Impact of Chronic Disease Shocks on Household Consumption in China: The Role of Multi-Level Health Insurance

Abstract

Expanding and enhancing the quality of consumption is crucial to satisfy the increasing demand for a higher standard of living. However, the rising risk of chronic diseases impacts not only the level of consumption but also impedes the optimization of consumption patterns. This research utilizes Staggered Difference-in-Differences (DID) and Difference-in-Differences-in-Differences (DDD) models to evaluate the effects of chronic disease shocks on both the levels and structures of household consumption, leveraging data from the China Family Panel Studies (CFPS) across four survey waves from 2012 to 2018. Additionally, the study examines the effects and mechanisms of the multi-level health insurance system in mitigating the impact of chronic disease shocks. The study discerns a significant decrease in household consumption expenditure, obstruction in the enhancement of consumption structures, and a constriction of consumption diversity attributable to chronic disease shocks. Heterogeneous analysis reveals that the impact of chronic shocks varies across factors such as age, capital endowment, urban versus rural residence, and the availability of healthcare resources. Moreover, the analysis highlights the heterogeneous mitigating effects of the multi-level health insurance system, particularly through diminishing out-of-pocket medical expenses and the propensity for precautionary savings. The findings accentuate the imperative for the enhancement and equalization of the multi-level health insurance system to fortify household consumption, offering significant policy insights for the realization of the “Healthy China” vision.

JEL Classification: P36, D12.

Keywords

chronic disease shocks consumption levels consumption structure multi-level health insurance system staggered DID method

Introduction

China’s rise as the world’s leading developing nation underscores the need for bolstering consumption growth and high-quality economic development. However, despite its economic progress, China’s savings rate as a percentage of GDP surpasses that of both developed and developing countries, indicating a potential shortfall in consumer demand, particularly in household consumption. The National Statistical Office reports a 0.2% decline in per capita household consumption expenditure in 2022, with significant drops in spending on education, culture, recreation, and clothing. Addressing this necessitates a strategy to expand consumption levels and upgrade structures, crucial for realizing the populace’s aspirations for improved living standards and developing a strong domestic market.

Health risks significantly influence consumption quality and expansion. Literature confirmed that health shocks can severely impact household consumption levels through increased medical costs, reduced labor capacity, and altered risk preferences (Cheng et al., 2019; Dong et al., 2021; Kolukuluri, 2023; Yaya Bocoum et al., 2018). However, the impact of health shocks on consumption structures and its underlying mechanisms requires further research, especially in China’s context of rapid aging and rising chronic disease prevalence. This demographic shift is linked to a surge in medical costs and labor force implications, necessitating effective risk-sharing mechanisms to prevent economic hardship (Alefan et al., 2019).

China is advancing a multi-level health insurance framework which encompasses basic medical insurance, supplemental medical insurance, commercial health insurance, and various other insurance mechanisms, each offering different levels of coverage to better meet diversified medical insurance needs and address health challenges. Despite this, the system’s design variations result in inconsistent impacts on household consumption, particularly following chronic disease shocks. However, the existing literature does not provide comprehensive analysis of the mitigating effects of different health insurance systems.

This study leverages data from the China Family Panel Studies (CFPS) spanning the years 2012 to 2018 and employs a DID approach to investigate the causal relationship between chronic disease shocks and household consumption levels. It also evaluates the multi-level health insurance system’s potential to mitigate these shocks, contributing to three areas:

(1) Expanding the investigation into the household welfare effects of chronic disease shocks. While existing literature primarily addressed the extensive marginal effect of whether a chronic disease shock occurred, this paper assesses both the extensive marginal effect and the intensive marginal effect (quantified by the number of household members affected) of chronic disease shocks, enhancing the understanding of the micro-level impacts of chronic disease shocks.

(2) Offering a comprehensive depiction of household consumption characteristics. The paper enriches the delineation of household consumption characteristics across three dimensions: The level of consumption expenditure, the evolution of consumption structure, and the diversity of consumption, offering detailed evidence to address the challenges of constrained household consumption due to aging and health issues.

(3) Elucidating the heterogeneous risk-buffering effects of the multi-level health insurance system and its components. The study conducts a comprehensive evaluation of the multi-level system’s effectiveness in mitigating the impact of chronic disease shocks, providing valuable insights for the ongoing improvement of health insurance policies.

Literature Review and Hypothesis

Chronic Disease Shocks and Household Consumption Behavior

Divergencies existed in the literature regarding the impact of health shocks on household consumption, with some studies suggesting no significant effect, positing that households can use intra-household labor reallocation, borrowing, and asset liquidation against such risks (Yilma et al., 2021). Others, however, reported adverse impacts, where the severity of health issues is negatively associated with consumption levels and self-insurance capacity (Dong et al., 2021; Kolukuluri, 2023; Han et al., 2024). Babiarz and Yilmazer (2017) specifically highlighted psychiatric health shocks leading to substantial consumption declines due to reduced labor income. Yet, the literature was scarce on how health shocks influence consumption structure. Cheng et al. (2019) noted a marked drop in non-medical expenditures post-health shock, particularly in discretionary areas like entertainment. Additionally, health shocks were found to shift consumption to online essential categories (Dong et al., 2021). Further research is needed to understand the micro-level effects of health shocks on consumption structure, particularly within the changing dynamics of expenditure patterns.

The aging population’s rise in chronic diseases poses substantial public health challenges, influencing healthcare costs, labor participation, income, and economic stability. Studies, such as Liu et al. (2020), revealed that chronic diseases increase healthcare expenditures by 13% for older adults, heightening the risk of financial catastrophe. These conditions also decrease labor force involvement, leading to income loss and early retirement, thereby restricting the labor supply (Qiao et al., 2021). Chronic diseases intensify work-caregiving conflicts, further impacting labor participation and income (Anand et al., 2022; Eriksen et al., 2021). Moreover, chronic health conditions are also a significant catalyst for poverty or increased vulnerability, with their financial impact being four times greater than general health shocks for middle-aged and elderly households (Zhang et al., 2022). Yet, research examining the comprehensive effects of chronic health shocks on consumption quality remains scarce.

Chronic disease shocks diminish household consumption by increasing medical expenses, constraining budgets, and decreasing labor participation and income. They also provoke emotional and cognitive responses, prompting precautionary savings and reduced consumption intentions. Kolukuluri (2023) found a 2% drop in non-food consumption among affected households, highlighting the need for stable income to mitigate the effects of chronic diseases. The burden of chronic illnesses significantly reduces the economic resources available to households for consumption, leading to a reduction in expenditures on subsistence, development, and enjoyment goods. In response to these shocks, households often curtail the diversity of their consumption, particularly in non-essential categories, while continuing to meet essential expenditures on items such as food and housing. This strategic reduction can lead to a pronounced decline in development and enjoyment consumption, thereby reducing the overall variety of consumption and contributing to the downgrading of household consumption patterns. Based on the preceding discussion, this research postulates the following:

Hypothesis 1. Chronic disease shocks have a substantial impact on household consumption by significantly reducing consumption levels, hindering the optimization of consumption structures, and constraining the diversity of consumption.

