1 Introduction
In recent years, the size of China's residents' debts has increased rapidly, while the saving rate has decreased, increasing the financial vulnerability of households. According to the 2019 Survey on Household Assets and Liabilities of Urban Residents in China, the average household debt of Chinese urban residents is RMB 52, 000, with a debt level of 54.3 percent and a 3.5 percent annual growth rate. The continued buildup of household loans will significantly increase household financial risks. In order to achieve a dynamic balance between stable growth and risk prevention, it is critical to pay close attention to the sustainability of household debts and prevent the increase in financial vulnerability of the household sector due to household debt risks in the current situation, China is facing severe economic downward pressure and strict prevention of systematic financial risks.
Commercial insurance is considered a household's investment behavior as a measure of risk prevention. The traditional social security system is no longer able to meet the growing needs for household asset preservation and risk avoidance. In August 2014, the State Council issued Opinions on Accelerating the Development of Modern Insurance Services, proposing the establishment of a multi-level social security system, making commercial insurance an important pillar, and clarifying the status of commercial insurance in China's insurance market.
In China, commercial insurance is becoming increasingly vital for promoting economic development and maintaining social stability. On the one hand, the growth of commercial insurance is conducive and beneficial to raising funds for infrastructure construction; improving the financing environment for enterprises; strengthening and stabilizing the entire financial system; and promoting and facilitating economic development. On the other hand, commercial insurance, by reducing financial pressure, helps the government to make better use of financial instruments to meet the most basic security needs, improve the overall level of social security, and promote social harmony and stability. The Chinese commercial insurance industry has developed rapidly since the re-opening of the commercial insurance market in 1980. In January and February 2020, the premium income of the Chinese insurance industry reached RMB 4, 526 trillion, reporting a growth of 6.12% as compared to the same period of the previous year and showing a per capita premium of approximately RMB 3, 200. The depth and density of the Chinese insurance are 66% and 53% of the world's average respectively. One of the main reasons for China's insurance industry's slow development is a lack of insurance awareness among Chinese citizens.
To empirically test the impact of commercial insurance on household financial vulnerability and further to explore the household's heterogeneity characteristics on the mitigation effect of commercial insurance. In order to accomplish the study's aims, this research paper used the probit regression model as a research method to examine the relevant data acquired through the China Household Financial Survey Project (CHFS) in 2017. A robustness test has been conducted in order to address the small sample error problem and to ensure the robustness of the findings. This study focuses on the following two issues: First, whether holding commercial insurance has an effect on reducing the financial vulnerability of households; second, whether heterogeneity has an impact on households with various characteristics. In comparison to the existing literature, the possible innovations and contributions of this paper are as follows.
1) This study adds to the body of knowledge on household financial vulnerability. Household financial vulnerability is an extension of the field of household financial research. It is a holistic evaluation of the household's ability to cope with the risk impact rather than an examination of household hazards generated by a particular risk source. There are few studies on financial vulnerability based on a commercial insurance perspective. Therefore, this study contributes to the existing body of literature in related fields.
2) This paper discusses in detail the differences and disparities among different groups to examine the impact of commercial insurance on financial vulnerability and analyzes the heterogeneity by taking into consideration human capital, social networks, income, urban and rural areas, and regions. Finally, based on the research findings and conclusions, it offers relevant policy recommendations to encourage citizens to participate in commercial insurance to mitigate possible financial risks and achieve the objectives of financial asset allocation, risk prevention, and household management.
2 Literature Review and Theoretical Hypothesis
2.1 Literature Review
The nature of vulnerability is considered a risk. Many scholars believe that household financial vulnerability is reflected in the household's ability to cope with financial risks. Current research shows various factors that affect household financial vulnerability. Financial literacy can significantly reduce the level of financial vulnerability of households
[1]. Li and Zhu
[2] believe that debt leverage ratio increases the financial vulnerability of households. Furthermore, Giarda
[3] argues that quantitative indicators of financial vulnerability boil down to the mix of assets and liabilities, with the risk associated with holding liabilities rising with the gearing ratio but being reduced by the presence of both real and financial assets in the household portfolio. The findings of Brunetti, et al.
[4] reported that the influencing factors of the financial vulnerability of households include households' asset portfolio, income, debt level, age, gender, financial literacy, marriage, etc. From the above research results, it is found that the impact on household financial vulnerability has its own economic and social characteristics, as well as the impact of its response to uncertain factors.
The findings of Lusardi, et al.
[5] showed that households use a variety of mechanisms to cope with financial shocks, with saving being the most common coping method, but not the only one, and that they can also rely on a wide range of support from family and friends, additional labor, and so on, to cope with shocks. Failure to cope with savings, on the other hand, may lead to reduced social ties, making it more difficult to receive credit from family and friends for low-income households that are especially exposed to financial shocks and lack savings themselves. Increased preventive savings and participation in insurance are the two most common approaches for households to deal with financial risks. According to Zhang
[6], research savings can only cover some households' future uncertain expenditures, such as education, pension, and house purchase, but cannot fully cover future uncertain expenditures, such as health shocks. Unexpected events, such as health shocks, may affect the financial status of households from both income and expenditure perspectives. On the one hand, unexpected events may result in a labor shortage and an inability to provide a consistent supply of funds. On the other hand, unexpected expenditures resulting from unforeseen events have risen sharply in the short term, exceeding preventive savings and forcing people to respond to unexpected events by borrowing, etc.
