Abstract
Background
The excessive urban‒rural income gap has been one of the main challenges facing developing countries. The establishment and improvement of health insurance is conducive to narrowing the income gap.
Methods
On the basis of the quasinatural experiment of urban–rural health insurance integration in China, we investigate the impact and mechanism of action of health insurance integration on the urban–rural income gap using the staggered difference–in-differences (DID) model.
Results
The research findings empirically indicate that health insurance integration has narrowed the urban–rural income gap by a large margin (average treatment effect on the treated (ATT)=-0.013, p < 0.01). The heterogeneity analysis reveal that in the central and western regions and in cities with large urban‒rural income gaps and high fiscal expenditure ratios, the policy effect of health insurance integration is more pronounced. Further research into the income effects of health insurance integration reveals that health insurance integration can increase the income growth of rural residents but has no significant influence on that of urban residents. We also observe through mechanism analysis that health insurance integration can narrow the urban‒rural income gap by improving rural residents’ health.
Conclusions
This study indicates that health insurance integration is an effective means of narrowing the urban‒rural income gap. The research findings of this paper provide empirical evidence for how to optimize the income distribution pattern through the reform of health insurance.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12889-026-26195-w.
Keywords: Health insurance, Health insurance integration, Urban–rural income gap, Income distribution
Introduction
The excessive urban‒rural income gap has severely retarded the sustainable development of society and is one of the main challenges faced by developing countries. Statistics from the OECD show that the income gap among countries worldwide has obviously narrowed, but such a gap has widened in different regions of a country, particularly in developing countries. The income gap in developing countries is caused mostly by the urban‒rural income gap [1]. As the world’s largest developing country, China saw its urban-rural income gap first widen and then gradually narrow following the implementation of the reform and opening-up policy. According to the “China Statistical Yearbook”, in the early years following the introduction of the reform and opening-up policy (1978–1983), the urban-rural income gap experienced a brief decline, with the income ratio between urban and rural residents falling from 2.6 to 1.8. Since the mid-1980s, the urban-rural income gap in China has progressively widened, reaching a ratio of 3.33 by 2009. Thereafter, it has gradually narrowed, with the urban-rural income ratio falling to 2.5 in 2021. However, compared to the income ratios of below 1.5 in developed countries, China’s current urban-rural income gap remains alarmingly large. How to narrow the urban‒rural income gap and how to enhance the well-being of people’s livelihoods have been major issues facing the urban‒rural integration process. The Chinese government has underscored the imperative to advance income redistribution reforms through the social security framework, with basic health insurance serving as a critical pillar of this strategy. By enhancing equitable access to healthcare services and mitigating financial burdens on vulnerable groups, this insurance mechanism functions as both a social stabilizer and a pivotal instrument in fostering inclusive prosperity.
Theoretically, health insurance exists as an effective means of income distribution [2]. Since China’s adoption of the reform and opening-up policy, it has achieved sound development of its basic health insurance system. To date, three principal health insurance systems have been subsequently established, including the Urban Employee Basic Medical Insurance (UEBMI), the New Rural Cooperative Medical System (NRCMS), and the Urban Resident Basic Medical Insurance (URBMI). However, affected by the differences in the traditional binary economic structure and institutional design of urban and rural health insurance systems, these systems are widely different in terms of internal funding and benefit standards [3]. Inadequate guarantees have been provided for vulnerable people. Therefore, China’s health insurance system has widened the income gap instead of narrowing it. To eliminate the urban‒rural differences in health insurance systems, China began integrating the URBMI and the NRCMS into the Urban‒Rural Resident Basic Medical Insurance (URRBMI) in 2016 [4]. Prior to this, some cities had already initiated the integration of the URBMI and NRCMS. The integration of urban and rural health insurance has improved rural residents’ health insurance. However, the following questions remain: Will health insurance integration impact the urban–rural income gap? If yes, will there be a difference in the effects? What is the mechanism of action? The existing research has paid little attention to the issues stated above.
On the basis of panel data from Chinese cities from 2012 to 2021, we employ the staggered difference-in-differences (DID) model to explore the effects of health insurance integration on the urban‒rural income gap. Additionally, we collect data from the China Labor Dynamics Survey (CLDS) to conduct a microlevel mechanism test. Our research contributes to the existing research findings in the following two ways. First, the existing literature has examined the influencing factors of the urban‒rural income gap from various perspectives, including industrial structure, fiscal expenditure, and financial development [5, 6]. In this study, we identify the causal relationship between health insurance integration and the urban–rural income gap on the basis of a quasinatural experiment. This can help expand the literature on factors influencing the urban‒rural income gap from the perspective of health insurance. Second, this research provides a theoretical basis and empirical support for understanding how health insurance impacts the urban‒rural income gap. This can help offer policy implications for developing countries to optimize income distribution structures through the establishment of a comprehensive health insurance system.
The remaining sections of this paper are organized as follows. The second section is a literature review; the third section introduces the institutional background of health insurance integration and conducts theoretical analysis; the fourth section outlines the research design; the fifth section presents empirical analysis; the sixth section provides further analysis; and the seventh section concludes with research findings and policy recommendations.
Literature review
The literature related to the research theme of this paper can be categorized into three main types. The first is the income effect of health insurance. The second is the income distribution effect of health insurance. The third is the policy effect of health insurance integration.
First, regarding the income effect of health insurance, the existing research shows that health insurance can increase insurance participants’ income. If any member of a low-income family not covered by health insurance has a major illness that can incur high medical expenses, the family will be forced to give up unnecessary medical treatment. This can lead to further deterioration of their living and health conditions and deepening poverty [7, 8]. Under these conditions, health insurance, which can effectively solve problems such as illness-induced poverty and the reoccurrence of poverty because of disease, is particularly important [9]. Health insurance influences income mainly through the following channels. First, reimbursement from health insurance can compensate for income loss. In this sense, health insurance exists as a form of cash transfer. Second, health insurance can improve access to medical services for participants [10] and improve their health status [11]. An increase in personal health capital can promote an increase in income [12]. In other words, health insurance can increase health capital, which can increase income levels. Together, these factors enable health insurance to effectively increase household income levels and prevent families from falling into poverty.