The influence of chronic disease shocks on consumption varying by individual and regional factors. In terms of age, middle-aged and elderly adults, due to declining health, face higher chronic disease prevalence, increased healthcare needs, and economic challenges from reduced labor participation and income (Zhang et al., 2022), thereby highlighting age-related heterogeneity in the consumption impact of chronic disease shocks. Secondly, household capital, including physical, human, and social resources, is pivotal for managing such risks (Chen & Wang, 2023). Physical capital supports living standards and productivity, human capital enhances health and resilience, and social capital aids in accessing post-shock resources. Thus, the impact of chronic disease shocks on consumption is likely to vary based on a household’s capital endowment. The urban-rural divide also plays a role, with rural areas experiencing a higher elderly population and less access to healthcare and insurance (Jiang et al., 2019). Moreover, certain labor groups may lose income-generating opportunities when they return to their hometowns to care for ill family members, making them less equipped to manage the effects of chronic illnesses. Additionally, the uneven distribution of healthcare resources across China can also intensify the negative effects of chronic diseases on consumption, particularly in areas with limited medical services (Feng et al., 2022). This study hypothesizes that the consumption impact of chronic diseases will differ based on healthcare resource accessibility. Based on the preceding discussion, this research postulates the following:

Hypothesis 2. The impact of chronic disease shocks on consumption demonstrates heterogeneity across households with diverse age groups, capital endowments, urban and rural residences, and access to medical resources.

The Moderating Effect of Health Insurance on Chronic Disease Shocks

Health insurance is recognized for its role in dampening the volatility in household consumption due to health shocks. It stabilizes consumption by increasing household income through improved labor supply and health status, thus encouraging human capital investment and reducing poverty (Fadlon & Nielsen, 2019; Liao et al., 2022). Moreover, it alleviates economic pressures by decreasing out-of-pocket medical expenses, despite debates on its extent (Jiang et al., 2019; Han & Song, 2024). By converting uncertain future healthcare expenses into predictable current costs, health insurance also reduces the need for precautionary savings, as argued by Zhao (2019), making it superior to household savings in smoothing consumption. Despite its benefits, the nuanced link between health insurance and consumption post-health shocks has been understudied. This paper posits that health insurance could play a protective role against consumption declines following chronic disease shocks, a gap in the current literature.

China’s fragmented health insurance system, with variations in contribution levels, reimbursement, and coverage, impacts household consumption patterns differently. For instance, Urban Employees Basic Medical Insurance (UEBMI) increases overall consumption, especially in medical and discretionary spending (Shao & Hao, 2020). The New Rural Cooperative (NRC) has demonstrated efficacy in facilitating long-term consumption smoothing among participating households (Wang et al., 2023), but its integration with Urban Residents Basic Medical Insurance (URBMI) shows no significant effect on healthcare consumption for urban and wealthier groups (C. Li et al., 2019). The supplemental insurance’s impact on consumption is less studied; however, Major Medical Insurance, which automatically enrolled URBMI and NRC participants in 2012, is seen to reduce medical costs and consumption risk (Jiang et al., 2019; Zhao, 2019), although C. Li et al. (2019) found it could increase medical expenses and negatively impact consumption. Commercial health insurance is recognized for alleviating financial stress and safeguarding post-illness consumption (Dong et al., 2021; Vo & Van, 2019). The literature predominantly assessed the performance of specific health insurance systems, indicating a need for further research to understand the multifaceted effects of China’s multi-level health insurance system on mitigating health shocks. Based on this, this research proposes the following:

Hypothesis 3. Multi-level health insurance aims to stabilize household consumption amidst chronic disease shocks by easing medical financial burdens and reducing the need for precautionary savings. Yet, the effectiveness of different insurance schemes in mitigating these shocks varies.

Data Variables and Models

Sample Data

This study utilizes the China Family Panel Studies (CFPS) database from 2012, 2014, 2016, and 2018 as its research sample. CFPS, administered by Peking University, constitutes a comprehensive and nationwide longitudinal survey, providing a rich dataset for analysis on a wide array of social phenomena in China (Xie & Hu, 2014). The initial survey, conducted in 2010, comprised 14,960 households and 42,590 individuals across 25 provinces in China, with follow-up surveys occurring biennially. The CFPS collects a wide range of information, including economic activity, insurance participation, family dynamics, population migration, and physical and mental health, offering a robust foundation for our research. The data processing procedure is as follows: Initially, “financial respondents” are identified as household heads, merges heads information with household data across four survey waves to construct a balanced panel dataset. Samples with zero or negative values for consumption, income, or asset holdings are excluded. Subsequently, a 5% trim is applied to the consumption distribution’s tails to minimize the impact of outliers. The analysis is further refined by excluding households where the head was under 16 or over 85 years of age. Consistent with Q. Li et al. (2022), the study focuses on households without chronic illnesses throughout 2012 to 2018, or those where such conditions emerged for the first time within this period, thus excluding those with continuous chronic illness reports or initial reports not followed by subsequent ones. The final study sample comprised 17,376 individuals across 4,344 households.

Variables

Dependent Variable

This study examines household consumption by assessing both the level and structure of consumption. The consumption level is operationalized as the natural logarithm of total annual household expenditure, with medical expenses excluded. The consumption structure is characterized by two metrics: the structural upgrading index and the diversity of consumption. The upgrading index, following Sun et al. (2022), is determined by the proportion of development and enjoyment consumption (share of transportation, communication, healthcare, cultural and educational, household equipment, and other goods and services). Refer to Yang et al. (2021), diversity of consumption is measured by the variety of goods purchased across nine categories, including food, clothing, housing, daily necessities, transportation and communication, education, travel, beauty and personal care, and entertainment. The diversity, ranging from 0 to 9, reflects the extent of the consumption spectrum within a household, where a higher score denotes greater diversity.

Independent Variable

This study examines the distinct effects of the extensive and intensive margins of chronic disease shock on household consumption. Following the approach of E. Liu et al. (2020), the extensive margin is operationalized as a binary variable, where it is assigned a 1 for households with new chronic disease diagnoses during the study period, and 0 for those without new diagnoses. In contrast, the intensive margin is quantified by the total number of household members affected by chronic diseases, with a scale from 0 to 3. For households with three or more members afflicted, the count is capped at 3.

Mediating Variables

The mediating variables is household participation in health insurance. Participation in health insurance is a dummy variable defined as the presence of at least one individual in the household enrolled in any of the employee health insurance, resident health insurance, supplemental health insurance, or commercial health insurance. Furthermore, participation in basic health insurance is characterized by the involvement of at least one member in either employee health insurance or resident health insurance.

Control Variables

Drawing from Zhang et al. (2022) and Q. Li et al. (2022), the analysis controls for a spectrum of individual and household characteristics. Individual attributes include gender, age, and the square of age divided by 100 for the household head to account for the non-linear relationship between age and consumption. Marital status, educational level, urban and rural domicile, and the health status of the head at the study’s onset are also considered. Household characteristics encompass family size, annual income, property value, financial asset value, and pension insurance participation. To address the skewed distribution, annual income, and asset values are log-transformed after a constant of 1 is added. Furthermore, nominal variables are adjusted using the 2012 provincial price index to maintain temporal consistency.

Table 1 provides descriptive statistics, revealing the distinct attributes of households subjected to chronic disease shocks (referred to as the treatment group). These households show significantly reduced consumption expenditure levels, diminished consumption structure upgrading, and a lower diversity of consumption, underscoring the adverse impact of chronic disease shocks on household consumption. The treatment group also exhibits higher participation in health insurance, notably employee health insurance, with no significant variation in other insurance types. The average age of the head in the treatment group is higher, with a higher proportion of females and a greater rural representation. There are no significant differences in education, marital status, or health self-assessment at the study’s beginning. While the treatment group has lower financial assets, there is no significant disparity in family size, annual income, property value, or pension insurance.

Table 1.

Descriptive Statistics.

