayana M R. [7] analyzed the impact of population aging on public finance in India for the period of 2005–2050 by constructing a budget forecasting model, and the results showed that the impact of population aging on fiscal sustainability in India is manageable. Azolibe C B et al. [8] showed that the impact of population aging on fiscal sustainability in Africa is insignificant. Dolls M et al. [9] similarly argued that in the economically developed EU region, this change in the age distribution of the population may not have as severe of an impact on fiscal expenditure pressures as currently thought.

However, mainstream research has concluded that population aging significantly increases the fiscal burden and leads to fiscal unsustainability. Jacobsen R H and Jensen S E H [10] found that lower birth rates and higher average life expectancy not only exacerbate population aging but also affect the family structure, and both the family structure and fiscal sustainability are negatively impacted by the aging of the population, but this impact is more severe than that of the family structure. Van E C et al. [11] studied the impact of population aging on public finances in the Netherlands and found that the cost of social pensions and healthcare for an aging population will rise in response to an increase in the old age dependency ratio, outpacing any gain in tax receipts, ultimately breaking the balance between future public spending and tax revenues. Cho S and Kim J R [12] indicated that in Korea, the aging of the population has a two-way impact on the economy and society, with the government facing an expansion in fiscal spending along with a contraction in government revenues, in which case the government’s fiscal deficit increases year by year, and the impact will be long-lasting. From a long-term perspective, Svaljek S. [13] argues that population aging will eventually cause fiscal unsustainability and that only active policies that guarantee the rate of economic growth can deal with the adverse effects of population aging. Liu B and Yang Z [14] argued that China’s fiscal sustainability is facing a huge challenge with the accelerated aging of its population, and Li Wang et al. [15] went further and found that the pension shortfall had exceeded $2 trillion by 2020, and without active countermeasures, a future fiscal crisis is inevitable.

As can be seen, a portion of the literature argues that population aging does not have a serious negative impact on fiscal sustainability, while mainstream Chinese and foreign literature studies show that population aging significantly worsens fiscal sustainability in Denmark, the Netherlands, Korea, and China. Synthesizing the intersection of the two views, this paper proposes that:

Hypothesis 1. 

Fiscal sustainability is significantly hampered by population aging; that is, the more aging there is, the less fiscal sustainability there is.

2.2. Heterogeneity in the Impact of Population Aging on Fiscal Sustainability across Regions

The problem of population aging is widespread and occurs not only in economically developed countries and regions but also in less developed regions, as found by Kubanová J and Linda B [16]. Wu Y et al. [17] argue that the aging population has an impact on fiscal sustainability, it is heterogeneous across countries and regions with different levels of economic development, and the level of impact of population aging on fiscal sustainability is closely related to the level of economic development of the region. Some scholars argue that the impact of the aging population on fiscal sustainability is smaller in economically developed regions. Carchano M et al. [18] argued that a region’s status of economic development is positively related to fiscal sustainability, and economically developed regions have higher levels of fiscal sustainability and are better able to cope with the fiscal sustainability problems caused by population aging compared to economically backward regions.

However, mainstream academics think that the negative financial effects of population aging are greater in regions with higher levels of economic development. Azolibe C B et al. [8] found that the impact of the aging population on public finance expenditure is not significant in Africa. Li J et al. [19] found that by measuring the global aging rate, the coefficient of variation of the global aging rate increased by 0.15 over the 58-year period from 1960 to 2017, and the annual growth rate of the elderly (65 years and older) also showed significant differences, with the economically developed European region reaching the highest at 0.153%, while the relatively less-developed African region only reached 0.0069%, which shows that the growth rate of the elderly population in developed regions is relatively high and the negative impact of the aging population on the fiscal sustainability of economically developed regions is greater. According to the data published by the NBC of China, the economic and social development level of eastern China is somewhat higher than that of central and western China, and provincial cities are somewhat higher than prefecture-level cities.

The above-mentioned studies show that population aging has different fiscal sustainability implications for reasons that emphasize, on the one hand, that developed regions have stronger financial resources to deal with population aging and, on the other hand, that population aging is more severe in developed regions. From China’s perspective, compared to China’s less-developed regions, China’s developed regions have certain financial resources to deal with population aging, but China is a developing country with limited financial resources in general, while population aging is more serious in all regions, that is, compared to China’s less-developed regions, China’s developed regions have less ability to deal with population aging. Based on the above studies and facts, this paper proposes that:

Hypothesis 2. 