As a result, the household's ability to repay the loan will deteriorate, increasing the household's financial vulnerability. The basic guarantee of insurance is the uncertain risks faced by the residents, which reduces the need for precautionary savings. Commercial insurance can significantly reduce the volatility of future risks
[7]. Although the cost of insurance is higher, it can compensate for unexpected expenditures that savings cannot cover. Therefore, insurance is more advantageous than savings since it can more effectively minimize the degree of household financial vulnerability. The mechanism of insurance's influence on household financial vulnerability is mostly accomplished by boosting solvency, as current premium payments account for a tiny part of household expenses. Economic compensation and loss allocation are two tasks of commercial insurance. It is a new wealth management model that works as an effective supplement to traditional social insurance coverage
[8]. Currently, commercial insurance literature focuses mostly on the factors that influence commercial insurance demand and studies the impact of insurance on residents' consumption of various types of commercial insurance. As a result, this study focuses on the impact of commercial insurance on household financial vulnerability.
2.2 Research Hypothesis
Purchasing insurance is a traditional and widely used risk management method by households. In the short term, external risk shocks will affect households' sustainable income, thus reducing households' precautionary savings. If savings are insufficient to balance the risk shock, households will have to borrow for disease or disaster. Commercial insurance can reduce the uncertainty of the household's primary source of income
[9]. The future income it generates can smooth out the future expenditures and offset the negative impact of external risks through economic compensation. Savings can only cover certain households' future fixed expenses, such as education, pension, and housing, but they cannot fully cover unpredictable costs, such as health issues. First of all, commercial insurance can play the role of risk protection. It can compensate the household for economic losses according to the insured amount in the commercial insurance contract. Second, commercial insurance can serve as credit enhancement financing, such as micro-loan guarantee insurance, which can help households avoid mortgages and reduce their financial vulnerability by applying for bank loans. Based on the above theoretical analysis, a hypothesis (H1) can be generated:
H1 Commercial insurance participation can reduce the financial vulnerability of households.
Because of the heterogeneity of household characteristics, commercial insurance has varying benefits in lowering financial vulnerability. First, households have different regional locations. The eastern region has a higher degree of financial development, with a large number of insurance intermediaries and third-party financial institutions, which can provide customers with diversified products and tailored services. In addition, various insurance companies will provide frequent insurance training and provide financial information to enhance residents' insurance awareness. However, the financial development in the western region is relatively low and the residents' insurance understanding is also limited. Therefore, commercial insurance cannot be used to effectively mitigate possible financial risks. Second, is the difference in residents' income levels. Because commercial insurance is based on contributions, a household's commercial insurance premium can only be afforded when its income reaches a certain level. With the increase in income levels of households, they can afford a variety of insurance products with a high level of protection against accidental risks. Third, the difference in social capital of residents. China is a traditional connected society, and many problems can be solved through social relationship networks. Therefore, social capital is an informal risk-sharing mechanism, and human costs have a significant negative impact on household commercial insurance participation
[10]. Therefore, the probability of high human costs of households participating in commercial insurance is reduced, and commercial insurance cannot fulfil the role of financial risk dispersion. Fourth, the difference in the education level of the household head. The higher the education level of the household head, the higher will be the income and financial literacy. They can choose the appropriate financial instruments to hedge the possible risks in the future. The probability of financial vulnerability is lower than that of households with low education levels. Therefore, the role of commercial insurance is limited. Fifth, urban and rural heterogeneity. Due to the large difference in income and financial literacy between urban and rural residents, the degree of financial market participation varies, urban residents have access to a wider range of financial instruments to spread out possible financial risks, and the impact of commercial insurance on household financial vulnerability is less severe in urban areas than it is in rural ones. Based on the above literature, a proposed hypothesis (H2) can be generated:
H2 The impact of commercial insurance on financial vulnerability varies due to the heterogeneity of household characteristics.
3 Empirical Research Design
3.1 Model Setting
In order to examine the impact of commercial insurance on household financial vulnerability, the following measurement model (Equation (1)) is used.
The probit model is used, where is a dummy variable, 1 indicates a household is financially vulnerable, and 0 indicates a household is not financially vulnerable. The main explanatory variable is the measurement index of participating in commercial insurance (Insurance), including whether the household is insured, the number of households insured, and the rate of participating in commercial insurance. is the control variable, including household characteristic variables, and is an unobservable error term reported in the model.
3.2 Data and Variable Description
3.2.1 Data Sources
This paper examines the cross-sectional data of the China Household Financial Survey (CHFS) in 2017 to empirically analyze the relationship between commercial insurance and household financial vulnerability. The data is representative at the national and provincial levels. The China Household Financial Survey collects information about household assets and liabilities, income and expenses, insurance and security, household demographic characteristics, and employment. The sample was taken from 29 provinces (autonomous regions and municipalities directly under the central government), 355 counties (districts and cities) and 1428 village (neighborhood) committees across the country, excluding missing variables and unobservable values. A total of 39875 samples were collected.