Variables	Description	Treatment group	Comparison group	Diff between groups
cons	The logarithm of the level of consumption expenditure	9.712	9.932	−0.022***
upgrade	The ratio of developmental and enjoyment consumption	0.329	0.367	−0.038***
diversity	Number of consumer goods categories involved in the household	5.632	5.698	−0.066**
insu	Participation in health insurance, yes is assigned a value of 1, otherwise 0	0.943	0.922	0.021**
bas	Participation in basic health insurance, yes is assigned a value of 1, otherwise 0	0.918	0.918	0.000
emp	Participation in employee health insurance, yes is assigned a value of 1, otherwise 0	0.165	0.156	0.009*
res	Participation in residential health insurance, yes is assigned a value of 1, otherwise 0	0.774	0.762	0.012
sup	Participation in supplementary health insurance, yes is assigned a value of 1, otherwise 0	0.012	0.012	0.000
age	Age at the time of the interview	55.970	47.933	8.037***
age ²	Age squared/100	31.326	22.976	8.350***
edu	Educational degree is assigned a value of 0 to 7 in ascending order	1.860	1.882	−0.022
gender	Males are assigned a value of 1 and females 0	0.484	0.538	−0.054*
marr	Assigns a value of 1 at the wedding, otherwise 0	0.873	0.884	−0.011
urban	Rural domicile is assigned a value of 1, otherwise 0	0.429	0.330	0.099**
health	Self-assessed health status at the beginning of the period is assigned a value of 1 to 7 in ascending order.	5.466	5.660	−0.194
familysize	Number of persons in the household	3.627	3.688	−0.061
income	Logarithmic annual household income	10.364	10.393	−0.029
property	Logarithmic household property values	2.525	2.538	−0.013
finance	Logarithmic household financial assets values	0.546	0.644	−0.098**
pension	Availability of old-age pension, yes is assigned a value of 1, otherwise 0	0.462	0.531	−0.069

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Empirical Models

The study initiates an evaluation of the influence that chronic disease shocks have on household consumption by employing the DID model. Widely recognized in the social sciences, the DID model is particularly adept at assessing the effects of interventions within quasi-experimental frameworks. It offers a robust approach to causal inference by comparing the pre- and post-treatment changes in both a treatment group and a control group, thereby controlling for time-invariant unobserved confounders. The model’s strength is its ability to provide a more nuanced understanding of causal relationships by accounting for these persistent unobserved factors. Considering the time difference in the onset of chronic diseases across households and to evaluate its impact on household consumption scientifically, following the studies of Athey and Imbens (2022), Xiao et al. (2023), a staggered DID model is constructed:

\begin{matrix} con s_{ijt} = α_{0} + α_{1} chr o_{ijt} + α_{2} X_{ijt} + λ_{i} + η_{t} + δ hea l_{ij 1} \times t \\ + ς cou n_{j} \times t + {ε_{ijt}}^{1} \end{matrix}

(1)

\begin{matrix} con s_{ijt} = β_{0} + β_{1} chronu m_{ijt} + β_{2} X_{ijt} + λ_{i} + η_{t} + ψ hea l_{ij 1} \times t \\ + φ cou n_{j} \times t + {ε_{ijt}}^{2} \end{matrix}

(2)

In Equation 1, cons_ijt encapsulates the consumption status of the ith household in county j during period t. This encompasses the logarithm of consumption level, the structural upgrading index, and consumption diversity. The independent variable chro_ijt serves as an indicator denoting whether the household is afflicted by a chronic disease, with the coefficient α₁ quantifying the extensive marginal impact of a chronic disease shock on household consumption. The control variable X_ijt incorporates individual and household characteristics. In Equation 2, the independent variable chronum_ijt signifies the count of members experiencing chronic diseases, and the coefficient β₁ gages the intensive marginal impact of chronic disease shocks on consumption. Robust standard errors are adjusted for clustering at the household level.

To alleviate omitted variable bias, two key strategies are employed. First, incorporate household fixed effects (λ_i,) to account for time-invariant household attributes and introduce year fixed effects (η_t) to control for macroeconomic variations or unforeseen shocks that can influence the overall consumption levels. Second, include an interaction term between the initial health status of the head and the temporal trend to mitigate the potential distortion caused by the predetermined health condition of individuals and introduce an interaction term between the county-level dummy variable and time to address potential time-varying impacts of county-level characteristics on household consumption.

To assess the efficacy of multi-level health insurance in alleviating the impact of chronic disease shocks on household consumption, a staggered DDD model is employed. A distinctive aspect of the DDD model is its enhancement of the traditional DID framework by incorporating an additional dimension of comparison. This extension allows for a more refined estimation of policy impacts, as it accounts for both time-invariant unobserved confounders and factors that fluctuate over time. By analyzing consumption patterns among households with varying levels of health insurance coverage before and after experiencing chronic disease shocks, the study can more precisely determine the extent to which health insurance acts as a buffer for household consumption. The model is structured as follows:

\begin{matrix} con s_{ijt} = θ_{0} + θ_{1} chr o_{ijt} \times ins u_{ijt} + θ_{2} chr o_{ijt} + θ_{3} ins u_{ijt} \\ \times pos t_{ijt} + θ_{4} ins u_{ijt} + θ_{5} X_{ijt} + ω hea l_{ij 1} \\ \times t + λ_{i} + η_{t} + {ε_{ijt}}^{3} \end{matrix}

(3)

In Equation 3, insu_ijt denotes whether the ith household in county j participates in health insurance in period t. The parameter θ₁ scrutinizes the mitigating effect of chronic disease shocks within the multi-level health insurance system. The variable post_ijt takes the value of 1 during the period when the ith household in county j experiences a chronic disease shock and in subsequent periods, and vice versa. Other variables remain consistent with those in equations 1 and 2. The rationale for easing the control of coun_j× t is the variability of employee health insurance and resident health insurance is primarily at the county level (C. Li et al., 2019), and excessive control may impact the assessment of the effectiveness of health insurance.

Empirical Result Analysis

The Impact of Chronic Disease Shocks on Household Consumption Levels

Table 2 displays the regression results on the impact of chronic disease shocks on household consumption levels. Columns (1)–(3) examine the extensive marginal effects of chronic disease shocks, progressively incorporating individual and household characteristics while controlling for household and time fixed effects. The coefficients of chro_ijt are consistently and significantly negative at the 5% significance level, signifying a substantial adverse effect of chronic illness shocks on household consumption levels. Sensitivity tests, which incrementally add control variables, confirm the robustness of these findings. Columns (4)–(6) explore the intensive margin effects, revealing that an increased number of household members affected by chronic disease shocks is associated with a significant decrease in total consumption expenditure.

Table 2.

Impact of Chronic Disease Shocks on Household Consumption Levels.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
chro_ijt	−0.053**	−0.048**	−0.045**
	(0.022)	(0.022)	(0.021)
chronum_ijt				−0.058**	−0.051**	−0.049**
				(0.023)	(0.025)	(0.023)
age		0.021**	0.009		0.022**	0.013
		(0.010)	(0.012)		(0.010)	(0.012)
age²		−0.023**	−0.013**		−0.019*	−0.014**
		(0.011)	(0.006)		(0.011)	(0.007)
gender		0.017	0.037		0.008	0.024
		(0.029)	(0.031)		(0.021)	(0.022)
edu		0.017	0.005		0.013	0.011
		(0.011)	(0.011)		(0.011)	(0.007)
marr		0.172***	0.051		0.123**	0.010
		(0.062)	(0.072)		(0.055)	(0.009)
urban			0.009			0.012
			(0.016)			(0.015)
familysize			0.085***			0.077***
			(0.015)			(0.020)
income			0.181***			0.201***
			(0.001)			(0.001)
property			0.078***			0.072***
			(0.011)			(0.011)
finance			0.048***			0.047***
			(0.009)			(0.009)
pension			0.049			0.034
			(0.033)			(0.027)
insu			0.026			0.039
			(0.057)			(0.045)
heal_ij1× t	YES	YES	YES	YES	YES	YES
coun_j× t	YES	YES	YES	YES	YES	YES
FamilyFE	YES	YES	YES	YES	YES	YES
YearFE	YES	YES	YES	YES	YES	YES
N	17,376	17,376	16,208	17,376	16,444	15,036
R²	0.114	0.235	0.338	0.103	0.211	0.232

Note. The values in brackets are robust standard errors after clustering at the household level.