Second, in response to the income distribution effect of health insurance, foreign scholars believe that the establishment and improvement of health insurance can help narrow the wealth gap. Compared with other redistribution methods, health insurance has more obvious income redistribution effects [2]. However, the opinions of Chinese scholars on the income distribution effect of health insurance vary. By adopting rural residents in Sichuan Province, Zheng et al. [13] found that the social security system, including health insurance, has an overall positive regulatory effect on income distribution for rural residents. Conversely, some Chinese scholars believe that health insurance, instead of positively regulating income redistribution, has, on the contrary, negatively regulated this process. Huang [3], utilizing data from the URBMI trial evaluation, observed that, as individuals’ social status increases, the income-promoting effect of the URBMI gradually strengthens, resulting in an obviously higher target. Like the “higher target” effect of the URBMI, the NRCMS has widened the internal income gap among rural residents. Ma et al. [14] believe that the difference between the type and internal design of the medical insurance leads to a certain gap between different income groups.
Third, urban–rural health insurance integration has emerged as a major effort of the Chinese government to reform its health insurance system, which has impacted individuals and families in many ways. Huang and Wu [15] noted that health insurance integration can significantly increase the utilization rate of inpatient care among rural elderly residents. This positive effect is particularly pronounced in impoverished areas. Zhou and He [16] reported that health insurance integration has improved financial risk protection and self-assessed health status among rural residents. A possible explanation for this finding is a reduction in out-of-pocket payments. Liu et al. [17] have found that the medical insurance integration has the positive impact on low-income residents. Huo et al. [18] have examined the impact of urban-rural medical insurance integration on medical impoverishment. Other scholars have examined the impact of health insurance integration on the urban‒rural gap in medical expenditures [19].
The urban‒rural income gap is a component of income distribution. Previous research has focused on the influence of health insurance on the income distribution gap within groups. Little attention has been given to the correlation between health insurance and the urban‒rural income gap. Additionally, some scholars have explored the influence of health insurance integration on the utilization of medical services and residents’ health. Nevertheless, the income distribution effects of urban‒rural health integration have not yet been widely examined. Hence, we conducted this research in an attempt to compensate for the research gaps mentioned above.
Institutional background and theoretical analysis
Institutional background
To satisfy the basic health needs of residents, China initiated a basic medical security system composed of the UEBMI, NRCMS, and URBMI after 1998. Statistics show that by the end of 2011, 97% of national permanent residents were participants in these three medical insurance systems. This means that China’s medical security system has basically achieved nationwide full coverage. However, the Urban Resident Insurance and the NRCMS artificially divided the participants on the basis of their urban–rural household registration. This factually reinforced the segmentation of household registration and its associated social security systems. As a result, the urban and rural health insurance systems differ widely in terms of coverage, funding levels and methods, coordination layers and reimbursement methods between urban and rural areas. Since the fragmented medical insurance system contradicts the principles of equity and justice, urban–rural health insurance integration has been imperative.
China’s health insurance integration can be divided into two stages. Before 2016, some cities spontaneously integrated the URBMI and NRCMS into the URRBMI. In January 2016, the State Council issued the “Opinions on Integrating the Basic Urban‒Rural Residents’ Health Insurance System” (Guo Fa [2016] No. 3). Different regions in China subsequently accelerated their urban‒rural health insurance integration. Health insurance integration primarily covers six aspects. First, the integrated basic medical insurance for urban and rural residents covers all urban and rural residents except urban employees. Second, an integrated funding standard is implemented for both urban and rural areas. Third, an integrated scope of protection and payment standards is established. Fourth, an integrated drug list and medical service project list for urban and rural residents’ health insurance is created. Fifth, management measures for designated medical institutions are integrated. Sixth, an integrated medical insurance fund management system is enforced. In this research, we compiled documents related to health insurance integration in nearly 300 cities. We find that only a few cities had spontaneously integrated health insurance before 2016. However, two years after the Central Government released a document on health insurance integration, the number of cities seeking urban–rural health insurance integration continued to increase, reaching 172. By 2021, a majority of cities had achieved health insurance integration. The gradual promotion of health insurance integration allows us to analyse the policy effect of health insurance integration under the “quasinatural experiment” research framework and the staggered DID model.
Theoretical analysis
Before health insurance integration, the urban and rural medical insurance systems were widely separated in terms of coverage, funding level and method, pooling level and reimbursement method. Although the NRCMS managed to cover the vast majority of rural residents, rural residents still had a low level of medical insurance benefits. For example, the number of items covered by the URBMI was approximately twice that covered by the NRCMS. The reimbursement rate for urban residents was higher than that offered by the NRCMS. The NRCMS operated at the country level, meaning that seeking medical care outside the country was classified as out-of-area treatment, which hindered both treatment and reimbursement. Research shows that the NRCMS did not provide sufficient protection for rural residents, thus failing to significantly reduce out-of-pocket medical expenses [20] or efficiently control the incidence of poverty caused by illness [21]. Therefore, it is difficult to judge whether the NRCMS comprehensively enhanced the health of rural residents. In contrast, the URBMI, which provides greater benefits, more effectively safeguards the health status of urban residents. According to health capital theory [22], the greater the health level of people is, the greater the probability that residents will work and the longer they will work. This can lead to greater labour productivity, thus increasing income. This is the income effect of health [23]. Therefore, differences in the health status of rural and urban residents resulting from differences in health insurance levels can increase the income differences between urban and rural residents.
In the actual process of health insurance integration, there are three main principles. First, the reimbursement scope and benefit items for urban and rural residents’ health insurance must be consistent. Second, the overall benefits for urban and rural residents should not decrease. Third, personal contributions should not be increased significantly under the prerequisite of not increasing the fiscal burden. This means that, compared with the NRCMS, the basic medical insurance for urban and rural residents formed after integration provides more comprehensive health insurance and greater benefits for rural residents. This can significantly influence rural residents. Health insurance integration aims mainly at improving the benefit level of medical insurance for rural residents, while the benefit level for urban residents does not changed significantly before and after integration.
This research proposes the fundamental logic of how health insurance integration influences the urban–rural income gap. After integration, the benefit level for rural residents improves, which can lead to a better health status. This can subsequently enhance their income level. Urban residents, who already enjoyed a high level of health insurance benefits, are not affected by the integration in terms of their health status. Therefore, their income status remains unchanged. This means that health insurance integration has different impacts on the health of urban and rural residents and that health has an income effect, thereby narrowing the income gap between urban and rural residents. Figure 1 shows the theoretical model diagram.
Fig. 1.
Theoretical model diagram
On the basis of the above analysis, we propose the following research hypothesis: Health insurance integration can narrow the urban‒rural income gap.