The inhibitory effect of population aging on fiscal sustainability is greater in developed regions compared to less-developed regions; the inhibitory effect of population aging on fiscal sustainability is greater in provincial cities compared to prefecture-level cities.

2.3. The Path of the Role of Population Aging on Fiscal Sustainability

Population aging is affecting fiscal sustainability through different paths. Echevarría C A. [20] found that the composition and proportion of fiscal expenditures are influenced by changes in population size and age structure. As to how these expenditure items affect fiscal sustainability, academics have different views. Okma K and Gusmano M K [21] argued that although the composition and proportion of fiscal expenditures need to be adjusted regularly due to the impact of population aging, the rate of population aging and health spending are not directly correlated.

However, as a mainstream view, Malačič J [22] argues that the adjustment of expenditure on public finances due to population aging is primarily in the area of public pensions, healthcare, and long-term care. Jimeno J F et al. [23] went on to argue that the increase in demand for public healthcare services by the aging population is objective, and the government, in response to a social problem such as population aging, will passively shift its focus to less productive or even public welfare pension programs, resulting in fiscal expenditures on pensions being greater than fiscal revenues and increasing the fiscal burden on local governments. In China, due to the long-standing pension and healthcare policies, social welfare fiscal expenditures such as pensions and healthcare have increased year by year as China’s population ages, as studied by Fu BY and Li HX) [24]. Chen Q et al. [25] further argued that the most direct fiscal expenditure items associated with population aging are fiscal expenditures on pensions and healthcare, and in the empirical analysis, the mediating effect of fiscal spending on old age and healthcare is significant. This situation has also been verified in Europe, where Cho D and Lee K) [26] found, using data from several European countries, that public spending on pensions and healthcare showed a significant positive relationship with the rate of population aging, and that population aging significantly reduced fiscal sustainability through these public expenditures. Synthesizing the above studies, to cope with population aging, some literature suggests directional paths, while the mainstream literature specifies two specific paths, namely social security and healthcare. this paper proposes that:

Hypothesis 3. 

Population aging affects fiscal sustainability through two paths: Social security spending and healthcare spending, respectively.

The main views of the literature are shown in

Table 1. Studies on the impact of population aging on fiscal sustainability and representative literature.

Research Hypothesis	Scholars’ Views	Representative Literature
The Impact of Population Aging on Fiscal Sustainability	The negative impact of population aging on fiscal sustainability is not severe	Narayana M R (2012) [7]; Azolibe C B et al. (2020) [8]; Dolls M et al. (2017) [9]
	Aging population can significantly inhibit fiscal sustainability	Jacobsen R H and Jensen S E H (2014) [10]; Van Ewijk C et al. (2006) [11]; Cho S and Kim J R (2021) [12]; Svaljek S (2005) [13]; Liu B and Yang Z (2018) [14]; Wang L et al. (2014) [15]
Heterogeneity in the Impact of Population Aging on Fiscal Sustainability	Developed regions have more secure financial resources to deal with aging populations	Kubanová J and Linda B (2014) [16]; Wu Y et al. (2019) [17]; Carchano M et al. (2021) [18]
	The impact of population aging on fiscal sustainability is positively correlated with the level of economic development	Azolibe C B et al. (2020) [8]; Li J et al. (2019) [19]
Mediating Effects of Population Aging on Fiscal Sustainability	There is a directional path for population aging to affect fiscal sustainability	Okma K and Gusmano M K (2020) [21];
	Significant health care and social security mediating effects	Malačič J (2008) [22]; Jimeno J F et al. (2008) [23]; Fu BY and Li HX (2009) [24]; Chen Q et al. (2022) [25]; Cho D and Lee K (2022) [26]

.

3. Research Design, Variable Descriptions, and Data Description and Sources

In this paper, we first performed a basic regression. Although the basic regression passed the significance test, there may be problems with unscientific variable indicators and endogeneity, so we performed robustness tests (replacing the dependent and independent variables) and two-step systematic GMM tests, respectively. After passing the robustness test and eliminating endogeneity, there may still be heterogeneity problems, so we performed heterogeneity tests at the region and city levels. After passing the heterogeneity test, there may still be a problem that the mechanism of the role between population aging and fiscal sustainability is unknown, so we analyzed the mediation effect of social security fiscal spending and health fiscal spending in depth. At this point, the empirical analysis is scientific, robust, and complete. The empirical analysis is concluded. The specific process is shown in Figure 1.