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Regression results on the control variables show that the linear term for age is positively correlated with household consumption, while the quadratic term demonstrates an inverse relationship, reaching statistical significance at the 5% level. This suggests a consumption pattern that aligns with life cycle theory, with peaks during middle age and subsequent declines. Furthermore, family size, measured by the number of members, is robustly associated with higher consumption levels, with a significant correlation at the 1% level. Financial assets and property values are also found to be positively and significantly associated with consumption levels, with both associations being statistically significant at the 1% level. This underscores the critical role of wealth in cushioning the effects of chronic disease shocks and its function as a stabilizer for household consumption. Incorporating these control variables into the regression model ensures that the observed effects of chronic disease shocks are not confounded by other concurrent socioeconomic factors, providing a clearer understanding of their impact.

The Impact of Chronic Disease Shocks on Household Consumption Structure

Table 3 provides empirical evidence on the effects of chronic disease shocks on household consumption structure. Columns (1) and (2), addressing the extensive margin effect, show that households facing chronic disease shocks significantly decrease their spending on development and enjoyment goods and reduce consumption diversity at the 1% significance level. This implies a noteworthy hindrance caused by chronic disease shocks in elevating the household consumption structure and curbing consumption diversity. Columns (3) and (4) investigate the intensive margin effect, demonstrating that an increase in the number of household members with chronic diseases leads to a significant decline in both consumption structure upgrading and diversity at the 1% significance level. This suggests that the negative impact on consumption quality is not only significant but also worsens with the severity of the chronic disease shock, as indicated by the number of affected members, thereby validating Hypothesis 1.

Table 3.

Impact of Chronic Disease Shocks on Household Consumption Structures.

	(1)	(2)	(3)	(4)
Variable	Consumption upgrading	Consumption diversity	Consumption upgrading	Consumption diversity
chro_ijt	−0.028***	−0.036***
	(0.007)	(0.006)
chronum_ijt			−0.035***	−0.041***
			(0.009)	(0.003)
Controls	YES	YES	YES	YES
FamilyFE	YES	YES	YES	YES
YaerFE	YES	YES	YES	YES
N	16,208	16,208	15,036	15,036
R²	0.247	0.258	0.217	0.214

Note. The values in brackets are robust standard errors after clustering at the household level.

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Heterogeneity Test

Family Heterogeneity Perspective

The study initially examines age-level heterogeneity in the context of chronic disease shock. While the incidence of chronic diseases has increased across all age groups, the nature of the disease and its economic impact vary significantly by age. The sample was thus categorized into four distinct age groups for analysis: under 30 years, 31 to 45 years, and 46 to 60 years, and over 61 years (Wu et al., 2021). Table 4a presents the findings, which show that both the extensive and intensive margins of chronic disease shocks have negative regression coefficients across different age groups, affecting consumption levels, consumption structure upgrades, and consumption diversity. Especially noteworthy is the more pronounced adverse impact of chronic disease shocks on consumption levels and quality among middle-aged and elderly households (46–60 years and over 61 years) relative to younger households, all sub-sample regressions passed the inter-group coefficient difference test. using the bootstraps inter-group difference test. This difference is likely due to the decline in physical function with age, increasing the incidence of chronic diseases, and the need for medical services in the older age groups. Furthermore, the reduced labor supply and employment income among the elderly lead to greater financial vulnerability and a constrained capacity to finance healthcare needs (Zhang et al., 2022). As a result, the economic effects of chronic disease shocks on consumption are more pronounced among middle-aged and older populations.

Table 4.

Heterogeneity Analysis of the Impact of Chronic Disease Shocks.

	(1)	(2)	(3)	(4)	(5)	(6)
	Whether or not experiencing a chronic disease			Number of family members affected by chronic diseases
Variable	Consumption level	Consumption upgrading	Consumption diversity	Consumption level	Consumption upgrading	Consumption diversity
a. Family heterogeneity perspective
≤30 years old	−0.013*	−0.008**	−0.024**	−0.014**	−0.011***	−0.027**
	(0.007)	(0.004)	(0.011)	(0.007)	(0.004)	(0.012)
31–45 years old	−0.020*	−0.021*	−0.029**	−0.045***	−0.006**	−0.028*
	(0.011)	(0.011)	(0.012)	(0.008)	(0.003)	(0.015)
46–60 years old	−0.041***	−0.038***	−0.045***	−0.053***	−0.047**	−0.046***
	(0.012)	(0.014)	(0.014)	(0.014)	(0.021)	(0.012)
≥61 years old	−0.047***	−0.039***	−0.055***	−0.052***	−0.015***	−0.071***
	(0.014)	(0.013)	(0.015)	(0.013)	(0.004)	(0.018)
chro_ijt× capi_ijt	−0.032**	−0.022**	−0.028***	−0.029**	−0.021*	−0.023*
	(0.016)	(0.011)	(0.009)	(0.013)	(0.011)	(0.013)
chro_ijt× huma_ijt	−0.024**	−0.019*	−0.021***	−0.023**	−0.017*	−0.020*
	(0.011)	(0.011)	(0.008)	(0.011)	(0.010)	(0.011)
chro_ijt× soci_ijt	−0.017**	−0.015	−0.020**	−0.019*	−0.016	−0.018**
	(0.007)	(0.011)	(0.009)	(0.011)	(0.012)	(0.009)
b. Regional heterogeneity perspective
rural	−0.038***	−0.044***	−0.042***	−0.041**	−0.047**	−0.044***
	(0.012)	(0.013)	(0.015)	(0.018)	(0.019)	(0.017)
urban	0.013	0.007	0.010	0.015	0.012	0.009
	(0.022)	(0.025)	(0.019)	(0.013)	(0.021)	(0.014)
chro_ijt× pcost_ijt	0.022**	0.023**	0.014***	0.021**	0.020*	0.013***
	(0.011)	(0.011)	(0.005)	(0.010)	(0.011)	(0.005)

Note. The values in brackets are robust standard errors after clustering at the household level.

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Second, an analysis of capital endowment heterogeneity is conducted. Physical, human, and social capital serve as crucial pillars for household resilience against risks, three dummy variables (capi, huma, soci) are established based on whether the sum of property and financial assets, annual social relationship expenditure, and the education level of the household head fall below the median of these variables for the entire sample of households in their region. These dummy variables represent low physical, low human, and low social capital, respectively, allowing an examination of the effects of their interaction terms with chronic disease shocks on both the level and structure of consumption. The outcomes, presented in Table 4a, indicate that households with low physical, human, and social capital endowments experience a further decline in household consumption expenditures, consumption structure upgrading, and consumption diversity at the time of a chronic disease shock.

Regional Heterogeneity Perspective

China’s economic landscape is marked by a pronounced imbalance in economic development between urban and rural regions. Table 4b presents a test for heterogeneity, differentiating the impact of chronic disease shocks on consumption based on the household place of residence. The findings indicate that rural residents, in contrast to their urban counterparts, experience a significant decline in consumption levels, a degradation of consumption structure, and a reduction in consumption diversity after a chronic disease shock. This discrepancy is attributed to several factors: a higher proportion of the elderly population in rural areas, which correlates with a greater prevalence of chronic diseases; increased exposure to uncertainties regarding employment, income, and social security; and a heightened vulnerability of rural households. The significant disparity in social security and public service supply between urban and rural areas means that rural families have less access to medical insurance benefits, leading to a diminished capacity to manage risks associated with chronic diseases. Thus, rural households are more susceptible to the adverse economic effects of chronic disease shocks. To mitigate the risk of poverty induced by chronic diseases in rural areas, it is recommended that efforts be made to equalize access to high-quality healthcare and social security for both urban and rural residents. Additionally, targeting chronic disease prevention and treatment initiatives in rural areas could be beneficial.