3.2.2 Main Explanatory Variables
Commercial insurance is the main explanatory variable, which primarily includes whether the household is insured, the number of participants in commercial insurance, and the rate of participation in commercial insurance. The CHFS questionnaire investigated the respondents' household members' participation in commercial insurance. If any household member holds more than one commercial insurance policy, the probability of the household purchasing commercial insurance is 1. On the contrary, it is 0. In addition, the questions about premium expenditure in the questionnaire will also become the main basis of this paper. Commercial property insurance holds the biggest amount in the categories of life insurance, health insurance, property insurance, and other insurance, while individual life insurance retains the lowest amount.
3.2.3 Interpreted Variables
Household financial vulnerability refers to the possibility of a household facing financial difficulties or the risk of falling into financial difficulties. It can be defined in three ways. The first method is to use the financial margin. If the financial margin is less than 0, it is deemed fragile. The financial margin is the sum of income and liquidity assets minus expected and unexpected expenditures. The second method is to refer to Brunetti, et al.'s
[4] definition of household financial vulnerability. The households' financial situation is divided into four kinds of situations; one is financial freedom (unconstrained). In this situation, the household income is greater than the expected consumption expenditures and the liquidity assets are greater than the unexpected expenditures. Second is being financially fragile. In this situation, the household's income is greater than expected expenditures but liquid assets are less than unexpected expenditures. Third, excessive consumption and liquidity. Here the household income is less than the expected expenditures but the liquid assets exceed the unexpected expenditures. The fourth is financial constraint. In this situation, the household income is lower than expected expenditures and liquid assets are lower than unexpected expenditures. Income refers to the total household income. Expected expenditures are a household's consumption expenditures, including daily consumption, rent, loans, insurance, etc. Liquid assets include cash, demand deposits, and time deposits. Unexpected expenses include medical expenses and transfer expenses. The household's financial vulnerability reflects the household's ability to cope with future uncertainty, which will be directly impacted by uncertain income and unexpected medical expenses. Therefore, the more likely the household is to be financially vulnerable, the higher the degree of financial vulnerability and the greater the likelihood that the family will slip back into poverty as a result of the impact of uncertainty
1. The third method is to use the definition of insolvency and not make ends meet.
1Yue W, Wang X, Zhang Q. Health risk, medical insurance and household financial vulnerability[J]. China Industrial Economics, 2021(10): 175–192.
3.2.4 Control Variables
Refer to Lu, et al.
[11] and Qin, et al.
[12], control variables are used in the empirical analysis. Such as household demographic characteristics (household location, number of elderly households, age of households' head, gender, education level, marital status) and household financial characteristics (financial market participation, whether to hold social security). Among them, the characteristic variables of the household's head include age, gender, education level (no primary school coded as 1, primary school and junior high school level coded as 2, senior high school and technical secondary school, and vocational high school coded as 3, junior high school, vocational high school and undergraduate high school coded as 4, graduate student and above education is coded as 5), health status (very good and good coded as 1, general, bad and very bad coded as 0), and marital status (married and remarried coded as 1, the rest is 0). The regional characteristic variable selects the virtual variable that the visited household's region belongs to the west, middle, or east (east assigned 1, middle assigned 2, and west assigned 3).
The data collected from 29 provinces of China is divided into eastern, central, and western regions based on their level of economic development. The eastern part refers to the provinces and cities with the earliest implementation of the coastal opening policy and a high level of economic development, including Beijing, Tianjin, Hebei, Liaoning, Shanghai, Zhejiang, Fujian, Shandong, Guangdong and Hainan. The central part refers to the economically underdeveloped areas, including Shanxi, Inner Mongolia, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei, Hunan, and Guangxi. The western part refers to the less developed regions, including Sichuan, Yunnan, Chongqing, Guizhou, Shaanxi, Ningxia, Gansu, Tibet, and Qinghai.
3.2.5 Descriptive Statistics
Table 1 gives descriptive statistics of the main variables. Among them, in terms of commercial insurance participation, 16.6% of households purchased commercial insurance, representing 28.14% of the total sample of number of participants in commercial insurance, with a participation rate of 8.83%. The definition of financial security is used in relation to household financial vulnerability. If the financial security is less than 0, it is termed vulnerable. The average value of household financial vulnerability is 0.3455, with a standard deviation of 0.4755, indicating that 34.55% of households have financial vulnerability characteristics. In terms of control variables, 90.53% of the households own at least one set of housing; the average age of the households is 55.21 years; based on the education level of the households, the average education level of the whole household is calculated as junior high school; 79.34% of the households' heads are men; 85.36% of the households are married; 24.7% of the households are aged over 65; and 52.45% of the households are unhealthy. The financial market participation rate is 10.64%, and the average household social security coverage rate is more than 88%, indicating that social insurance such as social medical insurance and social endowment insurance is owned by most households.