3.1. Research Design

For Hypothesis 1, according to the theoretical analysis, considering different regions and different period characteristics, this paper sets up a two-way fixed-effects model.

$${\mathrm{FinSus}}_{\mathrm{it}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{Elderly}}_{\mathrm{it}}+{\mathsf{\alpha }}_2{\mathrm{Bed}}_{\mathrm{it}}+{\mathsf{\alpha }}_3{\mathrm{Labor}}_{\mathrm{it}}+{\mathsf{\alpha }}_4{\mathrm{Income}}_{\mathrm{it}}+{\mathsf{\alpha }}_5{\mathrm{EndowInsur}}_{\mathrm{it}}+{\mathsf{\alpha }}_6{\mathrm{HouseholdSav}}_{\mathrm{it}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(1)

In Equation (1), FinSus denotes fiscal sustainability, Elderly indicates an aging population, and Bed, Labor, Income, EndowInsur, and HouseholdSav constitute a set of control variables denoting the number of hospital beds, labor force, disposable income per capita, number of pensioners participating in pension insurance, and residential savings, respectively. To reduce heteroskedasticity and autocorrelation, all control variables are taken as logarithms. μ represents urban fixed effects, δ represents year fixed effects, and $\mathsf{\epsilon }$ represents random perturbation terms. α0 to α6 are the parameters to be estimated.

For Hypothesis 2, this paper first divides the sample into three types of regions, East, West, and Central, and then divides the sample into two types of regions, prefecture-level cities and provincial cities (including provincial capitals and municipalities directly under the central government), and conducts a regression analysis using Equation (1).

For Hypothesis 3, this paper draws on the procedure for testing mediating effects proposed by Pan F et al. [27], using fiscal social security spending as a share of fiscal revenue (SeF) and fiscal health spending as a share of fiscal revenue (MhF) as mediating variables (${\mathrm{Media}}_{\mathrm{it}}$), and designing the econometric equation.

$${\mathrm{Media}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{Elderly}}_{\mathrm{it}}+{\mathsf{\beta }}_2{\mathrm{Bed}}_{\mathrm{it}}+{\mathsf{\beta }}_3{\mathrm{Labor}}_{\mathrm{it}}+{\mathsf{\beta }}_4{\mathrm{Income}}_{\mathrm{it}}+{\mathsf{\beta }}_5{\mathrm{EndowInsur}}_{\mathrm{it}}+{\mathsf{\beta }}_6{\mathrm{HouseholdSav}}_{\mathrm{it}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(2)

$${\mathrm{FinSus}}_{\mathrm{it}}={\mathsf{\gamma }}_0+{\mathsf{\gamma }}_1{\mathrm{Elderly}}_{\mathrm{it}}+{\mathsf{\gamma }}_2{\mathrm{Media}}_{\mathrm{it}}+{\mathsf{\gamma }}_3{\mathrm{Bed}}_{\mathrm{it}}+{\mathsf{\gamma }}_4{\mathrm{Labor}}_{\mathrm{it}}+{\mathsf{\gamma }}_5{\mathrm{Income}}_{\mathrm{it}}+{\mathsf{\gamma }}_6{\mathrm{EndowInsur}}_{\mathrm{it}}+{\mathsf{\gamma }}_7{\mathrm{HouseholdSav}}_{\mathrm{it}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(3)

3.2. Variable Selection

3.2.1. Explained Variables

Fiscal sustainability (FinSus) is used as the explanatory variable in this paper. Fiscal sustainability reflects the level of fiscal sustainability, and there are three main ways to measure fiscal sustainability in the existing literature: First, in the form of percentage or weight, such as the growth rate of fiscal revenue or the fiscal balance as a share of GDP; second, in the form of constructing linear functions or models, such as Markov transformation models, general dynamic equilibrium models, nonlinear fiscal response functions, interperiod budget constraint models, etc.; and third, using the indicator synthesis method, in which a set of indicators is designed to systematically study fiscal sustainability from four dimensions: Fiscal operation soundness, fiscal risk controllability, fiscal system scientificity, and fiscal development balance. Based on the research needs and data availability, this paper draws on the fiscal burden indicator of Aristovnik [4] and creatively proposes to express fiscal sustainability as $1-\frac{\mathrm{FinExp}-\mathrm{FinRev}}{\mathrm{GDP}}$ (FinExp denotes fiscal expenditure and FinRev denotes fiscal revenue). In the subsequent robustness test, the absolute amount of fiscal balance (FinBur) is selected as the explained variable.