Access to healthcare resources is vital for the preservation of human capital, and the uneven distribution of these resources across China’s regions is a realistic issue. Table 4b shows a significantly positive coefficient for the interaction between medical service supply levels and chronic disease shocks. This suggests that households with better medical service access experience a mitigated impact on consumption levels, structural upgrading, and diversity when facing chronic disease shocks. Conversely, households with limited medical service supply suffer a more severe effect on consumption quantity and quality. Therefore, it is imperative to optimize the allocation of medical resources and improve the standards of healthcare services in underserved areas and primary care units to reduce the risks posed by chronic diseases. The findings from the heterogeneity test, which robustly confirm Hypothesis 2, indicate that the effects of chronic disease shocks on consumption are heterogeneous, varying by age, capital endowments, urban and rural location, and the availability of medical resources.

Robustness Test

To ensure the reliability of the findings, several robustness tests have been conducted, utilizing the extensive marginal effects of chronic disease shocks as a case in point.

Goodman-Bacon Decomposition

This study confronts the issue of DID estimator bias due to the staggered onset of chronic disease shocks among households by applying the Goodman-Bacon (2021) decomposition technique. The research first determines weights for three comparative treatment effects: the initial group experiencing the shock against the unaffected group, the subsequent group experiencing the shock against the unaffected group, and the comparison between the two affected groups. We then analyze the DID estimates for bias by focusing on the weights and coefficients of the treatment effect between the groups that experienced the shock at different times. As shown in Table 5, these weights and coefficients are 27.1% and 0.004, respectively, markedly lower than those from the other two group comparisons, indicating that the potential bias in our DID estimation is not significant.

Table 5.

Goodman-Bacon Decomposition.

Groups	Weights	Treatment effect
Shock-first versus never-shocked groups	0.355	−0.056
Shock-second versus never-shocked groups	0.374	−0.063
Shock-first versus shocked-second groups	0.271	0.004

Parallel Trend Test

The validity of the DID approach hinges on the premise that, after controlling for factors that influence both household consumption and the incidence of chronic disease shocks, the treatment and comparison groups would follow congruent consumption paths in the absence of shocks. It is plausible to assume that if the consumption trends between the groups do not significantly diverge before the onset of chronic disease, they should remain parallel after the shock. To examine this, the article utilizes an event study method to assess the dynamic effects of chronic disease shocks. The model is formulated as follows:

\begin{matrix} con s_{ijt} = α_{0} + α_{1} \sum_{k = - 3}^{3} chr o_{ijt - k} + α_{2} X_{ijt} + λ_{i} + η_{t} \\ + δ hea l_{ij 1} \times t + ς cou n_{j} \times t + {ε_{ijt}}^{4} \end{matrix}

(4)

In Equation 4, the satisfaction of k = 0 indicates that the household experiences a chronic disease shock in the current period. Negative values of k denote periods prior to the shock, while positive values indicate subsequent periods. The regression analysis, based on Equation 4, with 95% confidence intervals, is presented in Figures 1 to 3. Estimates for consumption expenditure, consumption structure upgrading, and consumption diversity before the shock are statistically insignificant, with confidence intervals centered around zero, indicating no pre-shock difference in consumption levels and structures between the treatment and comparison groups. Post-shock, the coefficients are significantly negative, confirming that the shock negatively impacts both consumption levels and structures, with this effect persisting for three periods afterward, aligning with the baseline regression analysis.

Figure 1.

Parallel trend tests for household consumption expenditures.

Figure 2.

Parallel trend tests for consumption structural upgrading.

Figure 3.

Parallel trend tests for household consumption diversity.

Placebo Test

This study utilizes a nonparametric replacement test to counteract the possible influence of other policy shocks on consumption. The test randomly determines the incidence of chronic disease shocks within the sample households, with the randomization repeated 500 times to establish multiple sets of treatment and comparison groups. Figure 4 shows that the majority of coefficients from the 500 resultant regressions are centered around zero, with a near-zero likelihood of being beneath the coefficient from the benchmark regression. This substantiates the significant negative impact of chronic disease on household consumption levels identified in the benchmark analysis.

Figure 4.

Replacement test.

Other Robustness Test

Building on the robustness tests previously discussed, this paper further validates its findings through several additional checks: (1) by replacing the dependent variables, (2) employing the PSM-DID method to correct for potential sample selection bias, and (3) by modifying the comparison group. The household consumption level is redefined using total annual expenditures, with medical and healthcare costs deducted, as a proxy. The regression results, presented in Table 6, column (1), show a significant negative impact of chronic disease shocks on consumption levels. For the consumption structure, the household Engel’s index—measuring the share of food expenditure—is used as an inverse indicator of consumption upgrading; the significant positive impact at the 1% level, as shown in column (2), confirms the adverse effect on consumption structure optimization.

Table 6.

Other Robustness Test.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Replace dep var		PSM-DID			Change comparison group
Variable	Consumption level 2	Engel’s index	Consumption level	Consumption upgrading	Consumption diversity	Consumption level	Consumption upgrading	Consumption diversity
chro_ijt	−0.045**	0.020***	−0.058**	−0.030***	−0.026***
	(0.021)	(0.007)	(0.026)	(0.011)	(0.005)
chro_test_ijt						−0.043***	−0.029***	−0.027***
						(0.009)	(0.011)	(0.005)
Controls	YES	YES	YES	YES	YES	YES	YES	YES
FamilyFE	YES	YES	YES	YES	YES	YES	YES	YES
YaerFE	YES	YES	YES	YES	YES	YES	YES	YES
N	16,208	16,208	13,728	13,728	13,728	12,824	12,824	12,824
R ²	0.247	0.236	0.258	0.214	0.268	0.217	0.213	0.214

Note. The values in brackets are robust standard errors after clustering at the household level.

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Next, the PSM-DID method is applied using head and household-level matching variables, followed by 1:1 kernel matching for sample selection. The DID estimation on these samples, detailed in columns (3)–(5) of Table 6, reveals significant reductions in consumption levels at the 5% level and marked hindrance to consumption structure upgrading and diversity at the 1% level. Furthermore, by substituting the never-shocked sample with one that has consistently experienced shocks to create a new comparison group, the regression results in columns (6)–(8) of Table 6 corroborate the benchmark findings, with all coefficients for the variable chro_test being significantly negative.

Further Analysis: The Moderating Effect of Multi-Level Health Insurance

This section first examines the heterogeneous effects of multi-level health insurance in alleviating the impacts of health shocks on households by using a staggered DDD model. This is followed by an in-depth analysis of the mechanisms through which various health insurance schemes function to reduce the consumption fluctuations triggered by health shocks.

The Impact of Health Insurance on Consumption Level

This study examines the impact of multi-level health insurance on consumption levels during chronic disease shocks by introducing an interaction term between health insurance status and chronic disease occurrence. The results, detailed in Table 7, show that health insurance generally mitigates the decline in consumption expenditure caused by chronic disease shocks, as seen in column (1). Columns (2) and (6) reveal that both basic and commercial health insurance significantly alleviate the adverse effects of health shocks on consumption levels. However, column (5) indicates that supplemental health insurance does not effectively maintain consumption levels post-shock. This suggests that different health insurance models have heterogeneous effects on consumption level preservation during chronic disease shocks. The minimal perceived risk protection from supplemental insurance may be due to its automatic or enterprise-managed enrollment, leading to a lack of direct premium cost burden on families. Furthermore, the additional risk compensation from supplemental insurance, when combined with basic coverage, could result in overutilization of healthcare services, increasing out-of-pocket costs and reducing consumption levels.