Table 1 Descriptive statistics |
| Variable | Observed value | Average value | Standard deviation | Minimum value | Maximum value |
| Household financial vulnerability | 39, 875 | 0.3455 | 0.4755 | 0 | 1 |
| Commercial insurance (whether the household participates or not) | 39, 875 | 0.1660 | 0.3721 | 0 | 1 |
| Number of participants in commercial insurance | 39, 875 | 0.2814 | 0.7454 | 0 | 1 |
| Participation rate of commercial insurance | 39, 875 | 0.0883 | 0.2291 | 0 | 1 |
| Housing | 39, 875 | 0.9053 | 0.2929 | 0 | 1 |
| Age | 39, 875 | 55.2179 | 14.2261 | 17 | 117 |
| Square of age | 39, 875 | 32.5139 | 15.8801 | 2.89 | 136.89 |
| Education level | 39, 875 | 9.3227 | 4.2009 | 0 | 22 |
| Gender | 39, 875 | 0.7934 | 0.4049 | 0 | 1 |
| Marital status | 39, 875 | 0.8536 | 0.3535 | 0 | 1 |
| Number of the elderly (over 65 years old) | 39, 875 | 0.2470 | 0.4313 | 0 | 1 |
| Unhealthy condition | 39, 875 | 0.5245 | 0.4994 | 0 | 1 |
| Financial market participation | 39, 875 | 0.1064 | 0.3083 | 0 | 1 |
| Social security | 39, 875 | 0.8801 | 0.3249 | 0 | 1 |
4 Analysis of Empirical Results
4.1 Benchmark Regression
This paper uses a probit model to empirically test the relationship between commercial insurance and household financial vulnerability. The main findings showed that commercial insurance has a significant negative impact on household financial vulnerability, reporting a coefficient of (, ). Commercial insurance usually includes life insurance, health insurance, old-age insurance, and property insurance, considerably reducing the enormous unforeseen costs that face households. Therefore, high-income households will purchase commercial insurance as a supplement to social security to reduce the financial vulnerability.
Table 2 Benchmark regression: Probit estimation |
| | (1) | (2) | (3) |
| Commercial insurance | -0.0459***
(0.0066) | - | - |
| Number of participants in commercial insurance | - | -0.0224***
(0.0033) | - |
| Participation rate of commercial insurance | - | - | -0.0565***
(0.0108) |
| Housing | -0.0645***
(0.0078) | -0.0644***
(0.0078) | -0.0656***
(0.0078) |
| Age | -0.0033**
(0.0012) | -0.0033**
(0.0012) | -0.0033**
(0.0012) |
| Square of age | -0.0003
(0.0011) | -0.0002
(0.0011) | -0.0002
(0.0011) |
| Education level | -0.0203***
(0.0006) | -0.0204***
(0.0006) | -0.0203***
(0.0006) |
| Gender | 0.0067
(0.0177) | 0.0070
(0.0061) | 0.0068
(0.0061) |
| Marital status | -0.0557***
(0.0071) | -0.0550***
(0.0071) | -0.0571***
(0.0071) |
| Number of the elderly | 0.0252***
(0.0095) | 0.0255***
(0.0095) | 0.0251***
(0.0095) |
| Number of the unhealthy | 0.0921***
(0.0047) | 0.0919***
(0.0047) | 0.0920***
(0.0047) |
| Financial market participation | -0.1435***
(0.0089) | -0.1442***
(0.0089) | -0.1448***
(0.0090) |
| Social security | -0.0737***
(0.0070) | -0.0737***
(0.0070) | -0.0736***
(0.0070) |
| Sample size | 39875 | 39875 | 39875 |
| R2 | 0.0603 | 0.0603 | 0.0603 |
| Note: ***, ** are significant at a level of 1%, and 5%, respectively.
Robust standard errors are shown in brackets. |
From the perspective of individual level control variables, education level has a significant positive impact on household commercial insurance participation. This may be due to the household's higher education level, more knowledge, stronger perception of risks, and the ability to choose appropriate commercial insurance according to the household's needs. According to household characteristics, households with older members over 65 years old are more likely to purchase commercial insurance. This may be due to the complexity of the household structure, where their needs are more abundant and diversified, and they are more likely in need of commercial insurance. In addition, household loans have a significant impact on households' behavior of buying commercial insurance. Formal loans have a certain impact on reducing poverty in households, which will help to increase income and investment levels for current rural households and improve their ability to manage their assets.
The control variables, including education level, financial market participation, social security, endowment insurance, housing, have a significant negative inhibitory effect on the financial vulnerability of households.
4.2 Addressing Endogenous Problems Caused by Reverse Causality and Missing Variables: Instrument Variable Method
The main explanatory variable in the model, commercial insurance, may have endogenous problems caused by reverse causality and missing variables. There may be a reverse causality problem in commercial insurance participation. Families with higher financial vulnerability have a stronger willingness to participate in commercial insurance. In addition, there may be missing variables that have an endogenous bias towards household financial vulnerability and commercial insurance participation behavior. In order to overcome this issue, this paper chooses the mean value of the commercial insurance community as an instrument variable. The mean value of commercial insurance community reflects the coverage of commercial insurance and the public's insurance awareness to a certain extent. There is a high correlation between the mean value of the commercial insurance community and household participation of insurance, and it is not directly correlated to the explained variables. As a result, the mean value of the commercial insurance community is chosen as the instrument variable. The model is estimated using least squares and ivprobit. Table 3 shows the relevant test results for the endogeneity of the model and the effectiveness of the instrument variables.
