Promotion Incentive, Population Mobility and Public Service Expenditure

Abstract

Improving the level of public service expenditure is the basic condition to meet people’s growing yearning for a better life in a new era. This study aims to examine the influence of promotion incentive and population mobility on public service expenditure. With the panel data of 30 provinces (excluding Tibet, Hong Kong, Macao and Taiwan) in China from 2010 to 2018, a fixed effect model was employed to conduct the empirical research first, and then analyze the heterogeneity of the impact. In addition, the robustness was examined by substituting core explanatory variables and explained variables and using the spatial econometric model. Results indicated that promotion incentive and population mobility had significant negative effects on public service expenditure (medical care, education, culture, sports and media, social security and employment). By replacing the explained variables and the core explanatory variables and using the spatial economic model, the research conclusion was still stable and reliable. In different regions, promotion incentive and population mobility had different effects on different types of public service expenditure. In addition, it was revealed that the improvement of the urbanization level, economic development level, population density, residents’ education level and the proportion of tertiary industry had a significant positive impact on public service expenditure. The findings in this study can provide useful references for policy makers to allocate public service expenditure.

1. Introduction

For a long time, local governments’ financial and economic expenditure was large and rapidly increasing, while social expenditure on people’s livelihoods was relatively low; infrastructure, such as energy and transportation, is relatively complete, and expenditure on public services such as education, medical care, culture, and sports is obviously insufficient, causing an imbalance in public goods’ development and affecting the public’s rights. Since the 18th National Congress of the Communist Party of China, the Chinese government has been making significant efforts to maintain and improve people’s standard of living, and the building of China’s core public service system has been steadily strengthened. The proportion of financial input for people’s livelihood as a percentage of fiscal spending has been expanding, and numerous service facilities have been vastly enhanced, with individuals’ sense of well-being and sense of gain continuously improving. In the report of the 20th CPC National Congress of the Party, General Secretary Xi Jinping once again emphasized the need to pay greater attention to safeguarding and improving people’s livelihoods in development, to encourage joint efforts to create a better life, and to continually fulfill the aspirations of people for a better life.

Improving the level of public service expenditure is a crucial pillar for protecting and enhancing people’s standard of living. Promotion incentive and population mobility are two main elements that significantly impact public service expenditure. In the past, under the background of GDP assessment, local governments place a high priority on the economy but pay little attention to people’s ability to make a living. It is primarily manifested in the relatively perfect infrastructure, such as energy and transportation, and the obvious lack of public services for people’s livelihood, such as medical care, education, culture, sports and media, social security and employment. This will greatly affect the accessibility and equality of people’s access to public services [1]. Population mobility also has a significant effect on the cost of providing public services. Local governments allocate public service resources such as education, medical care, and culture to their administrative areas according to the registered population in their jurisdiction, which greatly restricts their actual public service supply capacity, and there is a certain competition and exclusivity for residents outside the space to enjoy public service resources. Regarding the population, it may result in contradictions such as an excess of public goods, resulting in idle assets and waste. Based on the aforementioned concerns, this article considers promotion incentive, population mobility, and public service expenditure within the same framework. This study plays an important role in improving the policies of the public service system, and people’s perceptions of fairness, gain, and pleasure.

Promotion incentive refers to the evaluation of the work completion of lower-level officials in higher-level departments according to relevant policy documents and performance appraisal schemes, and the evaluation results are used as the basis for official promotion [2]. The mainstream view of promotion incentive is that the promotion tournament of local government officials guided by the GDP assessment standard encourages local governments to invest limited financial funds in productive financial expenditure areas that are conducive to economic performance, such as capital construction investment, investment in tapping the potential of enterprises and other fields, while ignoring the unproductive financial expenditure areas with relatively small economic contributions such as medical care, education, culture, sports and media [3,4]. Population mobility is another important factor affecting public service expenditure. According to the data of the seventh national census in 2020, the amount of population mobility in China reached 376 million people, 1.7 times that of 2010 and 3.13 times that of 2000. A large amount of population mobility has an important impact on the economic and social development of inflow and outflow places. As far as the inflow location is concerned, population mobility not only brings demographic dividends and promotes the commodity circulation and economic development in the inflow location, but also increases the burden of public facilities and the difficulty of urban governance in the inflow location, profoundly affecting the scale and level of public service expenditures such as medical care, education, culture, sports and media. As far as the emigration location is concerned, the outflow of population plays a positive role in alleviating the local people’s conflicts and the shortage of public service expenditure. However, in the long run, the migration of population will lead to brain drain and labor shortage, which will ultimately affect the sustainable development of the economy and society in the emigration location.