Table 7.

Consumption Level Effects of the Multi-Level Health Insurance.

	(1)	(2)	(3)	(4)	(5)	(6)
Variable	HI	Basic HI	Employee HI	Residents HI	Supplemental HI	Commercial HI
chro_ijt× insu_ijt	0.124**	0.121**	0.112***	0.078*	0.082	0.074**
	(0.058)	(0.056)	(0.035)	(0.046)	(0.067)	(0.036)
FamilyFE	YES	YES	YES	YES	YES	YES
YaerFE	YES	YES	YES	YES	YES	YES
N	16,208	16,208	16,208	16,208	16,208	16,208
R ²	0.234	0.287	0.291	0.223	0.197	0.246

Note. Health insurance (HI) includes basic health insurance, supplementary health insurance, and commercial health insurance, of which basic medical insurance includes employee health insurance and resident health insurance. The values in brackets are robust standard errors after clustering at the household level.

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

In addition, columns (3) and (4) compare the efficacy of employee health insurance versus resident health insurance within the basic health insurance framework. The study finds that both types of insurance significantly reduce the adverse effects of chronic diseases on household consumption. Nonetheless, employee health insurance is identified as more effective, a finding that is consistent with the research by C. Li et al. (2019). This indicates that the capacity of resident health insurance to support consumption expenditure in households experiencing chronic disease shocks should be improved, as it has less robust institutional benefits.

The Impact of Health Insurance Based on Consumption Structure

The examination of health insurance’s impact on consumption structure, as presented in section 8a of Table 8, reveals a nuanced role of the multi-level health insurance system in mitigating the negative impact of chronic disease shocks on household consumption structure. The findings in columns (1), (2), (3), and (6) show that basic, employee, and commercial health insurance significantly attenuate the decline in consumption structure. In contrast, resident health insurance and supplemental health insurance, as indicated in columns (4) and (5), do not significantly improve the share of development and enjoyment consumption for households facing chronic diseases.

Table 8.

Consumption Structure Effects of the Multi-Level Health Insurance.

	(1)	(2)	(3)	(4)	(5)	(6)
Variable	HI	Basic HI	Employee HI	Residents HI	Supplemental HI	Commercial HI
a. Dependent variable: Consumption upgrading
chro_ijt× insu_ijt	0.007*	0.005**	0.006**	0.002	0.006	0.008**
	(0.004)	(0.002)	(0.003)	(0.004)	(0.007)	(0.004)
N	16,208	16,208	16,208	16,208	16,208	16,208
R ²	0.188	0.148	0.274	0.253	0.310	0.261
b. Dependent variable: Consumption diversity
chro_ijt× insu_ijt	0.106	0.061	0.144**	0.052	0.161	0.057***
	(0.099)	(0.098)	(0.062)	(0.078)	(0.129)	(0.021)
N	16,208	16,208	16,208	16,208	16,208	16,208
R ²	0.239	0.258	0.267	0.233	0.201	0.230

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Section 8b of Table 8 provides regression results on the influence of multi-level health insurance on household consumption diversity. It is observed that only employee and commercial health insurance, as shown in columns (3) and (6), have a significant positive effect on enhancing consumption diversity among households affected by chronic diseases. The magnitude of impact suggests that employee health insurance is more effective in increasing consumption diversity than commercial health insurance. This insight underscores the potential benefits of actively developing commercial health insurance to foster diversified consumption among residents. It also indicates a need to improve the risk protection offered by resident health insurance and supplemental health insurance to encourage optimal diversification of household consumption.

Mechanism Test

This study scrutinizes the effects of the multi-level health insurance system on total medical costs, out-of-pocket expenses, and the accumulation of precautionary savings among households experiencing chronic illness. The objective is to clarify how such a system can potentially stabilize consumption patterns in the face of chronic disease shocks.

Medical Cost Mechanism Test

Table 9 presents the results of a regression analysis examining the impact of multi-level health insurance on total medical costs (Table 9a) and out-of-pocket expenses (Table 9b) for households dealing with chronic disease shocks. The analysis reveals a significant positive correlation between health insurance coverage and total healthcare expenses at the 1% level when experiencing chronic illness shocks, as shown in columns (1) and (2). However, the impact on out-of-pocket medical costs is not statistically significant. This suggests that health insurance, including basic health insurance, does not substantially increase the financial burden on households managing chronic diseases, even with a rise in healthcare utilization, which is consistent with the findings of Zhao (2019). Upon reviewing column (3), it is observed that the influence of employee health insurance on total medical expenses is not significant. In contrast, the impact on out-of-pocket expenses is significantly negative at the 10% level, indicating that employee health insurance can significantly reduce the financial burden of medical costs for households. Conversely, column (4) shows that resident health insurance does not reduce out-of-pocket expenses and significantly increases total medical costs for households, highlighting its weaker risk mitigation function compared to employee health insurance, as discussed by C. Li et al. (2019). Column (5) indicates that supplemental health insurance significantly increases both total household medical costs and out-of-pocket medical costs at the 5% level. This could be due to the high coverage and cost-invisibility associated with supplementary insurance, which may reduce households’ cost awareness and increase their demand for higher-quality medical services. Lastly, column (6) demonstrates that commercial health insurance does not have a significant impact on either total medical costs or out-of-pocket medical costs for households facing chronic illnesses.

Table 9.

Medical Cost Mechanism Test.

	(1)	(2)	(3)	(4)	(5)	(6)
Variable	HI	Basic HI	Employee HI	Residents HI	Supplemental HI	Commercial HI
a. Dependent variable: Total medical costs
chro_ijt× insu_ijt	0.374***	0.415**	0.200	0.349***	0.398**	−0.084
	(0.143)	(0.190)	(0.137)	(0.134)	(0.170)	(0.089)
N	13,496	8,868	8,352	8,352	13,496	13,496
R ²	0.342	0.347	0.330	0.341	0.309	0.331
b. Dependent variable: Medical out-of-pocket costs
chro_ijt× insu_ijt	0.162	0.139	−0.148*	0.111	0.204**	−0.077
	(0.120)	(0.119)	(0.084)	(0.093)	(0.096)	(0.084)
N	13,496	8,352	8,352	8,352	13,496	13,496
R ²	0.353	0.362	0.347	0.341	0.322	0.333

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

The above analysis highlights the distinctive roles of different health insurance systems in reducing the economic impact of medical care. Employee health insurance effectively mitigates financial strain by lowering direct medical costs for households. In contrast, supplemental health insurance, while offering extensive coverage, may paradoxically increase financial burdens due to its hidden costs. Additionally, the study finds that the effects of resident and commercial health insurance on out-of-pocket expenditures are minimal.

Precautionary Savings Mechanism Test

Table 10 provides empirical data on precautionary savings, defined as the share of financial assets in cash and demand deposits held by households (X. Liu & Wang, 2021). The analysis shows that multi-level health insurance frameworks, encompassing basic, resident, supplemental, and commercial health insurance, each contributes to reducing precautionary savings, albeit to different extents. Specifically, resident health insurance stands out as the most effective in lowering the propensity for such savings. Interestingly, the influence of employee health insurance on precautionary savings is statistically insignificant. This lack of impact may be attributed to the demographic profile of employee health insurance participants, who are predominantly urban, with stable employment and higher levels of education, income, and wealth, thus having a reduced need for precautionary savings compared to other groups. Collectively, these results suggest that multi-level health insurance systems can promote consumption by reducing the necessity for precautionary savings. The findings to date support the validation of Hypothesis 3.