Research design
Data sources
Before 2012, only a limited number of cities had implemented integrated health insurance, whereas by 2021, full integration had been achieved nationwide. Accordingly, the study period was defined as 2012–2021. Information on urban-rural health insurance integration was collected from the official government websites of each city, while city-level economic and financial data were obtained from the statistical yearbooks published on the official websites of local statistical bureaus. All data were obtained through publicly accessible online resources. To ensure consistency and replicability, we followed a step-by-step approach to locate the relevant information. For instance, we used the Baidu search engine to identify the official websites of municipal governments and statistical bureaus, and then accessed the policy documents on health insurance integration as well as the statistical yearbooks available in the “Statistical Yearbook” sections of these websites. A detailed example of the procedure used to search, identify, and retrieve these data is provided in the supplementary materials. Our data cleaning steps are as follows. Before and after the year 2014, various cities modified the statistical standards for urban and rural residents’ income. Residents’ income before and after the modification of the statistical standard is not comparable, so we keep only the samples under the new statistical standard. In addition, if there is a significant lack of other indicator data, we delete the corresponding sample. Finally, we obtain the unbalanced panel data of 1,979 observations from 236 cities over a span of 10 years.
Model settings
The standard DID model is a method of analysing causal effects by using the change in the actual untreated results of the control group as the counterfactual of the change in the untreated results of the treatment group. The staggered DID model is applicable to policies with pre- and postintervention time differences [24]. There are differences in the time points of health insurance integration in different cities, so this article uses the staggered DID model to study the impact of health insurance integration on the urban‒rural income gap. The model is shown in Eq. (1).
 |
1 |
where i denotes the city, t denotes the year; Y denotes the explained variable, i.e., the urban‒rural income gap; 
denotes the core explanatory variable, which can explain whether the city has implemented health insurance integration; X denotes a series of control variables that affect the urban‒rural income gap; 
denotes the dummy variable of the city, which is used to control the city fixed effect; 
denotes the dummy variable of the effect, which is used to control the time fixed effect; and 
is the error item.
Selection of the variables and descriptive statistics
Dependent variables
In this paper, we define the ratio of urban residents’ per capita disposable income to rural residents’ per capita disposable income as a measure of the urban‒rural income gap. The smaller the urban‒rural ratio is, the narrower the urban‒rural income gap is. The urban‒rural income gap will be wider if the opposite is true, i.e., the urban‒rural ratio is larger. The urban‒rural income ratio reflects the ratio of urban residents’ income to rural residents’ income, but it ignores the demographic structure and income distribution. Comparably, the Theil index can accommodate both [25]. Therefore, the Theil index can be used as another measure of the urban–rural income gap for a robustness test. The Theil index is is shown in Eq. (2).
 |
2 |
where 
denote the urban area and the rural area, respectively; 
denotes the total income of urban residents or rural residents in city i in year t; 
denotes the total resident income in city i in year t; 
denotes the permanent urban or rural population in city i during year t; and 
denotes the permanent population in city i during year t. The value of the Theil index ranges from 0 to 1. Similar to the urban‒rural resident income ratio, a smaller Theil index indicates a narrower urban‒rural income gap. In contrast, the higher the urban‒rural resident income ratio is, the wider the income gap is.
Independent variable
The independent variable of this paper is the treated variable of health insurance integration. If the city has already implemented health insurance integration in the current year, the value of the core explanatory variable is assigned a value of 1; otherwise, it is assigned a value of 0.
Control variables
There are numerous factors affecting the urban‒rural income gap. Referring to previous representative research, we select the following control variables. First, economic development is represented as the logarithm of per capita regional GDP. The second is the industrial structure, which is the ratio of the value added of the secondary and tertiary industries to the regional GDP. The third is human capital, which is measured by the ratio of the number of students in higher education institutions to the permanent population. The fourth is the fiscal expenditure ratio, which is represented by the ratio of general public budget expenditures to regional GDP. The fifth is opening-up, which is indicated by the ratio of total exports to regional GDP. The sixth is financial development, which is measured by the ratio of the year-end balance of loans from financial institutions to the regional GDP. The definitions of all these variables and their descriptive statistics are presented in Table 1. As shown in Table 1, the average income ratio of the sample cities is 2.298, which is far higher than the reasonable level of 1.5. This means there is a wide urban‒rural income gap in China.
Table 1.
Descriptive statistics
| Variables | Definitions | Mean | SD |
|---|---|---|---|
| Urban‒rural income gap | Urban residents’ per capita disposal income/Rural residents’ per capita disposal income | 2.298 | 0.405 |
| Theil index | Refer to Eq. (2) above. | 0.075 | 0.036 |
| Integration | Whether the health insurance integration is adopted or not: Yes = 1; No = 0. | 0.528 | 0.499 |
| Economic development | Logarithm of per capita regional production | 10.779 | 0.510 |
| Industrial structure | Value added of the secondary and tertiary industries/GDP | 0.872 | 0.080 |
| Human capital | Number of students in general institutes of higher education/Permanent resident population | 0.018 | 0.020 |
| Fiscal expenditure ratio | General public budget expenditure/GDP | 0.214 | 0.099 |
| Opening‒up | Export value/Gross regional production | 0.074 | 0.111 |
| Financial development | Year-end financial institution loan balance/Gross regional production | 1.097 | 0.606 |
Analysis of the regression results
Benchmark regression
Table 2 shows the regression results for the impact of health insurance integration on the urban‒rural income gap. Column (1) of Table 2 does not include control variables and does not control for two-way fixed effects. On the basis of Column (1), Column (2) includes the control variables. Column (3), which does not include the control variables, controls for two-way fixed effects. On the basis of Column (3), Column (4) includes the control variables. According to the regression results in Columns (1) to (4), the coefficient of the independent variable, namely, health insurance integration, is negative, which is insignificant at the 1% level. This suggests that health insurance integration can reduce the urban‒rural resident income ratio and narrow the urban‒rural income gap. This provides solid evidence for our research hypothesis.
Table 2.