3.2.2. Core Explanatory Variables

In this paper, population aging ($\mathrm{Elderly})$ is used as the core explanatory variable. Referring to Liu B and Yang Z [14] and Ricciardi A M et al. [28], the logarithm of the absolute size of the elderly population in each region was selected as the core explanatory variable in the base regression and replaced with the ratio of old age dependency ($\mathrm{Elderly}$ R) as the prime explanatory variable ($\frac{\mathrm{Elderly}\mathrm{population}}{\mathrm{the}\mathrm{working}\mathrm{age}\mathrm{population}}$) in the subsequent robustness tests.

3.2.3. Mediating Variables

There are two mediating variables in this paper, SeF and MhF. SeF is the ratio of social security and fiscal employment to fiscal revenue, and MhF is the ratio of Medicare spending to fiscal revenue.

3.2.4. Control Variables

This study controls the effects of other factors on financial sustainability and tries to mitigate the bias of estimation results due to omitted variables. Referring to Liu C et al. [29] and Chen Q et al. [25], the control variables selected in this paper are (1) the number of hospital beds (Bed), which reflects the level of regional healthcare services; (2) per capita disposable income (Income), which reflects residents’ ability to pay; (3) labor supply level (Labor), which reflects the regional production factor level; (4) EndowInsur, which is an important basis for financial expenditure and is expressed in this paper as the number of pension insurance participants; and (5) HouseholdSav, which reflects a region’s level of economic development and has the role of transforming it into investment and promoting social and economic development, which is expressed as the absolute amount of household savings in this paper.

3.3. Data Description and Sources

Given the availability of data, this paper collects data from 4 municipalities directly under the central government, 8 provincial capitals, and a sample of 89 prefecture-level cities from 2010–2019, for a total of 101 sample data. However, among the 89 prefecture-level city samples, Gannan Tibetan Autonomous Prefecture in Gansu Province belongs to the western minority region, which may be influenced by China’s western development support policy, and the 10-year fiscal expenditure of this region exceeds the GDP, so it was excluded from the sample. Therefore, the sample selected in this paper includes 4 municipalities directly under the central government, 8 provincial capital cities, and 88 prefecture-level cities, for a total of 100. For the data that were still missing after collecting statistics, no technical means were taken to fill in the gaps so as to guarantee the authenticity and scientificity of the empirical analysis results as much as possible.

Table 2. Variable selection and description.

Variable Type	Variable Name	Number of Obs	Mean	Standard Deviation	Minimum	Maximum
Explained variables	FinSus	991	0.895	0.134	0.00400	1.018
	lnFinBur	981	4.786	0.863	1.073	7.906
Explanatory variables	lnElderly	935	3.768	0.867	1.316	5.901
	ElderlyR	966	0.204	0.0910	0.0600	0.487
Control variable	lnBed	962	9.540	0.826	7.449	11.75
	lnIncome	990	10.15	0.345	9.403	11.04
	lnLabor	936	5.392	0.800	2.968	7.631
	lnEndowInsur	889	13.17	1.305	5.094	16.62
	lnHouseholSav	978	16.09	1.091	13.48	19.28
Mediating variable	MhF	990	0.262	0.215	0.0380	1.585
	SeF	990	0.421	0.394	0.0580	3.896

Source: China Statistical Yearbook, Provincial Statistical Yearbook, China Urban Statistical Yearbook, Statistical Bulletin of National Economic and Social Development.

reports the statistical characteristics of each variable.

The economic and social data and the population aging data were acquired from the China City Statistical Yearbook and the statistical yearbooks of prefecture-level cities, and the missing parts were obtained from the national economic and social development bulletins of various regions.

4. Empirical Analysis

4.1. Basic Regression

We estimate Equation (1) to test hypothesis 1, and the findings are presented in

Table 3. Basic regression.

	(1)	(2)	(3)
	FinSus	FinSus	FinSus
lnElderly	−0.050 ***	−0.020 **	−0.047 ***
	(−4.62)	(−2.05)	(−2.72)
lnBed			0.012
			(0.58)
lnIncome			0.154 ***
			(3.37)
lnLabor			0.090 ***
			(2.72)
lnEndowInsur			-0.015
			(−0.87)
lnHouseholdsav			−0.003
_cons	1.083 ***	0.986 ***	−0.778
	(22.96)	(29.39)	(−1.54)
N	945	945	806
Adjusted ${\mathrm{R}}^2$	0.095	0.274	0.347
City FE	NO	Yes	Yes
Year FE	NO	Yes	Yes

Notes: *** and ** denote 1% and 5% levels of significance, respectively. T-values are given in parentheses.