Table 10.

Precautionary Savings Mechanism Test.

	(1)	(2)	(3)	(4)	(5)	(6)
Variable	HI	Basic HI	Employee HI	Residents HI	Supplemental HI	Commercial HI
chro_ijt×insu_ijt	−0.113*	−0.062*	−0.078	−0.111***	−0.094**	−0.072*
	(0.067)	(0.036)	(0.051)	(0.042)	(0.047)	(0.042)
N	14,496	14,496	14,496	14,496	14,496	14,496
R ²	0.326	0.364	0.367	0.338	0.322	0.343

*, **

, and *** indicate the 10%, 5%, and 1% statistical significance levels.

Discussion

This study provides insights into the intricate interactions between chronic disease shocks and their impact on the level and structure of consumption. Consistent with previous studies, the empirical results show that the extensive and intensive marginal effects of chronic disease shocks significantly reduce household consumption levels (Babiarz & Yilmazer, 2017; Cheng et al., 2019; Dong et al., 2021). A key contribution is the demonstration of the negative impact of chronic disease shocks on the evolution of consumption structures, operationalized through two indices: the consumption structure upgrading index and consumption diversity. By assessing both the quantity and quality of consumption, this paper provides a comprehensive analysis of how household consumption patterns are affected by chronic disease shocks, thereby enhancing the scholarly discourse on health-related economic impacts on households.

While existing literature primarily investigates the extensive marginal effects of chronic disease shocks on their broader implications (Anand et al., 2022; Eriksen et al., 2021; E. Liu et al., 2020; Qiao et al., 2021), this study uniquely examines both extensive and intensive marginal effects to elucidate the complex interplay between chronic diseases and household consumption behavior. The findings could guide policymakers in enhancing financial and employment assistance to households with increased chronic illness, thereby mitigating their economic risks.

Analysis of heterogeneity indicates that chronic disease shocks differentially affect consumption levels and structures based on age, capital resources, and the urban-rural gap, especially considering disparities in healthcare resource access. Households over 45, those with fewer capital resources, and those in rural regions with inadequate medical services face more severe declines in consumption levels and structure, as noted by Zhang et al. (2022) and Jiang et al. (2019). This finding calls for targeted interventions to address the multifaceted challenges of poverty and financial insecurity within at-risk groups, including the elderly and rural populations. The paper emphasizes the importance of equitable medical resource distribution and social security, with a strategic focus on rural and less developed areas to manage chronic diseases effectively.

The literature on health insurance primarily assesses the efficacy of single insurance schemes, such as Shao and Hao (2020), Wang et al. (2023), and C. Li et al. (2019). This paper offers a novel analysis by assessing the synthetic effects of a multi-level health insurance system, highlighting its differential effects in mitigating the adverse impact of chronic disease shocks on consumption. It identifies basic and commercial health insurance as effective in reducing consumption declines, while employee health insurance promotes consumption structure upgrades and diversity. However, the system’s heterogeneity can exacerbate consumption disparities, particularly disadvantaging households in lower-tier insurance schemes. Mechanistic tests reveal that these insurance systems protect household consumption by reducing healthcare out-of-pocket costs and the incentive for precautionary savings, thus serving as a buffer against the economic repercussions of chronic diseases. This suggests that efforts should be made to promote the equalization of the health insurance system and to alleviate the financial burden and risk concerns of the insurance participants.

Some opportunities for future research could extend this paper and address some of our limitations. Initially, our model featured the incidence of chronic disease and the count of affected household members as independent variables. Future work could benefit from incorporating specific disease types to capture the full spectrum of chronic illness impacts, ranging from transient to permanent effects. Additionally, while our heterogeneity analysis is limited to certain household and regional attributes, with further analysis focusing solely on health insurance, the inclusion of health literacy measures could reveal its significant moderating role on consumption patterns post-disease incidence. Other household heads and regional characteristics are also likely to influence household resilience to chronic disease shocks. Future research should expand on these dimensions to better understand the micro-welfare implications of chronic diseases.

Policy Recommendations

Our research highlights the critical need to enhance preventive and managerial strategies for chronic diseases, essential for achieving the “Healthy China” initiative. We propose a multifaceted approach:

(1) Strengthen prevention and early intervention: Recognizing the profound impact of chronic diseases on household consumption patterns, we recommend the adoption of extensive health education and lifestyle modification initiatives. These strategies should aim to preemptively control chronic disease risk factors. Emphasizing the importance of preventive healthcare in early detection and intervention, we suggest reinforcing the tiered diagnosis and treatment system and improving the quality of care in primary healthcare settings.

(2) Targeted support for vulnerable groups: Given the outsized impact of chronic disease on the elderly, those with scarce capital resources, and individuals in rural or underprivileged areas, we advocate for targeted interventions. These may include the establishment of medical expense waivers and diversified financing mechanisms to alleviate the relative poverty exacerbated by chronic illnesses. We also call for the equitable distribution of medical security services to enhance the resilience of these households.

(3) Health insurance system optimization: To address the protective yet inequitable nature of the multi-level health insurance system, we propose reforms to reduce disparities in benefits across different tiers, thereby ensuring more equitable coverage and support for households in lower-tier systems.

(4) Elevate consumption levels and quality: We contend that fostering consumption requires both an increase in spending levels and adaptation to evolving consumer preferences. This involves nurturing and satisfying the public’s demand for developmental and experiential consumption, promoting innovative consumer models and business formats, and integrating online and offline consumption channels. By bolstering medical security and public services, we can mitigate residents’ risk perceptions and enhance their consumption tendencies.

Conclusions

Employing staggered DID and DDD models, this research explores the extensive and intensive margins of chronic disease shocks on both the levels and structures of household consumption. It further assesses the protective role of multi-level health insurance, leveraging CFPS data from 2012 to 2018. The main findings from our analyses can be summarized as follows. First, chronic disease shocks significantly lower household consumption and hinder the upgrading and diversity of consumption structure. Second, the impact is more severe for older households, those with fewer capital resources, and those in rural or constrained by medical resources. Third, the multi-level health insurance system mitigates the adverse effects of chronic disease shocks. However, insurance disparities can widen consumption gaps, especially for households in lower-tier systems. Finally, the protective effect of insurance operates through lowering out-of-pocket medical costs and precautionary savings, yet the efficacy of different insurance schemes varies, highlighting the need for targeted improvements. Consequently, our findings suggest that appropriate policy interventions are needed to promote the prevention and early treatment of chronic diseases, to improve the construction of a multi-level health insurance system, and to reduce systemic gaps, especially for vulnerable households.

Footnotes

Acknowledgements

We are grateful to the editors for their initial reading of the manuscript and their suggestions for improvement. We also thank the reviewers for their helpful comments.

Author Contribution

This paper is written by the authors named on the title page. Linlin Han: Literature, Investigation, Resources, Writing original draft. Yongqing Nan: Project administration, Methodology, Writing-review and editing. Feng Yang: Methodology, software, Visualization. Xuheng Zang: Project administration, Funding acquisition, Supervision. Xiaoling Xue: Writing-review and editing.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by the Major Program of National Fund of Philosophy and Social Science of China (Grant No. 21&ZD088) and in part by the Undergraduate Teaching Reform Research Project of Qilu Normal University (Grant No. JG202212).

Ethics Approval

Not applicable.

Consent to Participate

Not applicable.

Consent to Publish

Not applicable.