Benchmark regression
| Variables | (1) | (2) | (3) | (4) |
|---|---|---|---|---|
| Urban‒rural income gap | ||||
| Integration | -0.121*** | -0.073*** | -0.013*** | -0.013*** |
| (0.018) | (0.016) | (0.003) | (0.003) | |
| Economic development | -0.279*** | -0.016 | ||
| (0.026) | (0.028) | |||
| Industrial structure | 0.029*** | 0.003* | ||
| (0.001) | (0.002) | |||
| Human capital | 3.261*** | 0.683 | ||
| (0.541) | (0.858) | |||
| Fiscal expenditure ratio | 2.044*** | -0.057 | ||
| (0.120) | (0.072) | |||
| Opening‒up | -0.367*** | -0.106* | ||
| (0.071) | (0.059) | |||
| Financial development | -0.076*** | -0.000 | ||
| (0.017) | (0.016) | |||
| Constant | 2.361*** | 2.389*** | 2.305*** | 2.219*** |
| (0.013) | (0.269) | (0.002) | (0.284) | |
| City fixed effect | NO | NO | YES | YES |
| Time fixed effect | NO | NO | YES | YES |
| Observations | 2,011 | 1,979 | 2,011 | 1,979 |
| R-squared | 0.022 | 0.328 | 0.985 | 0.985 |
The values in parentheses represent clustered robust standard errors, with, *,** and *** indicating significance at the 10%, 5%, and 1% levels, respectively
Validity test of the staggered DID model
Parallel trend test
A prerequisite for using the DID method for causal identification is the parallel trend assumption of the experimental group and the control group before the policy shock. In other words, before urban–rural health insurance integration, the urban–rural income gap between cities in the experimental group and cities in the control group was not significantly different. Thus, we adopt the parallel trend test according to the research of Huang and Liu [26]. The length of the event window involved in intervening in staggered data should be greater than the length of the sample time. The distribution of the observed values before and after the intervention is unbalanced, and the wider the event window is, the more severe the sample imbalance phenomenon is, which may lead to sample selection errors and sample consumption problems. Therefore, the width of the event window should not be too long. The observation samples used in this study were collected during the four periods before and after the intervention. Therefore, we take the current year and four years before and after health insurance integration as the inspection period. The fourth year before the policy implementation is defined as the baseline. A dummy variable is set such that if a city implements health insurance integration in the current year, the is assigned a value of 1; otherwise, is assigned a value of 0. Dummy variables are established for each corresponding year. Next, the treated variables in the baseline model are replaced with dummy variables, while the control variables remain the same. Then, the regression is conducted. The estimated results are shown in Fig. 2. Figure 2 shows that the 95% confidence intervals for the estimation coefficient in the three years preceding health insurance integration cover 0. This indicates that the coefficients are not statistically significant, which satisfies the preexisting parallel trend condition. Additionally, Fig. 2 shows that in the year of policy implementation and the following two years, the coefficient is statistically significant but has relatively small absolute values. In the third and fourth years after the policy implementation, the absolute values of the coefficients are larger and significant. This means that the policy effects of health insurance integration become more pronounced over time.
Fig. 2.
Parallel trend test
Placebo test
To avoid the influence of random factors on the baseline results, this study adopts the approach of Liu and Lu [27] by randomly assigning whether each city implements health insurance integration. The process is repeated with 500 resamples for the regression analysis, and the resulting kernel density function distribution of the estimation coefficient is shown in Fig. 3. Theoretically, with the random assignment of the experimental and control groups, the treatment variable should not have a significant effect on the outcome variable. As indicated in Fig. 3, the estimation coefficient from the random sampling is approximately symmetrically distributed around 0, whereas the true coefficient value from the baseline regression is -0.013, which is located strictly on the left side of the probability distribution of the estimated coefficients from the random sampling. This suggests that the placebo test results are significant. Consequently, these findings indicate that the baseline regression results are not driven by unobserved factors.
Fig. 3.
Placebo test
Robustness test
Replacing the dependent variable
We use the Theil index to replace the urban‒rural income ratio as a measure of the urban‒rural income gap. As shown in Column (1) of Table 3, the coefficient of health insurance integration is negative and significant at the 5% level. This indicates that the conclusions from the baseline regression remain unchanged after the dependent variable is replaced.
Table 3.
Robustness test
| Variables | (1) | (2) | (3) | (4) |
|---|---|---|---|---|
| Theil index | Urban‒rural income gap | |||
| Integration | -0.078** | -0.013*** | -0.009** | -0.012*** |
| (0.032) | (0.003) | (0.004) | (0.003) | |
| Economic development | -1.241*** | -0.015 | 0.021 | -0.012 |
| (0.341) | (0.027) | (0.026) | (0.027) | |
| Industrial structure | 0.068*** | 0.003** | 0.003* | 0.003 |
| (0.022) | (0.002) | (0.002) | (0.002) | |
| Human capital | -5.418 | 0.771 | 0.969 | 0.777 |
| (9.202) | (0.869) | (0.915) | (0.876) | |
| Fiscal expenditure ratio | -1.353 | -0.052 | 0.037 | -0.054 |
| (1.130) | (0.071) | (0.070) | (0.071) | |
| Opening‒up | -1.487 | -0.074 | -0.072 | -0.116* |
| (0.945) | (0.061) | (0.062) | (0.059) | |
| Financial development | -0.355 | 0.003 | 0.006 | -0.002 |
| (0.248) | (0.016) | (0.016) | (0.016) | |
| Port distance × Time trend | -0.000 | |||
| (0.000) | ||||
| Constant | 15.903*** | 2.181*** | 1.785*** | 2.225*** |
| (3.412) | (0.283) | (0.284) | (0.283) | |
| City fixed effect | Yes | Yes | Yes | Yes |
| Time fixed effect | Yes | Yes | Yes | Yes |
| Observations | 1,712 | 1,957 | 1,668 | 1,979 |
| R-squared | 0.980 | 0.985 | 0.986 | 0.985 |
The values in parentheses represent clustered robust standard errors, with, *,** and *** indicating significance at the 10%, 5%, and 1% levels, respectively
Excluding municipalities directly under the control of the central government
Compared with prefectural cities, municipalities directly under the control of the central government have greater health insurance benefits and a smaller urban‒rural income gap. The differences between municipalities and prefectural cities may affect the regression results. Therefore, this study excludes samples of municipalities directly under the control of the central government and re-runs the regression. As shown in Column (2) of Table 3, the coefficient of health insurance integration is significantly negative at the 1% level, suggesting that the regression results remain robust after these municipalities are excluded.
PSM-DID
To overcome systematic differences between the experimental and control groups and enhance their comparability, this study selects the control variables from the baseline regression as covariates. Using a 1-to-2 nearest neighbour matching method, propensity score matching is applied to the sample data, followed by DID estimation on the matched data. As noted in Column (3) of Table 3, the coefficient of health insurance integration is significantly negative at the 5% level, which is consistent with the results of the baseline regression.