There is a close relationship between promotion incentive, population mobility and public service expenditure. Since the 18th National Congress of the Communist Party of China, the Chinese government has been paying more attention to people’s livelihoods and has been gradually incorporating the improvement of people’s livelihoods and the equalization of public services into the performance evaluation system of local governments, which will inevitably affect the local government’s fiscal expenditure behavior and the bias of public goods investment. Furthermore, as the economy and society expand, there is a large-scale population flow between regions, causing the population size and structure of the immigration and emigration locations to alter appropriately. Population change will unavoidably impact the structure of fiscal expenditure and cause an adjustment in the supply bias of public goods. Furthermore, as the economy and society expand, there is a large-scale population flow between regions, causing the population size and structure of the immigration and emigration locations to alter appropriately. Population change will unavoidably impact the structure of fiscal expenditure and cause an adjustment in the supply bias of public goods. So, what influence does the central government’s policy direction have in the context of promotion championships and large-scale population movements? That is, will the local government pay attention to the input of public goods for people’s livelihood, and what kind of public goods for people’s livelihood would it prefer? Will there be geographical variations in the impact of this? In response to the above questions, this paper empirically tests the impact of promotion incentive and population mobility on public service expenditure by using the fixed effect model based on the panel data of 30 provinces (excluding Tibet, Hong Kong, Macao and Taiwan) in China from 2010 to 2018. This study is helpful in optimizing the incentive mechanism for the promotion of officials, and has important practical significance for formulating scientific and reasonable policies for population mobility.

This research is organized as follows: the Section 1 is the introduction; the Section 2 is the literature review; the Section 3 consists of measures and the analytical strategy; the Section 4 consists of empirical results and analysis (containing basic regression results, robustness test and heterogeneity test); the Section 5 is the conclusion.

2. Literature Review

2.1. The Relationship between Promotion Incentive and Public Service Expenditure

The performance evaluation of some local governments lacks citizen orientation, which contributes to the wrong view of local governments’ achievements. It is mainly manifested in the fact that local governments attach importance to the economy over people’s livelihood, growth over development, speed over quality. A few local governments ignore people’s livelihood issues that are closest to the public, and even pursue economic growth unilaterally at the expense of the long-term interests of the environment and society. One-sided emphasis on economic growth and neglect of coordinated development of the social environment distort the government’s value orientation, resulting in livelihood issues such as tension between doctors and patients, lack of social security and environmental pollution [5,6,7,8]. This is related to China’s vertical political system. In order to pursue political achievements, local government officials have greater incentives to improve the scale and level of financial expenditures in productive areas such as infrastructure, tapping the potential of enterprises, and ignoring the scale and level of financial expenditures in non-productive areas, such as medical care, education, culture, sports and media, that are difficult to improve in the short term [9,10,11,12,13]. The research based on the provincial panel data, using the systematic GMM model, investigated the impact of promotion incentive on human capital investment (including culture, education and health), which found that promotion incentive and local human capital investment showed a significant negative relationship, as well as cross-period and cross-regional effects that could not be ignored [14]. Another study found that the higher the preference of local officials for political promotion, the more certain the role of productive expenditure, and the lower the degree of local government consumption expenditure [15]. In cities with stronger incentives for officials’ promotion, the strategic interaction between regions of economic public goods investment is stronger [16]. This yardstick competition mechanism is not reflected in the investment of social public goods. The research found that promotion incentives crowded out local government’s public education investment [17]. The research of Hu and Peng also confirmed the conclusion that promotion incentive reduce the supply of rural medical and health public services [18]. Others discovered that the promotion incentive of officials had no significant impact on the efficiency of education expenditures, but had a positive impact on the efficiency of productive financial expenditures, such as infrastructure [19].

In recent years, the relevant policies issued by the Central Committee of the Communist Party of China have made it clear that the evaluation standard of government officials is no longer based on GDP as the only evaluation index. High-quality development goals not only require local governments to attach importance to economic growth, but also pay attention to improving people’s livelihood and making people’s lives better. In recent years, some related studies have confirmed the conclusion that promotion incentives have a positive effect on public service expenditure. Blanchard and Shleifer [20], Zou [21] found that promotion incentives are positively related to the supply of social public goods such as medical care and education. Using the data of German federal states, Fischer and Wigger [22] studied the relationship between local government competition and higher education expenditure. The analysis revealed a positive correlation between local government competition and investment on higher education. The research found that the increase in government competition leads to the increase in education expenditure [23]. Yang et al. [24] used the spatial Durbin model to examine the influence of the official promotion incentive on the supply of public goods. The study found that the political promotion incentive of local officials has a positive impact on the supply of public goods, but this impact has an obvious time lag effect. Hu et al. [25] found that the promotion incentive is conducive to improving the efficiency of financial social security expenditure. Based on the above analysis, the influence of the promotion incentive on public service expenditure needs to be further confirmed. This paper puts forward the following hypothesis, Hypothesis 1:

H1a: 

Promotion incentives have a negative impact on unproductive public service expenditure.

H1b: 

Promotion incentives have a positive impact on unproductive public service expenditure.