ORCID iD

Linlin Han

Data Availability Statement

Publicly available datasets were analyzed in this study. The datasets for this study can be found in the China Family Panel Studies (CFPS) database (isss.pku.edu.cn/cfps/). The CFPS project was conducted by the China Social Science Survey Center of Peking University and publicly available.

References

Alefan

Al-Issa

E. T.

Alzoubi

K. H.

Hammouri

H. M.

(2019). Association of smoking with direct medical expenditures of chronic diseases in north of Jordan: A retrospective cohort study. BMJ Open, 9(10), Article e031143. https://doi.org/10.1136/bmjopen-2019-031143

Anand

Dague

Wagner

K. L.

(2022). The role of paid family leave in labor supply responses to a spouse's disability or health shock. Journal of Health Economics, 83, 102621. https://doi.org/10.1016/j.jhealeco.2022.102621

Athey

Imbens

G. W.

(2022). Design-based analysis in Difference-In-Differences settings with staggered adoption. Journal of Econometrics, 226(1), 62–79. https://doi.org/10.1016/j.jeconom.2020.10.012

Babiarz

Yilmazer

(2017). The impact of adverse health events on consumption: Understanding the mediating effect of income transfers, wealth, and health insurance. Health Economics, 26(12), 1743–1758. https://doi.org/10.1002/hec.3496

Cheng

T. C.

Vaithianathan

(2019). Monthly spending dynamics of the elderly following a health shock: Evidence from Singapore. Health Economics, 28(1), 23–43. https://doi.org/10.1002/hec.3824

Chen

Wang

(2023). Social trust and income of rural residents: Evidence from China. Applied Economics, 55(45), 5352–5368. https://doi.org/10.1080/00036846.2022.2139804

Dong

Y. W.

Yang

Huang

Y. P.

(2021). Inclusive medical security, major illness shocks and consumption: Evidence from online medical security and mobile payment data. Financial Review, 13(06), 1–11.

Eriksen

T. L. M.

Gaulke

Skipper

Svensson

(2021). The impact of childhood health shocks on parental labor supply. Journal of Health Economics, 78, Article 102486. https://doi.org/10.1016/j.jhealeco.2021.102486

Fadlon

Nielsen

T. H.

(2019). Household labor supply and the gains from social insurance. Journal of Public Economics, 171, 18–28. https://doi.org/10.1016/j.jpubeco.2018.01.010

10.

Feng

Lyu

Wang

(2022). Medical resource sharing and patients' choice: Evaluation of the medical consortium construction policy in China. Management World, 38(10), 144–157. https://doi.org/10.19744/j.cnki.11-1235/f.2022.0142

11.

Goodman-Bacon

(2021). Difference-in-differences with variation in treatment timing. Journal of Econometrics, 225(2), 254–277. https://doi.org/10.1016/j.jeconom.2021.03.014

12.

Han

Song

(2024). Influence of paternal health on household consumption—A concurrent discussion of the mitigating effects of health insurance. Journal of Xiangtan University, 48(4), 112–119. https://doi.org/10.13715/j.cnki.jxupss.2024.04.019.

13.

Jiang

Chen

Xin

Wang

Zeng

Zhong

Xiang

(2019). Does the critical illness insurance reduce patients’ financial burden and benefit the poor more: A comprehensive evaluation in rural area of China. Journal of Medical Economics, 22(5), 455–463. https://doi.org/10.1080/13696998.2019.1581620

14.

Kolukuluri

(2023). Adverse health shocks, social insurance and household consumption: Evidence from Indonesia’s Askeskin Program. International Journal of Health Economics and Management, 23(2), 213–235. https://doi.org/10.1007/s10754-022-09329-6

15.

Liao

Zhang

(2022). Endogenous health risks, poverty traps, and the roles of health insurance in poverty alleviation. Health Economics Review, 12(1), 25. https://doi.org/10.1186/s13561-022-00370-2

16.

Tang

Wang

(2019). Effects of health insurance integration on health care utilization and its equity among the mid-aged and elderly: Evidence from China. International Journal for Equity in Health, 18(1), 166. https://doi.org/10.1186/s12939-019-1068-1

17.

Zhao

Zhang

(2022). Does the old-age insurance system mitigate the adverse impact of widowhood on health of the elderly? World Economy, 44(09), 180–206. https://doi.org/10.19985/j.cnki.cassjwe.2021.09.009

18.

Liu

Zhang

Feng

(2020). Old-age poverty risk of chronic diseases: theoretical mechanisms and empirical tests. Insurance Research, 11, 63–78. https://doi.org/10.13497/j.cnki.is.2020.11.005

19.

Liu

Wang

(2021). Impact of commercial health insurance on household risky financial assets investment. Finance and Trade Research, 32(05), 49–61. https://doi.org/10.19337/j.cnki.34-1093/f.2021.05.005

20.

Qiao

Wang

Guo

(2021). Chronic diseases and labor force participation among presenile and senile Chinese. Frontiers in Public Health, 9, 675927. https://doi.org/10.3389/fpubh.2021.675927

21.

Shao

Hao

(2020). Health risk, medical insurance and consumption. Insurance Research, 12, 18–37. https://doi.org/10.13497/j.cnki.is.2020.12.002

22.

Sun

Wang

Yao

(2022). The impact of housing prices on the upgrading of consumption structure: Mechanism and evidence. Journal of Shanghai University of Finance and Economics, 24(02), 61–77. https://doi.org/10.16538/j.cnki.jsufe.2022.02.005

23.

T. T.

Van

P. H.

(2019). Can health insurance reduce household vulnerability? Evidence from Viet Nam. World Development, 124, Article 104645. https://doi.org/10.1016/j.worlddev.2019.104645

24.

Wang

Gao

(2023). Long-term consumption smoothing and ex ante moral hazard effects of social medical insurance: Evidence from the new cooperative medical scheme implementation. Management Review, 35(5), 19–28. https://doi.org/10.14120/j.cnki.cn11-5057/f.2023.05.009

25.

Wang

Zheng

Liu

Zhong

(2021). Triple enzyme-regulated molecular hydrogels for carrier-free delivery of lonidamine. Advanced Functional Materials, 31(42), Article 2104418. https://doi.org/10.1002/adfm.202104418

26.

Xiao

Ding

(2023). Can farmers punch their tickets to wealth? The spillover effect of high-speed railway on agriculture development. Sage Open, 13(4), 21582440231214645. https://doi.org/10.1177/21582440231214645

27.

Xie

(2014). An Introduction to the China Family Panel Studies (CFPS). Chinese Sociological Review, 47(1), 29–33. https://doi.org/10.2753/CSA2162-0555470101

28.

Yang

Jiang

Yin

(2021). Social networks, shocks, and household consumption in China. Economic Analysis and Policy, 71, 111–122. https://doi.org/10.1016/j.eap.2021.04.008

29.

Yaya Bocoum

Grimm

Hartwig

(2018). The health care burden in rural Burkina Faso: Consequences and implications for insurance design. SSM - Population Health, 6, 309–316. https://doi.org/10.1016/j.ssmph.2018.10.012

30.

Yilma

Mebratie

A. D.

Sparrow

(2021). Economic consequences of ill-health in Rural Ethiopia. Health Systems & Reform, 7(2), Article e1885577. https://doi.org/10.1080/23288604.2021.1885577

31.

Zhang

Shi

Cao

(2022). Financial risk assessment of middle-aged and elderly families under dynamic health shocks. Journal of Finance and Economics, 48(02), 153–168. https://doi.org/10.16538/j.cnki.jfe.20210813.201

32.

Zhao

(2019). Does health insurance promote people's consumption? New evidence from China. China Economic Review, 53, 65–86. https://doi.org/10.1016/j.chieco.2018.08.007