. In Table 3, column (1) is regressed by mixed least squares, column (2) and column (3) are regressed by two-way fixed effects, column (1) and column (2) do not contain any control variables, and column (3) adds control variables.

According to Table 3’s regression results, the results of the mixed least-squares regression and the two-way fixed effects regression are similar in that the impact of the aging population on fiscal sustainability both pass the 5% significance test with negative coefficients, indicating that population aging is negatively related to fiscal sustainability in China, i.e., as the aging population increases, it significantly reduces the level of fiscal sustainability, and this finding verifies Hypothesis 1. In addition, in column (3), both disposable income per capita and labor force pass the 1% significance test with positive coefficients, indicating that both economic development and labor force growth contribute to fiscal sustainability.

4.2. Robustness Test

In order to evaluate the reliability of the results of basic regression research, three methods are adopted in this paper to evaluate the reliability of the regression estimation results. First, the absolute size of the elderly population (lnElderly), the core explanatory variable, was replaced by ElderlyR. The results are presented in column (1) of

Table 4. Robustness test.

	(1)	(2)	(3)
	FinSus	lnFinBur	lnFinBur
ElderlyR	−0.687 ***	1.851 **
	(−7.09)	(2.57)
lnElderly			0.159 ***
			(2.68)
lnBed	0.010	0.441 ***	0.432 ***
	(0.59)	(3.43)	(3.54)
lnIncome	0.117 ***	−0.008	−0.109
	(3.17)	(−0.06)	(−0.88)
lnLabor	0.012	0.023	−0.200
	(1.23)	(0.13)	(−1.46)
lnEndowInsur	−0.028	0.094 *	0.058
	(−1.65)	(1.90)	(1.13)
lnHouseholdsav	−0.019	−0.133	−0.172
	(−1.09)	(−1.22)	(−1.46)
_cons	0.376	0.690	3.740 *
	(0.84)	(0.30)	(1.71)
N	804	794	796
Adjusted ${\mathrm{R}}^2$	0.402	0.765	0.759
City FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes

Notes: ***, **, and * denote 1%, 5%, and 10% levels of significance, respectively. T-values are given in parentheses.

. We find that the ElderlyR also significantly inhibits the level of China’s fiscal sustainability after the change in the core explanatory variable. Second, the explained variable was replaced by financial burden (lnFinBur) instead of financial sustainability (FinSus). The regional financial burden was measured by the absolute amount of difference between fiscal expenditure and fiscal revenue. In order to reduce heteroscedasticity, the logarithmic processing of financial burden was carried out. The results are presented in column (2) of Table 4. We found that ElderlyR passed the significance level of 5% and the coefficient was positive, indicating that the financial burden was significantly increased by the old-age dependency ratio, thus inhibiting the financial sustainability. This conclusion is consistent with the results of column (1) of Table 4. Third, the core explanatory variable (lnElderly) was regressed to the financial burden (FinBur). The results are shown in column (3) of Table 4. We found that the elderly population passed the tests at the significance level of 1% and had a positive coefficient, indicating that with the increase in the elderly population, financial burden increased and financial sustainability decreased. In conclusion, these three conclusions prove that the regression results of Hypothesis 1 are robust.

4.3. Endogenous Processing

Fiscal sustainability has a lag, i.e., the level of fiscal sustainability in the previous period affects the fiscal sustainability in the current period. Therefore, the model is changed to a dynamic panel model by adding the previous period’s fiscal sustainability to the original regression model, so a two-step system GMM is used for estimation, and the results are reported in

Table 5. Endogenous test.