Considering the nonrandomness of pilot programs
The selection of cities for health insurance integration is not entirely random and may be related to certain inherent attributes of the cities, such as whether they are provincial capital cities, their altitude, or their distance from the nearest coastal port. Although it has been verified that the urban‒rural income gap between experimental and control cities had parallel trends before health insurance integration, the inherent attributes of cities may have differential impacts on the urban‒rural income gap over time. Following the approach of Li et al. [28], we select the distance of cities from the nearest coastal port as a nontime-varying attribute variable and include an interaction term between the port distance and time trends in the baseline model to mitigate potential estimation bias arising from the nonrandomness of health insurance integration implementation. The regression results, as shown in Column (4) of Table 3, indicate that the coefficient of health insurance integration is -0.012 and significant at the 1% level. This suggests that the conclusions from the baseline regression remain robust even after the nonrandomness of health insurance integration implementation is considered.
Heterogeneous treatment effects
Heterogeneous treatment effects are a potential issue when the staggered DID model is used. Specifically, in this study, heterogeneous treatment effects refer to the possibility that the impact of health insurance integration on the urban‒rural income gap varies across different cities. If conventional two-way fixed effect (TWFE) estimators are used for estimation, the estimated coefficients may be biased [29]. This study employs the method proposed by De Chaisemartin and D’Haultfoeuille to test for potential heterogeneous treatment effects in the baseline model results. The measured standard deviations of the regression coefficients under heterogeneous treatment effects are 0.006 and 0.014, which are significantly smaller than the value of 1 required for robust results. This finding indicates that the baseline model results may not be robust in the presence of heterogeneous treatment effects. Therefore, we refer to the “heterogeneous-robust” estimator proposed by Callaway and Sant’Anna [22] to correct for potential bias caused by heterogeneous treatment effects. As shown in Table 4, the coefficient from the “heterogeneous-robust” estimator is -0.011 and significant at the 5% level. This finding indicates that the results remain consistent with the baseline regression results even after heterogeneous treatment effects are considered.
Table 4.
Heterogeneous treatment effects
| Variables | Coefficient | Standard error | P‒value | Observations |
|---|---|---|---|---|
| ATT | -0.011 | 0.004 | 0.01 | 1502 |
All control variables, city fixed effects, and time fixed effects have been accounted for in the model, and the coefficient of control variables are not reported in the Stata results
Further analysis
Heterogeneous analysis
Heterogeneous analysis on the basis of geographical location
As mentioned above, health insurance benefits in eastern China are better than those in central and western China. We thus infer that health insurance integration can improve urban health insurance benefits in western China more obviously to further influence the urban‒rural income gap. In light of geographical differences, we study the heterogeneity of the effect of health insurance integration on the urban‒rural income gap across different geographical locations. First, the samples are divided into city samples in eastern China and central and western China. Then, we run regressions on these two groups of samples. The results are shown in Table 5. Column (1) of Table 5 suggests that the coefficient of health insurance integration is negative but not significant. These findings indicate that health insurance integration does not affect the urban‒rural income gap in eastern China. Column (2) of Table 5 shows that the coefficient of health insurance integration is significantly negative at the 5% level. This finding confirms that health insurance integration can significantly reduce the urban–rural income gap in areas of central and western China. One possible explanation is that the eastern areas are more developed in terms of the economy, so the health insurance benefits for rural residents before integration were already high. As a result, health insurance integration does not have a notable effect on the urban‒rural income gap in these regions.
Table 5.
Heterogeneous analysis
| Variables | (1) | (2) | (3) | (4) | (5) | (6) |
|---|---|---|---|---|---|---|
| East | Central and West | Large income gap | Small income gap | High fiscal expenditure ratio | Low fiscal expenditure ratio | |
| Integration | -0.002 | -0.010** | -0.012** | -0.005 | -0.019*** | -0.008* |
| (0.006) | (0.004) | (0.006) | (0.005) | (0.005) | (0.004) | |
| Economic development | 0.017 | -0.010 | -0.050 | 0.056** | -0.105*** | 0.145*** |
| (0.050) | (0.030) | (0.037) | (0.022) | (0.037) | (0.031) | |
| Industrial structure | -0.005 | -0.002 | -0.002 | -0.002 | 0.003* | -0.001 |
| (0.007) | (0.002) | (0.002) | (0.001) | (0.002) | (0.003) | |
| Human capital | 2.203** | 0.358 | 1.501* | -0.671 | -0.321 | 0.881 |
| (0.918) | (1.174) | (0.906) | (0.921) | (1.234) | (0.978) | |
| Fiscal expenditure ratio | -0.371** | -0.033 | -0.229** | -0.097 | -0.105 | -0.207 |
| (0.175) | (0.073) | (0.094) | (0.068) | (0.074) | (0.183) | |
| Opening‒up | 0.047 | -0.092 | -0.309** | 0.159** | -0.232*** | 0.317*** |
| (0.144) | (0.061) | (0.124) | (0.0r64) | (0.085) | (0.113) | |
| Financial development | -0.022 | 0.007 | 0.011 | 0.014 | -0.001 | 0.039* |
| (0.034) | (0.017) | (0.026) | (0.015) | (0.025) | (0.021) | |
| Constant | 2.051*** | 2.489*** | 3.212*** | 1.409*** | 3.636*** | 0.516 |
| (0.632) | (0.352) | (0.432) | (0.257) | (0.424) | (0.384) | |
| City fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Time fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Observations | 476 | 1,503 | 983 | 983 | 972 | 983 |
| R-squared | 0.989 | 0.985 | 0.977 | 0.972 | 0.991 | 0.985 |
The values in parentheses represent clustered robust standard errors, with, *, **and *** indicating significance at the 10%, 5%, and 1% levels, respectively
Heterogeneous analysis based on the urban–rural income gap
In general, a smaller urban‒rural income gap is associated with a greater level of economic development and better health insurance benefits. It can be further inferred that health insurance integration primarily enhances the health insurance benefits in cities with a significant urban‒rural income gap. This can help narrow the income disparity in those areas. In this research, we investigate whether the policy effects of health insurance integration are heterogeneous on the basis of differences in the urban‒rural income gap. The city samples are divided into high and low urban‒rural income gap groups by median income on an annual basis, and regressions are conducted separately for each group, with the results shown in Table 5. Column (3) of Table 5 reveals that the coefficient of health insurance integration is positive and significant at the 5% level. This finding indicates that if a city has a considerable urban‒rural income gap, health insurance integration will have a notable policy effect. Conversely, as shown in Column (4) of Table 5, the coefficient of health insurance integration is negative but not significant, suggesting that if a city has a small urban–rural income gap, health insurance integration does not have a clear policy effect. These results indicate that the impact of health insurance integration on the urban‒rural income gap is more akin to “providing timely assistance” rather than “adding embellishments”.