Table 3 Estimation of instrument variables |
| | 2SLS | ivprobit | 2SLS | ivprobit | 2SLS | ivprobit |
| Commercial insurance | -0.1341***
(0.0083) | -0.0828***
(0.0081) | - | - | - | - |
| Number of participants in commercial insurance | - | - | -0.0797***
(0.0050) | -0.0192***
(0.0048) | - | - |
| Participation rate of commercial insurance | - | - | - | - | -0.2609***
(0.0163) | -0.0912***
(0.0157) |
| Housing | -0.0621***
(0.0080) | -0.0754***
(0.0078) | -0.0599***
(0.0083) | -0.0724***
(0.0078) | -0.0634***
(0.0083) | -0.0724***
(0.0078) |
| Age | -0.0034***
(0.0012) | -0.0044***
(0.0012) | -0.0035***
(0.0012) | -0.0041***
(0.0012) | -0.0036***
(0.0013) | -0.0042***
(0.0012) |
| Square of age | -0.0004
(0.0011) | -0.0010
(0.0011) | -0.0004
(0.0012) | -0.0006
(0.0011) | -0.0003
(0.0012) | -0.0007
(0.0011) |
| Education level | -0.0201***
(0.0006) | -0.0171***
(0.0006) | -0.0201***
(0.0006) | -0.0177***
(0.0006) | -0.0197***
(0.0006) | -0.0177***
(0.0006) |
| Gender | 0.0048
(0.0061) | 0.0077
(0.0061) | 0.0053
(0.0061) | 0.0049
(0.0061) | 0.0042
(0.0061) | 0.0054
(0.0061) |
| Marital status | -0.0563***
(0.0073) | -0.0561***
(0.0071) | -0.0529***
(0.0075) | -0.0563***
(0.0071) | -0.0617***
(0.0075) | -0.0539***
(0.0071) |
| Number of the elderly | 0.0258***
(0.0095) | 0.0228***
(0.0071) | 0.0269***
(0.0097) | 0.0231***
(0.0094) | 0.0252***
(0.0097) | 0.0234***
(0.0094) |
| Number of the unhealthy | 0.0925***
(0.0049) | 0.0914***
(0.0048) | 0.0916***
(0.0048) | 0.0881***
(0.0047) | 0.0917***
(0.0048) | 0.0880***
(0.0047) |
| Financial market participation | -0.0962***
(0.0080) | -0.1305***
(0.0090) | -0.0932***
(0.0067) | -0.1305***
(0.0090) | -0.0884***
(0.0069) | -0.1328***
(0.0091) |
| Social security | -0.0797***
(0.0073) | -0.0697***
(0.0069) | -0.0799***
(0.0077) | -0.0705***
(0.0069) | -0.0791***
(0.0077) | -0.0705***
(0.0069) |
| Observed value | 39875 | 39875 | 39875 | 39875 | 39875 | 39875 |
| R2 | 0.0692 | 0.0692 | 0.0656 | 0.0656 | 0.0640 | 0.0640 |
| Note: *** is significant at a level of 1%. |
Table 4 reports the results of the test of commercial insurance participation on payout, which has a significant positive effect on payout, i.e., participation in commercial insurance increases payout and makes the liquidity constraint alleviated, thereby reducing household financial vulnerability. Drawing on Li, et al.
[13], this paper defines a liquidity constraint as being present if the value of financial assets is greater than two months of permanent income. The empirical results of this paper show that the participation of commercial insurance effectively alleviates the impact of loss on household mobility through compensation mechanism, thus effectively weakening the financial vulnerability of households. In other words, the family holding commercial insurance can smooth consumption, hedge the huge expenditure brought by emergencies, ease the household saving pressure to a certain extent, and reduce the impact of external risk on the family.
Table 4 Commercial insurance participation and commercial insurance payouts |
| | (1) | (2) | (3) | (4) |
| Commercial insurance | 0.0172***
(0.0013) | - | - | - |
| Commercial life insurance | - | 0.0169***
(0.0019) | - | - |
| Commercial health insurance | - | - | 0.0205***
(0.0021) | - |
| Commercial property insurance | - | - | - | 0.0336***
(0.0018) |
| Housing | 0.0039*
(0.0023) | 0.0044*
(0.0023) | 0.0046**
(0.0023) | 0.0016
(0.0023) |
| Age | 0.0002
(0.0004) | 0.0004
(0.0004) | 0.0004
(0.0004) | 0.0009
(0.0004) |
| Square of age | -0.0007
(0.0004) | -0.0010**
(0.0004) | -0.0009**
(0.0004) | -0.0004
(0.0004) |
| Education level | 0.0008***
(0.0002) | 0.0009***
(0.0002) | 0.0009***
(0.0002) | 0.0002
(0.0002) |
| Gender | -0.0001
(0.0015) | -0.0001
(0.0015) | -0.0001
(0.0015) | -0.0003
(0.0015) |
| Marital status | 0.0036*
(0.0021) | 0.0040*
(0.0021) | 0.0043**
(0.0021) | -0.0033
(0.0021) |
| Financial market participation | 0.0092***
(0.0016) | 0.0106***
(0.0016) | 0.0110***
(0.0016) | 0.0080***
(0.0015) |
| Number of the elderly | 0.0007
(0.0036) | 0.0007
(0.0036) | 0.0004
(0.0036) | 0.0007
(0.0035) |
| Number of the unhealthy | -0.0012
(0.0012) | -0.0012
(0.0013) | -0.0012
(0.0012) | 0.0025**
(0.0012) |
| Social security | 0.0002
(0.0018) | 0.0003
(0.0018) | -0.0001
(0.0018) | -0.0011
(0.0018) |
| Observed value | 39875 | 39875 | 39875 | 39875 |
| R2 | 0.1078 | 0.0904 | 0.0923 | 0.1877 |
| Note: ***, **, * are significant at a level of 1%, 5% and 10%, respectively. |
Table 5 Impact of commercial insurance payouts on liquidity constraints and financial vulnerability |