	(1)	(2)
	FinSus	FinSus
L.FinSus	0.743 ***	0.724 ***
	(12.26)	(12.12)
L2.FinSus	0.217 ***	0.237 ***
	(3.31)	(3.48)
lnElderly	−0.040 ***
	(−2.61)
ElderlyR		−0.237 ***
		(−3.54)
lnBed	0.054 ***	0.053 **
	(2.58)	(2.50)
lnIncome	0.199 ***	0.175 ***
	(5.38)	(5.15)
lnLabor	0.059 ***	0.009
	(2.68)	(0.54)
lnEndowInsur	−0.011	−0.005
	(−0.55)	(−0.26)
lnHouseholdsav	−0.064 ***	−0.054 ***
	(−3.12)	(−2.83)
_cons	−1.577 ***	−1.378 ***
	(−4.94)	(−4.60)
N	707	705
AR(1) in first differences	Pr > z = 0.000	Pr > z = 0.000
AR(2) in first differences	Pr > z = 0.188	Pr > z = 0.1
Hansen test	Prob > chi2 = 0.242	Prob > chi2 = 0.262

Notes: *** and ** denote 1% and 5% levels of significance, respectively. T-values are given in parentheses.

. We find that the lagged one period (L.FinSus) of fiscal sustainability is positively correlated at the significance level of 1%, and the lagged two periods (L2.FinSus) are also positively correlated at the significant level of 1%, indicating that the level of fiscal sustainability in the previous period significantly affects the level of fiscal sustainability in the current period and shows a significant positive correlation. The core explanatory variables Elderly and ElderlyR are both significant at the 1% level with negative coefficients, indicating that population aging significantly reduces the level of fiscal sustainability. The Arellano–Bond test proves the existence of first-order autocorrelation but not second-order autocorrelation for the differential disturbance term, and the Hansen test proves the validity of the instrumental variables, and this finding again verifies Hypothesis 1.

4.4. Heterogeneity Test

The 100 samples selected in this paper cover the eastern, central, and western areas of China, and there are great differences among different regions. Therefore, a more effective two-way fixed-effect model is selected for the above analysis. In order to further test the aging of the population’s impact on fiscal sustainability in different regions, this paper divides 100 samples into eastern, central, and western regions according to the regional division of the National Bureau of Statistics of China and conducts a heterogeneity analysis. According to

Table 6. Heterogeneity test: East, Central, and West.

	(1) East	(2) Central	(3) West
Variables	FinSus	FinSus	FinSus
lnElderly	−0.103 **	−0.046 ***	−0.104
	(−2.054)	(−2.848)	(−1.659)
lnBed	0.083 ***	−0.006	−0.047
	(2.826)	(−0.269)	(−1.177)
lnIncome	0.010	0.351 ***	0.045
	(0.319)	(4.562)	(0.292)
lnLabor	0.075 ***	0.080 ***	0.002
	(2.781)	(3.293)	(0.006)
lnEndowInsur	0.041 **	−0.047***	−0.034
	(2.143)	(−3.123)	(−1.464)
lnHouseholdsav	−0.128 ***	0.009	0.116
	(−3.395)	(0.693)	(1.515)
Constant	1.652 *	−2.402 ***	−0.397
	(1.869)	(−2.904)	(−0.135)
Observations	270	419	117
Adjusted ${\mathrm{R}}^2$	0.704	0.710	0.951
city FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
F	5.575	4.923	1.408

Notes: ***, **, and * denote 1%, 5%, and 10% levels of significance, respectively. T-values are given in parentheses.

’s regression results, the findings showed that the aging population in the eastern and central areas both passed the 5% significance test, while the western area failed to pass the significance test, possibly because the sample size was too small. The population aging coefficients are all negative, which verifies hypothesis 1 again. The coefficient of population aging in the east is -0.103 and in the middle is -0.046, indicating that population aging presents heterogeneity in different regions, which verifies the proposition in the first half of hypothesis 2.

Compared to the central region, the eastern region is more severely hit by the aging population. The reasons for this are that, on the one hand, the eastern region is the fastest growing region in China with relatively better medical conditions, and therefore the population’s average life expectancy in this region is relatively long; and on the other hand, it may be due to the faster pace of economically developed cities, where people are more concerned about the quality of life and have a more cautious attitude toward childbirth, thus resulting in a relatively low fertility rate. Combined with the dual effects, the aging population worsens the financial sustainability of the region.

Compared with the eastern area, the aging population in the central region has less impact on fiscal sustainability. The reasons for this are, on the one hand, the relatively less-developed economic and social development in the central region, the related poor medical conditions, and the relatively short average life expectancy of the population; on the other hand, it may be the relatively high fertility rate. The double impact is superimposed, and population aging has a lesser impact on fiscal sustainability in this region.