2.2. The Relationship between Population Migration and Public Service Expenditure

The push–pull theory of population migration divides the causes of population migration into two aspects: push and pull. The former prompts immigrants to leave their original place of residence. The latter attracts immigrants who want to improve their lives to move to new places of residence [26]. The particularity of China’s national conditions determines the difficulty and urgency of the transfer of rural surplus labor force in China. Since the 1980s, the intensification of rural reform, as well as the stability and perfection of the home contract responsibility system, have galvanized the passion of farmers and further emancipated social production. At the same time, surplus rural labor force problems are gradually emerging. The development of township enterprises has opened up a new way to solve the employment problem of rural surplus labor force, but this alone is not enough to absorb rural surplus labor force. In the twenty-first century, with the acceleration of urbanization in China, the transfer of surplus rural labor force has become a global issue. A large amount of population mobility has an impact on the economic and social development of the places of immigration and emigration, and then profoundly affect the scale and level of public service expenditure in the places of immigration and emigration [27,28]. Previous studies have found that population mobility is negatively related to the supply of basic public services of local governments [4,29,30,31]. For the places where the population flows in, population mobility brings the crowding-out effect of public goods [30]. Due to the existence of the household registration system, it is difficult for the floating population to enjoy the same public services as the registered population [32,33], while there will be surplus supply of public goods in the places where the population flows out, resulting in a waste of financial resources [34]. However, the floating population with housing property rights have a greater probability of obtaining relatively complete basic public services [35]. The emigration of young people from rural areas led to a de-crease in the supply of rural public goods and caused a lack of public services in rural areas [36,37]. According to Chen [38], immigration has a negative competitive effect and a positive financial effect on the supply of public goods. The type of influence immi-gration has on the supply of public goods depends on the relative magnitude of these two effects. Based on the panel data of China’s prefecture-level cities in 2010–2014, Yang et al. [39] examines the impact of the floating population size on the allocation of basic public service resources. It is found that the larger the floating population size is, the lower the per capita general education expenditure, social security and employment expenditure, and medical and health expenditure are, and the impact at the regional level shows obvious heterogeneity. Based on the above analysis, the following hypothesis, Hypothesis 2, is put forward:

H2. 

Population mobility has a negative impact on unproductive public service expenditure.

3. Measures and Analytical Strategy

3.1. Measures

This paper focuses on the impact of the promotion incentive and population mobility on public service expenditure (medical care, education, culture, sports and media, social security and employment). The magnitude of such public service expenditures is represented in public goods for people’s survival. Public goods refer to goods and services that, unlike private goods, cannot be effectively produced and supplied by enterprises and individuals through the market mechanism, but are instead primarily provided by the government and other public organizations to meet society’s public needs. Products in education, culture, radio, television, medical treatment, applied scientific research, sports, and other disciplines are examples of public goods. Given the data available, this article chooses the most representative financial expenditures in the categories of healthcare, education, culture, sports and media, social security, and employment to assess the scope and degree of regional basic public service. Explained variables, core explanatory variables and control variables involved in this paper are as follows:

This paper selects per capita medical and health expenditure, per capita education expenditure, per capita culture, sports and media expenditure, per capita social security and employment expenditure as explained variables to reflect the scale and level of regional basic public service expenditure. The core explanatory variables that this paper focuses on are promotion incentive and population mobility. Promotion incentive is measured by the per capita utilization of foreign direct investment and the level of opening to the outside world (the total import and export trade volume is higher than the GDP of the previous region). Population mobility is expressed by the number of permanent residents compared with the number of registered residents, which is used to reflect the proportion of the floating population in this area. When the ratio is greater than 1, it indicates that this area is an inflow place; when the ratio is less than 1, it indicates that this area is an outflow place.

This paper selects five control variables: urbanization rate, economic development level, population density, residents’ education level and industrial structure. Specifically: (1) Urbanization rate is measured by the proportion of urban population to the total population at the end of each region (population urbanization); (2) Per capita GDP of each region is used to measure the regional economic development level; (3) Population density is measured by the ratio of resident population to geographical area in each region; (4) The educational level of residents is measured by the average years of education of the residents; (5) Industrial structure is measured by comparing the value-added of the tertiary industry with the GDP of the last region. For variables other than ratios, logarithmic processing is used. On the one hand, the absolute difference between the data is minimized in order to avoid the influence of individual extreme values on the empirical conclusions. Simultaneously, using logarithms might limit the avoidance of multiple collinearity and heteroscedasticity between variables. On the other hand, using the logarithm makes the interpretation more economical. The above data comes from the Statistical Yearbook of China in the relevant years and the Statistical Yearbook of the provinces and cities where it is located.

Table 1. Descriptive statistical results of variables.

Category	Meaning	Variable Name	Sample Number	Mean Value	Standard Deviation	Minimum Value	Maximum Value
	Per capita medical and health expenditure (logarithmic)	PMexp	270	6.6167	0.4350	5.5666	7.7614
Explained variables	Per capita expenditure on education (logarithmic)	PEexp	270	7.4226	0.3853	6.4192	8.4682
	Per capita expenditure on culture, sports and media (logarithmic)	PCexp	270	5.2240	0.5648	3.9427	7.0383
	Per capita social security and employment expenditure (logarithmic)	PSexp	270	7.1405	0.5211	5.9373	8.3866
Explanatory variables	Per capita utilization of foreign direct investment (logarithmic)	Pro1	270	6.4904	1.3444	3.8197	11.5272
	Opening to the outside world	Pro2	270	6.4904	0.3225	0.0170	1.5480
	Proportion of population mobility	Pms	270	1.0404	0.1960	0.7900	1.6900
Control variables	Urbanization rate	Urb	270	0.5653	0.1258	0.3380	0.8960
	Per capita GDP (logarithmic)	Pgdp	270	11.3561	0.5806	9.4818	12.0594
	Population density (logarithmic)	Den	270	5.4553	1.2784	2.0530	8.2490
	Education level of residents	Edu	270	9.0510	0.9247	6.7640	12.6750
	Industrial structure	Ind	270	0.4486	0.0959	0.2860	0.8100

reports the descriptive statistical results of each variable.