Heterogeneous analysis based on the fiscal expenditure ratio
In this research, we investigate whether the policy effects of health insurance integration are heterogeneous on the basis of differences in the fiscal expenditure ratio. The city samples are divided into high and low fiscal expenditure ratio groups according to the median fiscal expenditure ratio on an annual basis, and regressions are conducted separately for each group, with the results shown in Table 5. According to Columns (5) and (6) of Table 5, the coefficient of health insurance integration in the group with a high fiscal expenditure ratio is -0.019, which is significant at the 1% level. The coefficient of health insurance integration for the low fiscal expenditure ratio group is -0.008 and significant at the 10% level. Because the absolute value and significance of the coefficient in the high fiscal expenditure ratio group are greater than those in the low fiscal expenditure ratio group, the effect of health insurance integration on narrowing the urban‒rural income gap is significant in the high fiscal expenditure ratio group.
Income effect of health insurance integration
This study further examines the income effects of health insurance integration. The per capita disposable income of urban residents and the per capita disposable income of rural residents are logged and used as dependent variables in regression Model (1), and the results are presented in Table 6. Column (1) of Table 6 shows that the coefficient of health insurance integration is positive but not significant, indicating that health insurance integration does not have a noticeable effect on urban residents’ income. Column (2) of Table 6 presents the regression results for rural residents’ income, where the coefficient of health insurance integration is significantly positive at the 1% level. These findings indicate that health insurance integration can significantly increase rural residents’ income. This finding aligns with the theoretical analysis mentioned earlier. Therefore, health insurance integration primarily improves health insurance benefits for rural residents, which in turn has a substantial effect on their income. In contrast, the level of health insurance benefits for urban residents has not changed significantly before and after health insurance integration. This has no effect on their income. In summary, health insurance integration primarily reduces the urban‒rural income gap by increasing the income of rural residents.
Table 6.
Income effect of health insurance integration
| Variables | (1) | (2) |
|---|---|---|
| Urban residents’ income | Rural residents’ income | |
| Integration | 0.003 | 0.007*** |
| (0.002) | (0.002) | |
| Economic development | 0.102*** | 0.098*** |
| (0.010) | (0.010) | |
| Industrial structure | -0.001 | -0.002** |
| (0.001) | (0.001) | |
| Human capital | -0.219 | -0.342 |
| (0.369) | (0.388) | |
| Fiscal expenditure ratio | -0.059* | -0.001 |
| (0.033) | (0.032) | |
| Opening‒up | 0.074*** | 0.086*** |
| (0.027) | (0.021) | |
| Financial development | 0.025*** | 0.023*** |
| (0.008) | (0.009) | |
| Constant | 9.268*** | 8.584*** |
| (0.103) | (0.120) | |
| City fixed effect | Yes | Yes |
| Time fixed effect | Yes | Yes |
| Observations | 1,979 | 1,979 |
| R-squared | 0.995 | 0.996 |
The values in parentheses represent clustered robust standard errors, with, *, **and *** indicating significance at the 10%, 5%, and 1% levels, respectively
Mechanism analysis
In examining the causal pathways, a common approach in the relevant literature is to propose one or several mediating variables. These variables theoretically have a clear causal relationship with the dependent variable and are closely aligned in terms of logic and temporal or spatial context, focusing solely on the impact of the policy variable [26]. Given that health has a significant income effect, this section aims to demonstrate the relevant mechanisms by investigating the health effects of health insurance integration.
Owing to a lack of macrolevel data measuring the health status of urban residents compared with that of rural residents at the prefectural level, this section utilizes data from the China Labor Dynamics Survey (CLDS) for the years 2014, 2016, and 2018. This study explores the microlevel mechanisms through which health insurance integration affects the urban–rural income gap from the perspective of household health. The CLDS team conducted data surveys in three stages: 2014, 2016, and 2018. The three periods of data contain a total of 55,735 samples, including 23,293 samples from 2014, 20,452 samples from 2016, and 11,990 samples from 2018. The dependent variable is residents’ health status, with self-reported health (SRH) used as the indicator of health condition. In the CLDS, respondents are asked, “How do you perceive your current health status?” The five options are “1. Very healthy, 2. Healthy, 3. Average, 4. Fairly unhealthy, 5. Very unhealthy.” The article assumes that if the interviewee chooses 1 or 2, their health status is set to 0; if the interviewee chooses 3, 4, or 5, their health status is set to 1. Thus, a lower numerical value indicates better health, whereas a higher value reflects worse health. The independent variable is a binary variable for health insurance integration, where a value of 1 is assigned if the respondent’s city has implemented health insurance integration; otherwise, it is assigned a value of 0. At both the individual and family levels, this study includes eight control variables: gender, marital status, age, education level, smoking status, drinking status, exercise habits, and degree of family tidiness. The specific definitions of these variables and their descriptive statistics are provided in Table 7. As the dependent variable, health status is a dummy variable ranging from 1 to 0; thus, this study employs a panel probit model to conduct the regression analysis.
Table 7.
Descriptive statistics of major variables in CLDS
| Variables | Definitions | Mean | SD |
|---|---|---|---|
| Health status | Unhealthy = 1; Healthy = 0 | 0.340 | 0.490 |
| Gender | Male = 1, Female = 0 | 0.477 | 0.500 |
| Marital status | Married = 1, Unmarried = 0 | 0.822 | 0.382 |
| Age | Age upon interview (Year) | 44.819 | 14.547 |
| Education level | High school and above = 1; Junior high school and below = 0 | 0.309 | 0.462 |
| Smoking status | Smoking = 1; Non-smoking = 0 | 0.267 | 0.442 |
| Drinking status | Drinking = 1; Non-drinking = 0 | 0.188 | 0.391 |
| Exercise habits | Exercise = 1; Non-exercise = 0 | 0.276 | 0.447 |
| Degree of family tidiness | Very Untidy = 1; Very Tidy = 1 | 6.358 | 1.730 |
Table 8 shows the regression results for the mechanism analysis. Column (1) of Table 8 shows that the coefficient of health insurance integration is negatively significant at the 1% level. This finding suggests that health insurance integration can significantly increase residents’ health level. Column (2) of Table 8 shows that the coefficient of health insurance integration is -0.003, but it is not significant. These findings suggest that urban‒rural health insurance integration has not significantly influenced urban residents’ health status. According to Column (3) of Table 8, the coefficient of health insurance integration is negatively significant at the 1% level. This means that urban‒rural health insurance integration has significantly improved rural residents’ health levels. The health effects of health insurance integration on urban and rural residents are different because health insurance integration mainly improves rural residents’ health insurance level. This undoubtedly leads to a positive effect of health insurance integration on rural residents’ health status. Considering the significant income effects of health, we conclude that health insurance integration improves the health conditions of rural residents, subsequently increases their income levels, and ultimately narrows the urban‒rural income gap.