The sample of this paper contains 4 municipalities directly under the central government, 8 provincial capitals, and 88 prefecture-level cities, and there are large differences among cities at different levels. In order to further test the variations in the impact of population aging on fiscal sustainability in cities of different levels, we include municipalities directly under the central government and provincial capitals as one group and prefecture-level cities as another group for regression analysis, and the regression findings are presented in

Table 7. Heterogeneity test: Provincial cities and Prefecture-level cities.

	(1)	(2)
Variables	FinSus	FinSus
lnElderly	−0.058 ***	−0.043 ***
	(−2.660)	(−2.781)
lnBed	−0.004	0.006
	(−0.150)	(0.298)
lnIncome	−0.024	0.166 ***
	(−0.743)	(3.344)
lnLabor	0.010	0.078 ***
	(0.156)	(2.767)
lnEndowInsur	0.021	−0.016
	(1.099)	(−1.248)
lnHouseholdsav	−0.011	0.013
	(−0.926)	(1.220)
Constant	1.350 **	−1.121 **
	(2.302)	(−1.994)
Observations	102	704
Adjusted ${\mathrm{R}}^2$	0.842	0.879
city FE	Yes	Yes
Year FE	Yes	Yes
F	1.396	3.469

Notes: *** and ** denote 1% and 5% levels of significance, respectively. T-values are given in parentheses.

. Column (1) shows the data for provincial cities and column (2) for prefecture-level cities. From the regression results, it is evident that population aging is significant at the 1% level and the coefficients are negative. The coefficient of the influence of population aging in provincial capitals and municipalities directly under the central government is −0.058, while the coefficient of other cities is −0.043, which indicates that the impact of population aging on fiscal sustainability is greater in provincial cities compared to prefecture-level cities, thus verifying the latter half of the proposition of Hypothesis 2.

4.5. Mediation Effect Test

In this paper, the paths of population aging on fiscal sustainability are analyzed using a mediation model with healthcare fiscal spending (MhF) and social security and employment fiscal spending (SeF) as mediating variables, respectively, and the results are reported in Table 7.

4.5.1. Population Aging, HealthCare Fiscal Spending, and Fiscal Sustainability

From

Table 8. Mediation effect test.

	(1)	(2)	(3)	(4)
	MhF	SeF	FinSus	FinSus
lnElderly	0.056 **	0.126 ***	−0.035 **	−0.028 **
	(2.59)	(2.64)	(−2.43)	(−2.16)
MhF			−0.213 ***
			(−4.09)
SeF				−0.148 ***
				(−7.64)
lnBed	−0.019	−0.071	0.008	0.001
	(−0.47)	(−0.86)	(0.38)	(0.09)
lnIncome	−0.183 ***	−0.764 ***	0.115 ***	0.041
	(−2.84)	(−3.40)	(2.79)	(1.27)
lnLabor	−0.053	−0.288 **	0.079 ***	0.047 ***
	(−1.45)	(−2.07)	(2.86)	(2.71)
lnEndowInsur	0.011	−0.036	−0.012	−0.020
	(0.46)	(−0.84)	(−0.75)	(−1.37)
lnHouseholdsav	0.009	−0.073	−0.001	−0.014
	(0.43)	(−1.14)	(−0.11)	(−1.01)
_cons	1.920 **	11.113 ***	−0.368	0.865 **
	(2.32)	(4.13)	(−0.81)	(2.07)
N	806	806	806	806
Adjusted ${\mathrm{R}}^2$	0.3172	0.3322	0.4006	0.4813
City FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes

Notes: *** and ** denote 1% and 5% levels of significance, respectively. T-values are given in parentheses.

, we observe these three links: First, as reported in column (1), the effect of population aging on healthcare fiscal spending passes the significance test at the 5% level, stating clearly that the higher the rate of population aging, the higher the cost of healthcare fiscal spending, and the two show a significant positive correlation, and this result verifies the path of healthcare fiscal spending in hypothesis 3. Second, putting both healthcare fiscal spending and population aging into Equation (3), the results are presented in column (3), and the effects of population aging and healthcare fiscal spending on fiscal sustainability pass the tests at the significance levels of 5% and 1%, respectively, indicating that the effect of healthcare fiscal spending as a mediating variable is significant and shows a negative correlation, i.e., the higher the aging population’s rate, the greater the fiscal expenditure on healthcare and the lower the fiscal sustainability. The main reason for this phenomenon is that with the implementation and continuous improvement of China’s resident healthcare policy, the amount of its financial investment and the area covered both increase positively with the deepening of population aging.

