Economics
The impact of basic public services on residents' consumption in China
X. Xiong, X. Yu, et al.
Discover how basic public services influence consumption patterns in China in this insightful study by Xing Xiong, Xinghou Yu, and Yuxin Wang. The research reveals that enhanced public services not only boost consumption but also help bridge the gap between urban and rural areas, offering vital policy recommendations for equitable resource distribution.
~3 min • Beginner • English
Introduction
The paper addresses why household consumption in China remains relatively low despite substantial income growth and economic expansion, and examines whether improving the supply of basic public services can stimulate residents’ consumption and reduce the urban–rural consumption gap. Contextually, China’s growth model has relied on investment and exports, but faces diminishing returns and external headwinds (global instability and COVID-19). Household consumption has been persistently below 40% of GDP, far lower than in developed economies. Preventive savings due to uncertainty and unequal access to education, healthcare, and social security—stemming from unbalanced development—may suppress consumption. The study posits that improved public services can increase effective demand via income effects (reducing private outlays on education, health, social security, etc., thus raising disposable income) while possibly exerting a substitution effect (shifting spending toward publicly provided services). It aims to empirically identify the net effect on overall, urban, and rural consumption and on the urban–rural consumption gap, providing evidence relevant to Keynesian demand management and development policy.
Literature Review
The study situates itself within several strands of literature: (1) Keynesian fiscal policy and effective demand, emphasizing that public expenditure can stimulate aggregate demand when private demand is insufficient. (2) Preventive savings theory, which links high savings and low consumption to future uncertainty; enhanced social security and public services can reduce precautionary motives. (3) Equal opportunity theory (Roemer), which frames unequal access to education, healthcare, and social security as opportunity inequality; expanding basic public services can promote equality of opportunity and potentially raise consumption. The paper also references empirical works on public goods and consumption interactions (Fiorito & Kollintzas, 2004; Song, 2022), the role of public goods in welfare and income (Czyzewski et al., 2021), impacts of public goods provision on rural consumption (Yang & Sun, 2020), and methodological literature on TOPSIS and entropy weighting for multi-criteria evaluation (Olson, 2004; Tao & Feng, 2018; Ocampo et al., 2019). It contributes by providing evidence from a developing economy context with relatively low public service levels and low consumption shares, and by examining effects on the urban–rural consumption gap.
Methodology
Design: Panel-data econometric analysis using provincial-level data for 31 Chinese provinces from 2014 to 2019 (China Statistical Yearbook 2015–2020). The study constructs a composite index of basic public services (PS) via an entropy-weighted TOPSIS approach and estimates ordinary least squares (OLS) regressions to assess effects on consumption outcomes.
Public services index (PS): Built from five domains—public basic education; medical and health care; social security and employment services; public welfare infrastructure; public ecological environment; public cultural services. Secondary indicators include: student–staff ratios (primary, junior), government education funds per student; per capita local government spending on health, number of health technicians per 10,000, beds per 10,000; per capita spending on social security and employment, health insurance coverage, registered urban unemployment rate (negative indicator); per capita transport spending, highway density, per capita public transport vehicles; per capita environmental protection spending, harmless household waste treatment rate, green coverage in built-up areas; per capita culture/sports/media spending, per capita public library holdings, and cable TV subscription share. Indicator weights are computed annually via entropy method, then averaged over 2014–2019 for comparability (Table 2 shows domain weights; e.g., mean weight: public ecological environment ≈ 0.2525, public cultural services ≈ 0.1791, public basic education ≈ 0.176, etc.).
Entropy-weighted TOPSIS steps: (1) Align indicator directionality (reciprocal transform for negative indicators). (2) Standardize to form Z. (3) Compute information entropy e_j and utility d_j. (4) Derive weights w_j = d_j / sum d_j. (5) Form weighted decision matrix R. (6) Identify positive and negative ideal solutions S+ and S−. (7) Compute separations D_i+ and D_i− (Euclidean distances). (8) Calculate closeness C_i = D_i− / (D_i+ + D_i−), used as the PS score (closer to 1 indicates better services).
Variables:
- Dependent variables: CR (overall household consumption rate/level), URCR (urban residents’ consumption rate/level), RUCR (rural residents’ consumption rate/level), GAPCR (difference between urban and rural consumption rates; treated as a negative indicator—higher implies larger gap).
- Key explanatory variable: PS (composite public services index).