3.2. Analytical Strategy

This study utilized the panel fixation effect model based on relevant research by Ding and Deng Kebin [40], Li et al. [41], and Liu and Zhang [42]. In geographic panel data, fixed effect regression is a type of variable approach that varies with individuals but not over time. One of the n distinct intercepts in the fixed effect model corresponds to an individual. A succession of binary variables may be used to express these intercepts. The individual fixed effect model, time-point fixed effect model, and time-point individual fixed effect model are the three classifications of the fixed effect model. In this paper, the formula for the double fixed effect model is as follows:

PMexpit = β0 + β1Proit + β2Pmsit + δXit + εit

(1)

PEexpit = β0 + β1Proit + β2Pmsit + δXit + εit

(2)

PCexpit = β0 + β1Proit + β2Pmsit + δXit + εit

(3)

PSexpit = β0 + β1Proit + β2Pmsit + δXit + εit

(4)

In the above formulas, Formulas (1)–(4), the meanings of each symbol is as follows: I is a panel individual variable with a value of 1–30 (excluding Tibet, Hong Kong, Macao and Taiwan). t is a time variable, and the corresponding sample interval is from 2010 to 2018. The explained variables are local public service expenditure, including medical and health expenditure (PMexp), education expenditure (PEexp), culture, sports and media expenditure (PCexp), social security and employment expenditure (PSexp). Pro represents promotion incentive. Pms represents floating population. X are control variables, including urbanization rate, economic development level, population density, residents’ education level and industrial structure, which affect the level of regional public service expenditure, and ε are random error terms. If the regression coefficient of β1 is significantly negative, then H1a is established, which states that promotion incentive has a negative impact on unproductive public service expenditure. If the regression coefficient of β1 is significantly positive, then H1b is established, which states that the promotion incentive has a positive impact on unproductive public service expenditure. β2 represents the regression coefficient for the percentage of the population. If the regression coefficient of β2 is significantly negative, then H2 is established. That is to say population mobility is negatively associated with public service expenditure.

4. Empirical Results and Analysis

4.1. Basic Regression Results

The estimated results of the influence of population mobility and promotion incentive on regional public service expenditures are presented in

Table 2. The estimated results of the influence of promotion incentive and population mobility on public service expenditure.

Variable	PMexp		PEexp		PCexp		PSexp
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Pro1		−0.0426 *** (−2.74)		−0.0634 *** (−4.79)		−0.0620 *** (−2.90)		−0.1106 *** (−6.01)
Pms		−3.0186 *** (−6.01)		−1.4593 *** (−3.41)		−2.9921 *** (−4.33)		−2.4531 *** (−4.12)
Urb	4.0855 *** (8.68)	2.8865 *** (6.06)	2.7757 *** (7.03)	2.0807 *** (5.12)	1.0077 (1.60)	−0.2341 (−0.36)	3.0641 *** (5.42)	1.8850 *** (3.34)
Pgdp	0.0999 (4.43)	0.1033 *** (4.91)	0.1725 *** (9.12)	0.1688 *** (9.41)	0.1739 *** (5.78)	0.1748 *** (6.04)	0.0641 ** (2.36)	0.0574 ** (2.30)
Den	2.0661 *** (4.76)	4.0061 *** (7.86)	1.3931 *** (3.83)	2.4355 *** (5.60)	1.9709 *** (3.40)	3.9422 *** (5.62)	2.5594 *** (4.91)	4.3213 *** (7.16)
Edu	0.2325 *** (7.38)	0.2033 *** (6.81)	0.2096 *** (7.93)	0.1810 *** (7.11)	0.2739 *** (6.51)	0.2383 *** (5.81)	0.2117 *** (5.59)	0.1622 *** (4.59)
Ind	1.8859 *** (8.43)	2.4362 *** (10.77)	0.0289 (0.15)	0.4370 *** (2.27)	0.5040 * (1.69)	1.1151 *** (3.58)	2.3279 *** (8.66)	3.0272 *** (11.30)
Cons	−11.0479 *** (−4.96)	−17.5588 *** (−7.55)	−5.6148 *** (−3.01)	−8.8610 *** (−4.47)	−10.7771 *** (−3.63)	−17.2764 *** (−5.40)	−12.2416 *** (−4.58)	−17.7076 *** (−6.42)
Obs	270	270	270	270	270	270	270	270
model	FE	FE	FE	FE	FE	FE	FE	FE

Note: ***, ** and * indicate that the significance level reaches 10%, 5% and 1%, respectively. The values in brackets are T, and FE is fixed effect model.