| | Financial Vulnerability | Liquidity Constraints |
| | OLS | probit | OLS | probit |
| Commercial Insurance payouts | -0.0627***
(0.0192) | -0.0685***
(0.0204) | -0.0992***
(0.0161) | -0.0676***
(0.0131) |
| Housing | -0.0702***
(0.0080) | -0.0687***
(0.0078) | -0.0044
(0.0072) | -0.0021
(0.0072) |
| Age | -0.0040***
(0.0012) | -0.0037***
(0.0012) | -0.0013
(0.0011) | -0.0019*
(0.0011) |
| Square of age | 0.0004
(0.0011) | 0.0001
(0.0011) | 0.0006
(0.0010) | 0.0011
(0.0010) |
| Education level | -0.0237***
(0.0006) | -0.0235***
(0.0006) | -0.0149***
(0.0005) | -0.0145***
(0.0005) |
| Gender | 0.0133**
(0.0061) | 0.0141**
(0.0061) | 0.0375***
(0.0051) | 0.0312***
(0.0049) |
| Marital status | -0.0606***
(0.0073) | -0.0586***
(0.0071) | -0.0087
(0.0061) | -0.0096
(0.0062) |
| Number of the elderly | 0.0254***
(0.0095) | 0.0245***
(0.0095) | 0.0030
(0.0077) | 0.0003
(0.0080) |
| Number of the unhealthy | 0.0961***
(0.0048) | 0.0954***
(0.0048) | 0.0198***
(0.0039) | 0.0186***
(0.0039) |
| Social security | -0.0851***
(0.0073) | -0.0796***
(0.0070) | -0.0202***
(0.0051) | -0.0173***
(0.0054) |
| Observed value | 39875 | 39875 | 26783 | 26783 |
| R2 | 0.0674 | 0.0537 | 0.0437 | 0.0624 |
| Note: ***, **, * are significant at a level of 1%, 5% and 10%, respectively. |
4.3 Testing the Mechanism
4.4 Heterogeneity Analysis
4.4.1 Analysis of Urban and Rural Heterogeneity
The model estimate is done for both the urban sample and the rural sample, respectively, taking into account any potential heterogeneous impacts of urban and rural areas. The results reported in Table 6 demonstrate the negative impact of commercial insurance program participation on household financial vulnerability in both rural and urban households. However, insurance has a better mitigation effect on the financial vulnerability of rural households, as rural households have a lower ability to deal with financial emergencies, leading to a higher probability of financial vulnerability than urban ones. As a result, they can benefit more from the protection of commercial insurance. The understanding of financial situations is also quite different among urban and rural residents. In addition, the urban financial system is much more comprehensive than that in rural areas. As a result, urban residents are more involved in the financial market, and there are a variety of financial instruments to be used in diversifying the financial risks. Therefore, commercial insurance has less negative financial vulnerability impact on urban households than it does on rural areas.
Table 6 Analysis of urban and rural heterogeneity |
| | (1) | (2) | (3) | (4) | (5) | (6) |
| | Rural
Sample | Urban
Sample | Rural
Sample | Urban
Sample | Rural
Sample | Urban
Sample |
| Commercial insurance | -0.0898***
(0.0139) | -0.0330***
(0.0073) | - | - | - | - |
| Number of participants in commercial insurance | - | - | -0.0446***
(0.0073) | -0.0173***
(0.0037) | - | - |
| Participation rate of commercial insurance | - | - | - | - | -0.1472***
(0.0271) | -0.0350***
(0.0117) |
| Control variable | Control | Control | Control | Control | Control | Control |
| Observed value | 12700 | 27175 | 12700 | 27175 | 12700 | 27175 |
| R2 | 0.0225 | 0.0488 | 0.0224 | 0.0489 | 0.0219 | 0.0485 |
| Note: *** is significant at a level of 1%. |
4.4.2 Analysis of Income Heterogeneity
This study also investigated possible heterogeneity arising from household income aspects. The mean value is initially chosen as the dividing line, and all samples are then split into high-income and low-income groups depending on the degree of household income. The results are shown in Table 7. It is reported that commercial insurance has a significant negative impact on the financial vulnerability of low-income households, while it has no significant impact on high-income households. A possible explanation is that low-income families have less wealth and less ability to resist risks. Low-income families may be more likely to fall into poverty due to unexpected expenditures caused by medical treatment or emergencies. Therefore, to reduce or shift the risk to a third party, insurance is considered one of the best ways to protect ourselves. These families will gain more if they are protected by commercial insurance. High-income families are more active in the financial markets due to their higher income and the efficiency of their various financial instruments for risk hedging, like credit cards and insurance. Moreover, the number of insurance products they hold is not large. Therefore, the impact of commercial insurance is not obvious.