Table 8.
Mechanism analysis
| Variables | (1) | (2) | (3) |
|---|---|---|---|
| Residents’ health | Urban residents’ health | Rural residents’ health | |
| Integration | -0.051*** | -0.003 | -0.073*** |
| (0.017) | (0.027) | (0.023) | |
| Gender | -0.232*** | -0.183*** | -0.265*** |
| (0.021) | (0.033) | (0.027) | |
| Marital status | -0.097*** | -0.081** | -0.116*** |
| (0.025) | (0.038) | (0.033) | |
| Age | 0.347*** | 0.343*** | 0.349*** |
| (0.008) | (0.013) | (0.009) | |
| Education level | -0.271*** | -0.203*** | -0.276*** |
| (0.020) | (0.028) | (0.031) | |
| Smoking status | 0.098*** | 0.144*** | 0.072** |
| (0.023) | (0.038) | (0.029) | |
| Drinking status | -0.064*** | -0.037 | -0.079*** |
| (0.022) | (0.037) | (0.028) | |
| Exercise habits | -0.068*** | -0.045* | -0.068*** |
| (0.018) | (0.026) | (0.025) | |
| Degree of family tidiness | -0.091*** | -0.093*** | -0.089*** |
| (0.005) | (0.008) | (0.006) | |
| Time effect | Yes | Yes | Yes |
| Observations | 5,5735 | 2,0435 | 3,5300 |
The values in parentheses represent robust standard errors, with *, **, and *** indicating significance at the 10%, 5%, and 1% levels, respectively. As panel probit model is used, individual fixed effects were not controlled
Conclusions and implications
Like a majority of developing countries, China faces a pronounced urban‒rural income gap. This gap has severely retarded sustainable social development. Hence, the Chinese government has initiated various policies to narrow the urban‒rural income gap. Theoretically, as an effective means to redistribute income, health insurance plays a significant role in realizing urban‒rural integration. Nevertheless, the fragmented and undeveloped health insurance system in China has widened the urban‒rural income gap. In recent years, due to reform efforts by the Chinese government, China’s health insurance integration has made great strides. However, research into the income distribution effects of these policy reform efforts is still insufficient. We believe that health insurance integration can enhance the health insurance benefits for rural residents, thereby narrowing the urban‒rural gap. In this study, we conduct a quasinatural experiment on health insurance integration and use city panel data and household microsurvey data to empirically analyse the effects of health insurance integration on the urban–rural income gap and its mechanisms.
Our research reveals an obvious narrowing of the urban–rural income gap after health insurance integration. This finding was further demonstrated by a series of robustness tests. Therefore, further improvements in the basic urban–rural health insurance system and residents’ health insurance benefits are essential. For example, a dynamic financing mechanism tailored to the economic development level should be established, transitioning the current fixed-amount financing mechanism to a proportional one. Where financially feasible, public finance should increase subsidies for health insurance contributions to alleviate the payment burden on rural residents. Health insurance coverage should be expanded to include more chronic and major diseases, ensuring that rural residents can receive timely and effective treatment when confronted with significant health issues. For rural residents, the health insurance reimbursement ratio should be appropriately increased to reduce their medical burden and increase their willingness and ability to seek medical care. Increased investment should be made in medical and health infrastructure in rural areas to guarantee that rural residents can access high-quality medical services nearby.
In this study, we also observed that the effect of health insurance integration on the urban‒rural income gap is heterogeneous. Health insurance integration can significantly narrow the urban‒rural income gap in central and western China. However, this effect is insignificant in eastern China. Furthermore, the health insurance integration effects are stronger in cities with a larger urban–rural income gap or high fiscal expenditure ratio. Conversely, the effect weakens in regions with a smaller urban‒rural income gap or low fiscal expenditure ratio. Therefore, we believe that for central China and western China, as well as for economically underdeveloped regions, the central government should increase fiscal transfer payments to support their medical insurance integration and the construction of medical and health service systems. Through policy guidance and market mechanisms, medical resources should be tilted towards the central and western regions, enhancing the service capabilities of grassroots medical institutions and reducing the gap in medical resources between regions.
In addition, there was a notable health effect of health insurance integration. However, health insurance integration significantly improved the health of only rural residents and not urban residents. In terms of the health income effect, health insurance integration positively promoted the income level of rural residents, but its effect on urban residents’ income was not evident. Therefore, health insurance integration influences the urban‒rural income gap by improving rural residents’ health. We suggest improving the health awareness level of rural residents through various means, including enhancing the level of medical insurance coverage and conducting health education activities, which will reduce the phenomenon of poverty caused by illness and the return to poverty due to illness. In summary, health insurance integration can effectively reduce the urban‒rural income gap. This finding has positive policy implications for other low-income and medium-income countries. In low-income and medium-income countries, health insurance programs are usually fragmented. Speeding up health insurance integration in these countries can bridge the urban‒rural gap, ensuring inherent equity for all.
It is worth noting that this study has certain limitations. We believe that health insurance integration may also indirectly impact income distribution through other channels, such as reducing out-of-pocket medical costs. Unfortunately, there are no relevant data on residents’ medical expenses in the “China City Statistical Yearbook” or the CLDS data. This gap in data availability constitutes a significant limitation, as it prevents comprehensive assessment of the cost-sharing mechanisms’ distributional consequences across urban-rural divides and income strata.
Supplementary Information
Abbreviations
- DID
Difference-in-differences
- UEBMI
Urban employee basic medical insurance
- NRCMS
New rural cooperative medical system
- URBMI
Urban resident basic medical insurance
- URRBMI
Urban-rural resident basic medical insurance
- CLDS
China labor dynamics survey
Authors’ contributions
ZY designed the study, performed the statistical analysis and drafted the manuscript. ZL and RX organized this study and contributed to the study design, interpretation of the analysis, and revision of the manuscript. The authors read and approved the final manuscript.