. The fixed effect model is chosen for each test equation after the Hausman test has been completed. Models 1, 3, 5, and 7 only include the results of regressions on control variables; Models 1, 3, 5, and 7 do not include the main explanatory variables of promotion incentive or population mobility. The regression results of two primary explanatory variables, namely population mobility and promotion incentive, as well as control variables, can be found in Models 2, Model 4, Model 6, and Model 8. Each model corresponds to the equations of spending on healthcare and education, spending on entertainment, sports, and the media, as well as spending on social security and employment. According to the findings of the regression, the regression coefficients for each promotion incentive are as follows: −0.0426, −0.0634, −0.0620, and −0.1106. All of these values are significant at a level of 1%. This demonstrates that the promotion incentive has a negative impact on the expenditure of regional public services. Therefore, Hypothesis 1a is confirmed. The primary reason is that China evaluates officials based on the level of economic development in the local area. As the proverb says, “whatever type of assessment system exists, the government will respond accordingly.” Local governments’ financial spending on various public goods is prioritized differently as a result of promotion. Economic growth can be promoted directly and effectively by productive expenditures such as infrastructure construction, enterprise tapping and transformation, and rural production expenditures. Increasing local government spending on soft public goods such as healthcare and education has a spillover effect of supporting economic and social growth in the long term, but it does not benefit local politicians with limited tenure in the short term. Therefore, as a result of the incentive, the local government would lower its expenditures in the medical and educational sectors. Comparing the absolute values of regression coefficients reveals that the promotion incentive has the most significant impact in reducing social security and employment expenses, followed by education, culture, sports, media, and medical and health expenditures. The regression coefficients of population mobility for the test equations are −3.0186, −1.4593, −2.9921, and −2.4531, all of which pass the 1% significance test. The results of this study confirm Hypothesis 2 and are consistent with the findings of Yang et al. [39], showing that the rising amount of population mobility has a considerable negative impact on public service expenditure. This illustrates the occurrence and challenges of the supply-and-demand mismatch between the amount of mobile population and regional public service expenditures. Short term, it is difficult for public service delivery to adjust positively and swiftly to fluctuating population levels [39]. Unlike the promotion incentive, population mobility has the greatest effect on reducing medical and health expenditures, followed by culture, sports, and media expenditures, social security expenditures, employment expenditures, and education expenditures.

Except for the insignificant expenditure on culture, sports, and media, the urbanization rate has significant positive effects on medical and health expenditures, education expenditures, social security expenditures, and employment expenditures, according to the regression results of each control variable. Currently, the Chinese government is actively supporting a new urbanization focused on people. As a result of growing urbanization, the government must immediately modify its public service functions. Its primary objective is to expedite the reversal of the imbalance between the supply and demand of public services to realize the level of equalization of public services and to safeguard the fundamental rights and interests of citizens. The present literature confirms that the regression coefficient of the economic development level to all types of public service expenditure is significantly positive. This is mostly due to the fact that an increasingly developed economy results in an increased demand for public services. It is easier to put pressure on local governments and encourage them to raise financial expenditure on public services when there is a higher demand for public services [43]. In addition, the level of economic development is directly correlated to the amount of adequate financial resources that are available to the local government. Governments with sufficient financial resources enable the expansion of public service expenditures and the fulfillment of public needs. The regression coefficients of population density are all significantly positive. This conclusion is consistent with the research findings of Pan [44]. The greater the population density, the greater the population’s demand for public services, which encourages the local government to boost expenditures on public services. The regression coefficients of residents’ educational level are all significantly positive, indicating that improving residents’ educational level has a positive effect on promoting public service expenditure. On the one hand, the higher the level of public education, the more it can stimulate people’s sense of political participation, and the more conducive it is for people to make choices about the behavior of local governments. On the other hand, the improvement of public education can encourage people to participate in the local government’s anti-corruption construction, strengthen their awareness of supervision, and then help local governments to optimize the scale and structure of public service expenditure. The regression coefficients of the industrial structure are all significantly positive, which is consistent with the research results of Pan [44]. The greater the proportion of tertiary industry, the greater the expenditures on public services. The tertiary industry is divided into four departments, including the circulation department, the department that serves production and life, the department that serves to improve the scientific and cultural level and residents’ quality, and the department that serves public needs. The medical care, education, culture, sports and media, and social security and employment expenditure examined in this research belongs to the departments that serve to improve the scientific and cultural level and residents’ quality. Consequently, the greater the proportion of the tertiary industry, the greater the public service expenditure invested by local governments. Both of these requirements are necessary and sufficient.

4.2. Robustness Test

In order to further confirm the reliability of the conclusion, this paper attempts to verify it from the following two aspects.

Models 1–4 replace the core explanatory variables for new ones, measure promotion incentive on the degree of opening to the outside world, and keep the explained variables unchanged. Additionally, the factors that were explained in Models 5 and 8 have also been replaced. The medical expenditure, education expenditure, culture, sports and media expenditure, social security and employment expenditure are replaced, respectively, by the number of medical and health technicians per thousand people, the teacher-student ratio of ordinary colleges and universities, the stock of public libraries per ten thousand people and the ratio of maternity insurance participants. The regression results are shown in

Table 3. Estimated results of promotion incentive and population mobility’s impact on public service expenditure.