Table 7 Analysis of income heterogeneity |
| | (1) | (2) | (3) | (4) | (5) | (6) |
| | High
Income | Low
Income | High
Income | Low
Income | High
Income | Low
Income |
| Commercial insurance | 0.0056
(0.0062) | -0.0491***
(0.0066) | - | - | - | - |
| Number of participants in commercial insurance | - | - | 0.0022
(0.0028) | -0.0249***
(0.0034) | - | - |
| Participation rate of commercial insurance | - | - | - | - | 0.0077
(0.0099) | -0.0598***
(0.0111) |
| Control variable | Control | Control | Control | Control | Control | Control |
| Observed value | 11992 | 27883 | 11992 | 27883 | 11992 | 27883 |
| R2 | 0.0061 | 0.0516 | 0.0061 | 0.0517 | 0.0061 | 0.0512 |
| Note: *** is significant at a level of 1%. |
4.4.3 Analysis of Regional Heterogeneity
Table 8 Analysis of regional heterogeneity |
| | (1) | (2) | (3) |
| | Eastern China | Central China | Western China |
| Commercial insurance | -0.0229**
(0.0092) | -0.0696***
(0.0128) | -0.0700***
(0.0161) |
| Number of participants in commercial insurance | -0.0130***
(0.0045) | -0.0354***
(0.0069) | -0.0345***
(0.0083) |
| Participation rate of commercial insurance | -0.0242*
(0.0147) | -0.0865***
(0.0227) | -0.0913***
(0.0285) |
| Control variable | Control | Control | Control |
| Observed value | 18183 | 11711 | 7640 |
| R2 | 0.0493 | 0.0382 | 0.0596 |
| Note: ***, **, * represent significance level of 1%, 5% and 10% respectively,
other control variables are the same as above. |
The heterogeneity analysis of eastern, central, and western China is carried out to test the heterogeneous effects of different regions. The eastern part includes the provinces and cities that first implemented the coastal opening-up policy and have a higher level of economic development. The central region comprises less developed areas. The western region includes the underdeveloped provinces and autonomous regions. Since households in the central and western regions of China earn less than those in the eastern part of the country. Therefore, their financial vulnerability is greater. As a result, the risks are higher, and participating in insurance plays a crucial role for these households. Families in the eastern region rely less on commercial insurance, which is less important because they earn more money and have more options for coping with financial difficulties, such as loans.
4.4.4 Analysis of Social Capital Heterogeneity
There are notable disparities between the effects of commercial insurance on financial vulnerability for various types of social relationship expenditures. Social relationship expenditure is an important part of total household expenditures. In real life, the principle of "reciprocity'' has become a rule of mutual influence
[14]. Giving each other in the form of money, gifts, or other presents allows people in social relationships to share risks. This makes social relationships, as an informal risk-sharing mechanism, and to some extent, it has a negative impact on the participation of commercial insurance. In other words, the higher the social relationship expenditure, the lower the demand for commercial insurance and the less motivation for families to participate in commercial insurance to deal with financial risks. The risk-sharing mechanism based on human feelings, however, has serious drawbacks. Once an emergency occurs, the risk-sharing ability of social relations will be greatly limited. It is far from sufficient to rely just on the assistance of family members and friends. In the event of an emergency, commercial insurance offers a large amount of guaranteed significantly reducing the financial hardships. As a result, households with lower social expenses have a higher percentage of insurance, which makes it more difficult for those households to manage their finances.
Table 9 Analysis of social network heterogeneity |
| | (1) | (2) | (3) | (4) | (5) | (6) |
| | High social
expenses | Low social
expenses | High social
expenses | Low social
expenses | High social
expenses | Low social
expenses |
| Commercial insurance | -0.0429***
(0.0106) | -0.0700***
(0.0161) | - | - | - | - |
| Number of participants in commercial insurance | - | - | -0.0265***
(0.0053) | -0.0345***
(0.0083) | - | - |
| Participation rate of commercial insurance | - | - | - | - | -0.0603***
(0.0172) | -0.0913***
(0.0285) |
| Control variable | Control | Control | Control | Control | Control | Control |
| Observed value | 10076 | 29799 | 10076 | 29799 | 10076 | 29799 |
| R2 | 0.0652 | 0.0562 | 0.0573 | 0.0472 | 0.0562 | 0.0468 |
| Note: *** is significant at a level of 1%. |
4.4.5 Heterogeneity Analysis of Years of Education of Householder
According to years of education, the households are divided into groups. Those with less than 9 years of schooling are considered to be in the lower education group, and those with more than 9 years of education are considered to be in the higher education group. From the heterogeneity regression results in
Table 10, the findings showed that commercial insurance has a significant difference in financial vulnerability for families with different educational levels. Higher educated households have a better understanding of finance and participate more actively in the financial market
[15]. The financial risk can be dispersed by selecting appropriate financial instruments. Therefore, commercial insurance has a less mitigating effect on household financial vulnerability. Commercial insurance has a significant negative impact on financial vulnerability on households with lower education levels, and the impact on financial vulnerability will increase as the degree of insurance coverage increases (premium increases).
Table 10 Heterogeneity analysis of years of education of householder |
| | (1) | (2) | (3) | (4) | (5) | (6) |
| | Higher
Education | Lower
Education | Higher
Education | Low
Education | Higher
Education | Lower
Education |
| Commercial insurance | -0.0232***
(0.0086) | -0.0641***
(0.0094) | - | - | - | - |
| Number of participants in commercial insurance | - | - | 0.0165***
(0.0043) | -0.0312***
(0.0049) | - | - |
| Participation rate of commercial insurance | - | - | - | - | 0.0246*
(0.0133) | -0.1007***
(0.0172) |
| Control variable | Control | Control | Control | Control | Control | Control |
| Observed value | 14106 | 25769 | 14106 | 25769 | 14106 | 25769 |
| R2 | 0.0469 | 0.0388 | 0.0394 | 0.0284 | 0.0386 | 0.0282 |
| Note: ***, **, * represent significance level of 1%, 5% and 10% respectively. |
5 Robustness Test
1) Taking the regional insurance participation rate as the instrument variable of commercial insurance.