Funding
This study benefited from research grants of Social Science Foundation of Anhui Provincial Education Department (Grant No. 2024AH052734; Grant No. 2023AH030071), Research Initiation Funding of Anhui University of Chinese Medicine (Grant No. 2024rcyb06; Grant No.2023rcyb015), Quality Engineering Project of the Anhui Provincial Department of Education (2023xjzlts028; 2024zyxwjxalk115), Key Priority Research Project of the Evergreen Program of Anhui University of Chinese Medicine (Grant No. CQT20250207), Anhui Province Social Science Innovation and Development Research Project (Grant No. 2022CX057), Science and Technology Innovation Strategy and Soft Science Fund of Anhui Province (Grant No. 202506f02050013) and National Social Science Fund of China (Grant No. 24BJY016). The funding agencies had no role in the design, analysis, interpretation, or writing of this study.
Data availability
The data is available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
Ethical approval was obtained from Anhui University of Chinese Medicine (AHUCM-HSS-2024010), and the study was conducted in strict adherence to the principles outlined in the Declaration of Helsinki, including its subsequent amendments, or equivalent ethical standards.
Consent for publication
All authors declare consent for publication.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Zhiguang Li, Email: lizhiguang0731@163.com.
Ruijin Xie, Email: xierj@ahtcm.edu.cn.
References
- 1.Lagkos D. Urban-rural gaps in the developing world: does internal migration offer opportunities? J Economic Perspect. 2020;34(3):174–92. [Google Scholar]
- 2.Conde-Ruiz J, Profeta P. The redistributive design of social security systems. Econ J. 2007;117:686–712. [Google Scholar]
- 3.Huang W. Impact of china’s urban resident basic medical insurance on targeted poverty alleviation. Economic Res J (in Chinese). 2017;52(09):117–32. [Google Scholar]
- 4.Liu T, Gao Y, Li H, Zhang L, Sun J. Analysis of the operational efficiency of basic medical insurance for urban and rural residents: based on a three-stage DEA model. Int J Environ Res Public Health. 2022;19(21):13831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Demirgüç-Kunt A, Levine R. Finance and inequality: theory and evidence. Annual Rev Financial Econ. 2009;1:287–318. [Google Scholar]
- 6.Su X, et al. Reforms on county-level fiscal governance in china: impact on urban-rural income inequality. Int Rev Public Adm. 2019;24(2):81–100. [Google Scholar]
- 7.Das J, Hammer J, Leonard K. The quality of medical advice in low Icome countries. J Economic Perspect. 2008;22:93–114. [DOI] [PubMed] [Google Scholar]
- 8.Zhou Y, Guo Y, Liu Y. Health, income and poverty:evidence from china’s rural household survey. Int J Equity Health. 2020;19:36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chen, Yuyu. Jin ginger Zhe. Does health insurance coverage lead to better health and educational outcomes? Evidence from rural China. J Health Econ. 2012;31:1–14. [DOI] [PubMed] [Google Scholar]
- 10.Miller G, Pinto DM, Vera-Hernandez M. Risk protection, service use, and health outcomes under colombis’s health insurance program for the poor. Am Econ Journal: Appl Econ. 2013;5(4):61–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Woolhandler S, Himmelstein DU. The relationship of health insurance and mortality: is lack of insurance deadly? Ann Intern Med. 2017;167(6):424–31. [DOI] [PubMed] [Google Scholar]
- 12.Schultz TP, Tansel A. Wage and labor supply effects of illness in Cote d i’voire and ghana: instrumental variable estimates for days disabled. J Dev Econ. 1997;53(2):251–86. [Google Scholar]
- 13.Zheng S, Du X, Chen C. Research on the regulatory effect of social security on income distribution of rural residents: a case study of Sichuan Province. Rural Economic (in Chinese). 2015;393(07):80–4. [Google Scholar]
- 14.Ma M, Li Y, Wang N, et al. Does the medical insurance system really achieved the effect of poverty alleviation for the middle-aged and elderly people in china? Characteristics of vulnerable groups and failure links. BMC Public Health. 2020;20:435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Huang X, Wu B. Impact of urban-rural health insurance integration on health care: evidence from rural Chian. China Econ Rev. 2020;64:101543. [Google Scholar]
- 16.Zhou Q, He Q, Eggleston K, Liu G. Urban-rural health insurance integration in china: impact on health care utilization, financial risk protection, and health status. Appl Econ. 2022;54:2491–509. [Google Scholar]
- 17.Liu H, Zhu H, Wang J, et al. Catastrophic health expenditure incidence and its equity in china: a study on the initial implementation of the medical insurance integration system. BMC Public Health. 2019;19:1761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Huo J, Hu M, Li S. The impact of urban-rural medical insurance integration on medical impoverishment:evidence from China. Int J Equity Health. 2023;22:245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Yang D, Acharya Y, Liu X. Social health insurance consolidation and urban-rural inequality in utilization and financial risk protection in China. Soc Sci Med. 2022;308:115200. [DOI] [PubMed] [Google Scholar]
- 20.Lei X, Lin W. The new cooperative medical scheme in rural china: does more coverage mean more service and better health? Health Econ. 2009;18:25–46. [DOI] [PubMed] [Google Scholar]
- 21.Yip W, Hsiao WC. Non-evidence-based policy: how effective is china’s new cooperative medical scheme in reducing medical impoverishment. Soc Sci Med. 2009;68(2):201–9. [DOI] [PubMed] [Google Scholar]
- 22.Mushkin SJ. Health as an investment. J Polit Econ. 1962;70(5):129–57. [Google Scholar]
- 23.Xie M, Huang Z, Zang W. The inequality of health-income effect in employed workers in china:a longitudinal study from China family panel studies. Int J Equity Health. 2020;19:96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Callaway B, Sant’Anna P. Difference-in-Differences with multiple time periods. J Econ. 2021;225(2):200–30. [Google Scholar]
- 25.Sicular T, Yue X, Gustafsson B, Li S. The Urban-rural income gap and inequality in China. Rev Income Wealth. 2007;53(1):93–126. [Google Scholar]
- 26.Huang W, Liu H. Early childhood exposure to health insurance and adolescent outcomes: evidence from rural China. J Dev Econ. 2023;160:1–18. [Google Scholar]
- 27.Liu Q, Lu Y. Firm investment and exporting: evidence from china’s value-added tax reform. J Int Econ. 2015;97(2):392–403. [Google Scholar]
- 28.Li P, Lu Y, Wang J. Does flattening government improve economic performance? Evidence from China. J Dev Econ. 2016;123:18–37. [Google Scholar]
- 29.De Chaisemartin C, D’Haultfoeuille X. Two-way fixed effects estimators with heterogeneous treatment effects. Am Econ Rev. 2020;110(9):2964–96. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data is available from the corresponding author on reasonable request.