Variable	PMexp	PEexp	PCexp	PSexp	PDoc	PTec	PLib	Pins
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Pro2	−0.5740 *** (−5.92)	−0.5459 *** (−6.48)	−0.5717 *** (−4.13)	−0.6823 *** (−5.57)	−2.0505 *** (−7.21)	−0.0176 *** (−7.29)	−0.2874 *** (−3.90)	−0.1233 *** (−6.67)
Pms	−1.7077 *** (−3.35)	−0.1223 (−0.28)	−1.6093 ** (−2.21)	−0.6587 (−1.02)	−5.8607 *** (−3.92)	0.0033 (0.26)	−1.0619 *** (−2.74)	−0.4344 *** (−4.47)
Urb	4.3823 *** (8.77)	3.5974 *** (8.29)	1.3362 * (1.87)	3.9095 *** (6.19)	16.5015 *** (11.26)	0.0076 (0.61)	0.3492 (0.92)	−0.2388 ** (−2.51)
Pgdp	0.1089 *** (5.48)	0.1767 *** (10.24)	0.1826 *** (6.45)	0.0708 *** (2.82)	0.0412 (0.71)	−0.00071 (−1.33)	−0.0011 (−0.07)	0.0046 (1.23)
Den	2.0331 *** (3.61)	0.4483 (0.92)	1.8824 ** (2.34)	1.6861 ** (2.37)	3.4069 ** (2.06)	−0.0472 *** (−3.37)	0.6011 (1.41)	0.4747 *** (4.43)
Edu	0.1804 *** (6.33)	0.1674 *** (6.76)	0.2225 *** (5.47)	0.1565 *** (4.35)	0.1848 ** (2.21)	0.0009 (1.33)	0.0739 *** (3.41)	0.0246 *** (4.53)
Ind	1.7670 *** (7.90)	−0.2878 (−1.48)	0.3731 (1.17)	2.0007 *** (7.09)	2.1587 *** (3.29)	−0.0083 (−1.49)	0.6207 *** (3.65)	0.1177 *** (2.76)
Cons	−8.6762 *** (−3.31)	−0.1691 (−0.07)	−8.2210 ** (−2.20)	−6.5095 ** (−1.97)	−18.6147 ** (−2.42)	0.3191 *** (4.90)	−2.5635 (−1.29)	−2.1705 *** (−4.35)
Obs	270	270	270	270	270	270	270	270
Model	FE	FE	FE	FE	FE	FE	FE	FE

Note: ***, ** and * indicate that the significance level reaches 10%, 5% and 1%, respectively. The values in brackets are T, and FE is a fixed effect model.

. According to the results, the regression coefficients of Model 1–Model 4 are −0.5740, −0.5459, −0.5717 and −0.6823, respectively, and all the regression coefficients pass the significance test of 1%. From the test results of Model 5–8, the regression coefficients of promotion incentive are −2.0505 and −0.0176, respectively. Based on the regression coefficient of population mobility, the regression coefficients of population mobility in Model 1, Model 3, Model 5, Model 7 and Model 8 are significantly negative at different test levels. The regression coefficients of population mobility in Models 2 and Model 4 are also negative, but they fail to pass the test. Except for Model 6, the above-mentioned conclusions generally affirm that the increase in floating population proportion has a negative impact on unproductive public service expenditure. Therefore, the robustness test demonstrates once again that the promotion incentive and population mobility have a negative impact on the unproductive public service expenditure. Due to space limitations, the regression coefficients of each controlled variable are not described in detail.

In fact, public service expenditure is affected not only by local explanatory variables, but also by variables in neighboring areas [39]. Therefore, it is necessary to include spatial factors for analysis, so as to make the estimation results of the model more robust and reliable. Before building the spatial econometric model, it is necessary to test the autocorrelation of all kinds of public service expenditures.

Table 4. Global Moran index value of public service expenditure.

Year	PMexp		PEexp		PCexp		PSexp
	I	p-Value	I	p-Value	I	p-Value	I	p-Value
2010	0.178 **	0.028	0.308 ***	0.002	0.159 **	0.042	0.038	0.214
2011	0.152 **	0.046	0.256 ***	0.008	0.176 **	0.034	0.111 *	0.094
2012	0.110 *	0.096	0.217 **	0.017	0.134 *	0.051	0.101	0.113
2013	0.114 *	0.095	0.272 ***	0.004	0.190 **	0.018	0.178 **	0.038
2014	0.117 *	0.095	0.289 ***	0.003	0.139 *	0.055	0.209 **	0.024
2015	0.130 *	0.077	0.293 ***	0.003	0.124 *	0.067	0.205 **	0.023
2016	0.078	0.170	0.268 ***	0.005	0.118 *	0.072	0.126 *	0.090
2017	−0.015	0.433	0.215 **	0.015	0.079	0.156	0.137 *	0.076
2018	−0.007	0.406	0.216 **	0.015	0.056	0.197	0.218 **	0.019

Note: ***, ** and * represent significance levels of 10%, 5%, and 1%, respectively.

reports the global Moran index test results of various public service expenditures. The results show that the overall Moran index of medical and health expenditure in 2010–2015 passed the significance test (5% or 10% significance level), while the Moran index in 2016–2018 failed the test. The overall Moran index of education expenditure is significantly positive (1% or 5% significance level). The significance of the overall Moran index of sports and media expenditure is similar to that of the overall Moran index of medical and health expenditure. The Moran index passed the test in the first and middle period, but failed in the last period. Except for 2010 and 2012, the overall Moran index of social security and employment expenditure in each year is also significantly positive. The findings presented above demonstrate that there is a steady and obvious positive spatial autocorrelation between education expenditure. In most years, there is a positive spatial autocorrelation between medical care expenditure, culture, sports, and media expenditure; social security and employment expenditure. Through the above analysis, it is necessary to further use the spatial econometric model to verify the impact of promotion incentives and population mobility on public service expenditure.

The global Moran index test findings presented in Table 4 reveal that regional public service expenditure shows significantly positive spatial autocorrelation. Based on the findings of the Lagrange multiplier test, the spatial lag regression model is better than the spatial error model when examining the impact of promotion incentive and population mobility on public service expenditure. Therefore, this research uses the spatial lag model.

Table 5. Results of spatial lag model estimation of the impact of promotion incentive and population mobility on public service expenditure.