4.5.2. Population Aging, Social Security, and Employment Fiscal Spending and Fiscal Sustainability

Three links are observed in Table 8: First, as shown in column (2), the influence of population aging on social security employment fiscal spending passes the tests at the significance level of 1%, indicating that the higher the rate of population aging, the greater the social security employment fiscal spending, showing a significant positive correlation, and this result verifies the path of social security employment fiscal spending in hypothesis 3. Secondly, putting social security employment fiscal expenditure and population aging into Equation (3) at the same time, the findings are shown in column (4), and the effects of population aging and social security employment on fiscal sustainability pass the tests at the significance level of 5% and 1%, respectively, indicating that the impact of social security employment as a mediating variable is significant and shows a negative correlation, i.e., the higher the aging population’s rate, the greater the fiscal expenditure on social security employment and the lower the fiscal sustainability. The primary cause of this phenomenon is that China’s population age structure is changing, and population aging has taken on a long-term nature in which the imbalance between income and the expenditure of social security funds has become a common phenomenon, which requires the government to continuously increase the fiscal investment in social security to resolve the issue of old-age security.

5. Conclusions, Policy Recommendations, and Future Research Directions

Based on the existing population aging scenario in China and the availability of data, this article analyzes the impact of population aging of 100 cities on fiscal sustainability in China from 2010 to 2019. This paper proposes three hypotheses and tests the hypotheses by constructing a two-way fixed-effects model and a mediation model, and the main conclusions are:

(1)

Population aging significantly inhibits fiscal sustainability in China. From the empirical analysis, each 1% increase in population aging reduces fiscal sustainability by −0.047%.

(2)

There is heterogeneity in the effect of the elderly population on fiscal sustainability across regions. The impact of population aging on fiscal sustainability is more severe in the eastern region than in the central region; the effect of the aging population on fiscal sustainability is more severe in provincial cities than in prefecture-level cities. From the empirical analysis, every 1% increase in population aging in the eastern region reduces fiscal sustainability by −0.103%, and every 1% increase in population aging in the central region reduces fiscal sustainability by −0.046%. Relative to prefecture-level cities, regarding population aging in municipalities directly under the central government and provincial capitals, the impact on fiscal sustainability is more severe, as the empirical analysis shows that for every 1% increase in population aging, the fiscal sustainability of provincial and prefecture-level cities decreases by −0.058% and 0.043%, respectively.

(3)

The aging of the population inhibits fiscal sustainability through two paths: Fiscal expenditure on healthcare and fiscal expenditure on social security and employment. From the empirical analysis, each 1% rise in population aging increases fiscal spending on healthcare by 0.056% and fiscal spending on social security by 0.126%, thereby reducing fiscal sustainability by −0.035% and −0.028%, respectively.

Based on the findings mentioned above, the following recommendations are made in this study:

(1)

Delaying the retirement age. The increasing per capita life expectancy leads to population aging, therefore delaying retirement can slow down population aging, and according to China’s national conditions, it is recommended to implement a gradual delayed retirement policy. A gradual delay in retirement policy can postpone fiscal expenditure on pensions, alleviate fiscal pressure, and improve fiscal sustainability.

(2)

Safeguard the fiscal expenditures of eastern regions, provincial capitals, and municipalities directly under the central government to cope with population aging. In conventional thinking, developed eastern regions and provincial cities are fiscally sustainable but their population aging problems are more serious than those manifested in central regions and prefecture-level cities, and population aging has a greater impact on fiscal sustainability, which should attract academic thought and policy attention.

(3)

Improve the performance of the use of funds for fiscal expenditures on healthcare and social security and employment, address the paths through which population aging affects fiscal sustainability, and improve the performance of the two types of expenditure paths. Actively introduce the concept of zero-based budgeting, innovate the budgeting model, break the inertia and rigid pattern of expenditures, reasonably set the levy and expenditure standards for financial expenditure on healthcare and social security employment through a precise screening system, improve the efficiency of fund use, smooth the path of population aging affecting fiscal sustainability, and mitigate the impact of population aging on fiscal sustainability.

There are some research shortcomings in this paper. First, although the data meet the randomness requirement, only some cities are included, and the sample is not sufficient. Second, although the indicators of population aging in this paper are commonly used in the literature, more scientific metrics may exist. In the future, we will expand the range of sample data regions and optimize the indicator measurement method of aging.