- Controls: IS (industrial structure, tertiary/secondary output ratio), UR (urbanization rate, %), PGDP (per capita GDP, yuan).
Models: Province-level OLS regressions:
(1) HC_it = α0 + α1 PS_it + Controls_it + ε_it
(2) UC_it = α0 + α1 PS_it + Controls_it + ε_it
(3) RC_it = α0 + α1 PS_it + Controls_it + ε_it
(4) URG_it = α0 + α1 PS_it + Controls_it + ε_it
Sample and descriptive statistics: N = 186 province-years (31 provinces × 6 years). PS mean 0.1637 (SD 0.1735; min 0.0275; max 0.9456), indicating low average public service levels and large regional disparity (highest ≈34.39× lowest). CR mean 0.7264; URCR mean 0.6912; RUCR mean 0.8240; GAPCR mean 0.4647. Controls show strong regional heterogeneity (PGDP max 153,095 yuan vs min 23,151 yuan).
Key Findings
- Public services promote overall household consumption: PS coefficient on CR = 0.0947 (p < 0.01).
- Stronger effect for urban residents: PS coefficient on URCR = 0.1348 (p < 0.01) versus rural residents: PS coefficient on RUCR = 0.0445 (p < 0.05).
- Public services reduce the urban–rural consumption gap: PS coefficient on GAPCR = −0.1340 (p < 0.01); since GAPCR is a negative indicator, this indicates a narrowing gap as PS improves.
- Controls:
- Industrial structure (IS) generally suppresses consumption for both urban and rural residents (negative and significant), while helping to narrow the urban–rural gap.
- Urbanization (UR) significantly promotes overall and urban consumption and contributes to narrowing the gap.
- Economic development (PGDP) significantly raises consumption for both urban and rural residents and narrows the gap.
- Model fit: R^2 approximately 0.696 (CR), 0.345 (URCR), 0.808 (RUCR), 0.377 (GAPCR).
- Descriptive context: Average public service level is low (mean PS = 0.1637), with pronounced east–west disparities in both PS and economic development.
- Mechanism interpretation: The net positive impact suggests that the income effect of public services (raising disposable income by reducing private outlays on education/health/social security, etc.) outweighs the substitution effect (shifting consumption toward lower-priced publicly provided services).
Discussion
The findings address the central question by demonstrating that enhancing basic public services is an effective lever to stimulate household consumption in China and to reduce the urban–rural consumption disparity. The larger elasticities for urban residents suggest heterogeneity consistent with differences in resource endowments, access, and income levels. The negative coefficient on the consumption gap implies that more equitable and accessible public services contribute to convergence in consumption behavior between urban and rural residents. These results align with Keynesian perspectives that public spending can bolster effective demand during periods of insufficient private consumption, and with theories suggesting that improved social insurance and public goods provision mitigate precautionary savings. The roles of industrial structure, urbanization, and per capita GDP underscore broader structural and developmental drivers: upgrading industrial structure may restrain immediate consumption yet supports gap reduction through employment, output variety, and demonstration effects; urbanization fosters agglomeration and scale effects that lift consumption and narrow gaps; economic growth increases purchasing power and diversifies consumption, though equitable distribution is important to ensure balanced gains. Overall, the results suggest that strategic public service investment is a viable policy mechanism to unlock consumption potential and support sustainable growth.
Conclusion
The study shows that improving the provision of basic public services significantly raises residents’ consumption in China, with a stronger effect for urban households than for rural households, and helps narrow the urban–rural consumption gap. Methodologically, it constructs a composite public service index using entropy-weighted TOPSIS and links it to consumption outcomes via panel OLS for 31 provinces (2014–2019). The work contributes to Keynesian and development economics by identifying public services as an effective tool to address insufficient demand and by offering evidence from a large developing economy. Future research should differentiate types of public services to assess their heterogeneous effects on consumption and explore mechanisms in greater depth, including distributional impacts and dynamic responses across regions and income groups.
Limitations
- The study treats basic public services as a composite index; it does not disaggregate and estimate the separate effects of specific service categories (education, health, social security, infrastructure, environment, culture), which the authors note may have different impacts on consumption.
- The analysis uses provincial-level panel data for 2014–2019, which may mask intra-provincial heterogeneity and limits temporal scope; causal identification beyond OLS associations is not pursued.
Related Publications
Explore these studies to deepen your understanding of the subject.