Variable	PMexp	PEexp	PCexp	PSexp
Pro1	−0.0078 (−0.83)	−0.0598 *** (−5.58)	−0.0546 ** (−2.57)	−0.0977 *** (−6.25)
Pms	−1.2787 *** (−4.09)	−0.7063 ** (−2.00)	−2.7161 *** (−3.93)	−1.2106 ** (−2.31)
Urb	0.8627 *** (2.86)	1.5932 *** (4.80)	−0.1727 (−0.27)	0.7426 (1.50)
Pgdp	0.0286 ** (2.18)	0.0747 *** (4.34)	0.1486 *** (4.70)	0.0320 (1.50)
Den	1.0686 *** (3.17)	0.8891 ** (2.32)	3.3088 *** (4.32)	2.6240 *** (4.79)
Edu	0.0422 ** (2.16)	0.0598 ** (2.52)	0.2162 *** (5.17)	0.0597 * (1.85)
Ind	0.5134 *** (3.12)	−0.0968 (−0.59)	0.8966 *** (2.75)	1.7214 *** (6.31)
ρ	0.7495 *** (20.47)	0.5366 *** (10.19)	0.1437 * (1.84)	0.4755 *** (8.61)
R2	0.9635	0.9321	0.8099	0.9203
LogLikelihood	315.9201	294.3398	143.2606	206.6516
Observations	240	240	240	240

Note: ***, **and * indicate that the significance level reaches 10%, 5% and 1%, respectively, and the Z value is in brackets.

displays the influence of the promotion incentive and population mobility on public service expenditure as estimated by the spatial lag model. Table 5 indicates that estimated autoregressive coefficients all passed the significance test, demonstrating that the spatial spillover effect of public service expenditure is evident. The regression coefficients of the promotion incentive are −0.0078, −0.0598, −0.0546, and −0.0977. Except for the medical and health expense equation, the promotion incentives are insignificant, whereas they are significantly positive in the other public service spending equations. This demonstrates that the promotion incentive has a significant negative impact on public service expenditure in general. At the present stage, the promotion incentive still has an obvious inhibitory effect on the unproductive fiscal expenditure, which means that although the new performance appraisal system does not take the explicit GDP as the standard, in the short term, it is still difficult to put an end to the phenomenon that local officials secretly compete with the GDP. The regression coefficients of population mobility are −1.2787, −0.7063, −2.7161 and −1.2106, and pass the significance test at the 1% or 5% level. The bigger the share of population mobility, the greater the crowding effect, which drastically reduces the scale and level of public service expenditure. Based on the above analysis, promotion incentive and population mobility have significant negative effects on public service expenditure. The regression results considering spatial factors are consistent with the previous estimation results, which proves that the research results of this paper are robust and reliable again. Due to space restrictions, the regression coefficient of each controlled variable is not explained in detail.

4.3. Heterogeneity Analysis

The disparity in the levels of economic development in China’s various regions is rather glaring given the country’s size as a whole. There are clear disparities between the levels of public service spending in each of the several areas. In order to conduct an in-depth examination of the regional differences, the sample data from each of China’s thirty provincial administrative units (with the exception of Tibet, Hong Kong, Macao, and Taiwan) are first separated into the eastern, central, and western regions, and then the panel fixed effect model is applied to the data. The estimated results are shown in

Table 6. Estimated results of promotion incentive and population mobility on medical and health expenditure.

Variable	PMexp		PEexp		PCexp		PSexp
	Eastern Region	Midwest Region	Eastern Region	Midwest Region	Eastern Region	Midwest Region	Eastern Region	Midwest Region
Pro1	−0.0822 ** (−2.31)	−0.0133 (−0.83)	−0.1183 *** (−3.98)	−0.0286 * (−1.91)	−0.1596 *** (−3.44)	−0.0023 (−0.09)	−0.0449 (−1.40)	−0.1365 *** (−7.21)
Pms	−3.4624 *** (−4.92)	−0.3010 (−0.40)	−1.8224 *** (−3.10)	0.1655 (0.24)	−2.7884 *** (−3.04)	−2.2264 * (−1.88)	−3.9033 *** (−6.16)	3.7472 *** (4.20)
Urb	0.8825 (1.06)	6.0593 *** (9.44)	1.5630 ** (2.24)	3.8120 *** (6.40)	−0.4429 (−0.41)	1.8328 * (1.83)	0.9731 (1.29)	5.5907 *** (7.40)
Pgdp	0.2579 *** (4.56)	0.0803 *** (4.06)	0.2254 *** (4.76)	0.1682 *** (9.16)	0.1878 ** (2.55)	0.1883 *** (6.10)	0.0721 (1.42)	0.0464 ** (1.99)
Den	4.5006 *** (5.50)	0.1109 (0.14)	2.8204 *** (4.12)	−0.3156 (−0.44)	4.0381 *** (3.78)	0.1752 (0.15)	4.6253 *** (6.27)	0.9789 (1.08)
Edu	0.1123 ** (2.06)	0.1553 *** (4.69)	0.1385 *** (3.04)	0.1473 *** (4.79)	0.2145 *** (3.02)	0.1848 *** (3.58)	0.1620 *** (3.30)	0.0638 (1.64)
Ind	3.4414 *** (6.46)	1.3282 *** (4.92)	0.9173 ** (2.06)	−0.1256 (−0.50)	1.8546 *** (2.67)	0.6444 (1.53)	4.5481 *** (9.48)	0.9001 *** (2.83)
Cons	−24.2593 *** (−5.42)	0.5882 (0.18)	−13.3032 *** (−3.55)	3.8790 (1.29)	−21.3889 *** (−3.67)	1.4778 (0.29)	−23.2591 *** (−5.77)	−4.4649 (−1.17)
Obs	99	171	99	171	99	171	99	171
Model	FE	FE	FE	FE	FE	FE	FE	FE

Note: ***, ** and * indicate that the significance level reaches 10%, 5% and 1%, respectively. The values in brackets are T. FE is the abbreviation of fixed effect model.

.

The regression results indicate that, with the exception of social security and employment expenditures, the promotion incentive has a considerable negative impact on all types of public service spending in the eastern region. In the central and western areas, the promotion incentive has a considerable negative impact on education expenditures, social security expenditures, and employment expenditures, but not on medical and health expenditures, cultural, sports, and media expenditures. The preceding results demonstrate that the role of the promotion incentive varies by area and public service category. Population mobility has a considerable negative impact on all types of public service expenditures in the eastern region, meaning that the rising share of floating population reduces expenditures on medical and healthcare, education, culture, sports and media, social security and employment. This conclusion is comparable to the findings of Yang [39]. The argument is that the comparatively developed eastern region has relatively flawless fundamental public services, and the surge of population has led to improvements in healthcare, education, culture, sports and the media. Different sorts of public service expenses in the central and western regions are affected differently by the increased share of mobile citizens. Population mobility has no substantial impact on medical and educational expenditures in the central and western areas, but has a negative impact on cultural, sports, and media expenditures and a positive impact on social security and employment expenditures.

The reason may be that the majority of the floating population belongs to vulnerable groups in the society, and their general characteristics are high mobility, poor employment conditions and low income level, which cause their most urgent needs to be in the fields of social security and employment, whereas their needs in the fields of culture, sports, and media are relatively slow to develop. To properly protect the rights and interests of the floating population, local administrations in the central and western areas have increased spending on social security and employment while delaying spending on culture, sports and the media. Due to space limitations, the regression coefficients of each controlled variable are not described in detail.

5. Discussion and Conclusions

Based on macro data at the provincial level, this work empirically investigates the influence of the promotion incentive and population mobility on public service expenditure using the panel fixed effect model. Previously, most studies focused solely on the impact of promotion incentive or population mobility on public service expenditure. This study examines the impact of the promotion incentive and population mobility on public service expenditure simultaneously. This paper tests the robustness by replacing the explained variables, the core explanatory variables, and the transformation model during the testing process, which confirm the conclusion that the promotion incentive and population mobility have a significant negative effect on public service expenditure, but there are still some limitations. First, this study merely discusses the influence of the promotion incentive and population mobility on public service expenditure from a theoretical and empirical standpoint, and the economic principles underlying it remain debatable. Second, this paper employs macro data at the provincial level, which is limited by the data itself. It fails to investigate the impact of officials’ promotion incentives and population mobility on public service expenditure using data from a smaller scale level (prefecture-level cities and county-level cities), a longer time span, and a micro level. It remains to be seen whether the research structure of this paper is still reliable. Third, this paper uses per capita foreign direct investment and opens up indicators to measure the promotion incentive. The selection of indicators is one-sided, which is not enough for the promotion incentive mechanism. We can build an indicator system of the promotion incentive from multiple dimensions such as economic growth, official age and whether to promote. Of course, other indicators, such as population mobility, also have similar problems.

Improving the level of public service expenditure is the basic condition to meet people’s growing yearning for a better life in the new era. Based on the panel data of 30 provinces (excluding Tibet, Hong Kong, Macao and Taiwan) in China from 2010 to 2018, this paper empirically tested the influence of the promotion incentive and population mobility on public service expenditure by using the fixed effect model. The conclusions are as follows: promotion incentive and population mobility had significant negative effects on public service expenditures (medical care, education, culture, sports and media, social security and employment). By changing the model and replacing the explained variables and core explanatory variables, the research conclusions are generally stable. The promotion incentive and population mobility had different effects on different types of public service expenditure in different regions. The urbanization level, economic development level, population density, residents’ educational level and the proportion of the tertiary industry had a significant positive impact on public service expenditure.

According to the conclusion of this paper, the following policy suggestions are put forward: on the one hand, a service-oriented government should be actively built. The government needs to establish service-oriented and people-oriented management concepts, continuously expand channels, build a platform for communication and interaction with the people, make more on-site visits and research, understand the sentiments of the people, listen to the opinions of the people, pay attention to the livelihood of the people and actively respond to the demands of the people. On the other hand, there should be a transfer of the core of performance appraisal to the overall regional strength evaluation. The central government can include livelihood indicators into the assessment criteria, and reduce the weight of economic performance to improve the incentive behavior of local government officials, so as to encourage local government officials to make more investment in social livelihood. At the same time, due to historical, political, geographical and other factors, various regions have shown different resource endowments and development conditions in the current society, and a single performance evaluation standard does not apply to all regions. Therefore, the central government needs to make the local leaders in power practice, investigate and analyze the local development, formulate practical work plans, implement them on time after being reviewed and approved by the superior government, and take the work progress assessment as a part of the performance assessment. This measure can not only weaken the pressure brought by competitors at the same level, but also bring real benefits to local development.