The regional insurance participation rate is used as an instrument variable for commercial insurance to ensure that the results are robust and that the same conclusions as those reached previously are achieved. Based on the results of the robustness test, which are shown in Table 11, the findings showed that the impact coefficient of commercial insurance, the number of participants in commercial insurance and the participation rate of commercial insurance on the household financial vulnerability is significantly negative. This indicates that insured households are less likely to be in a state of financial vulnerability, and the significance of other variables is basically consistent with the regression result, showing that the results are relatively robust.
Table 11 Regional insurance participation rate as an instrument variable for commercial insurance |
| | 2SLS | ivprobit | 2SLS | ivprobit | 2SLS | ivprobit |
| Commercial insurance | -0.0409***
(0.0061) | -2.4198***
(0.0098) | - | - | - | - |
| Number of participants in commercial insurance | - | - | -0.0243***
(0.0036) | -0.0156***
(0.0040) | - | - |
| Participation rate of commercial insurance | - | - | - | - | -0.0779***
(0.0112) | -0.0099***
(0.5596) |
| Housing | 0.0664***
(0.0083) | -0.0649***
(0.0085) | -0.0658***
(0.0083) | -0.0643***
(0.0078) | -0.0668***
(0.0083) | -0.0647**
(0.0078) |
| Age | -0.0034**
(0.0012) | -0.0034**
(0.0012) | -0.0035**
(0.0012) | -0.0033**
(0.0013) | -0.0035**
(0.0012) | -0.0033**
(0.0012) |
| Square of age | 0.0001
(0.0012) | 0.0001
(0.0012) | 0.0001
(0.0012) | 0.0003
(0.0011) | 0.0001
(0.0012) | 0.0003
(0.0011) |
| Education level | -0.0207***
(0.0006) | -0.0194***
(0.0006) | -0.0207***
(0.0006) | -0.0203***
(0.0007) | -0.0206***
(0.0006) | -0.0203***
(0.0006) |
| Gender | 0.0066*
(0.0061) | 0.0058
(0.0069) | 0.0068*
(0.0061) | 0.0060**
(0.0064) | 0.0065**
(0.0061) | 0.0068
(0.0061) |
| Marital status | -0.0581***
(0.0075) | -0.0557***
(0.0071) | -0.0571***
(0.0075) | -0.0551***
(0.0078) | -0.0597***
(0.0075) | -0.0556***
(0.0071) |
| Number of the elderly | 0.0261
(0.0097) | 0.0241*
(0.0094) | 0.0264
(0.0097) | 0.0254*
(0.0095) | 0.0260
(0.0097) | 0.0253*
(0.0095) |
| Number of the unhealthy | 0.0935**
(0.0048) | 0.0843**
(0.0047) | 0.0932**
(0.0048) | 0.0920**
(0.0049) | 0.0932**
(0.0048) | 0.0923**
(0.0047) |
| Financial market participation | -0.1128***
(0.0065) | -0.1365***
(0.0089) | -0.1119***
(0.0066) | -0.1437***
(0.0092) | -0.1107***
(0.0068) | -0.1447***
(0.0090) |
| Social security | -0.0799**
(0.0078) | -0.0731***
(0.0069) | -0.0800***
(0.0077) | -0.0730*
(0.0078) | -0.0797***
(0.0077) | -0.0737*
(0.0070) |
| Observed value | 39875 | 39875 | 39875 | 39875 | 39875 | 39875 |
| R2 | 0.0737 | 0.0737 | 0.0736 | 0.0736 | 0.0731 | 0.0731 |
| Note: ***, **, * represent significance level of 1%, 5% and 10% respectively. |
2) Changing the regression method
The probit regression method is replaced with logit regression. When the dependent variable is a nominal variable, logit and probit have no essential difference and can be changed in general. The distribution functions used to compare the two are different. A probit model is based on a normal distribution, where a logit model is based on a logical probability distribution. After changing the regression model as reported in Table 12, commercial insurance has a significant negative impact on household financial vulnerability, with a coefficient of 0.0496. Each regression variable's estimated parameter direction and significance are consistent with the parameter values predicted by the probit model, showing that the empirical results are robust.
Table 12 Changing regression method (Logit) |
| | (1) | (2) | (3) |
| Commercial insurance | -0.0496*** | - | - |
| (0.0067) |
| Number of participants in commercial insurance | - | -0.0248*** | - |
| (0.0035) |
| Participation rate of commercial insurance | - | - | -0.0649*** |
| (0.0113) |
| Control variable | Control | Control | Control |
| Sample size | 39881 | 39881 | 39881 |
| R2 | 0.0528 | 0.0528 | 0.0528 |
| Note: *** is significant at a level of 1%. |
3) Winsorizing
As there are many variables used in the empirical analysis, in order to prevent some extreme values from affecting the estimation bias, the study has performed winsorizing of 1% for commercial insurance, the number of participants in commercial insurance, and the participation rate of commercial insurance. The value of 1% is assigned to the number less than 1%, and the value of 99% is assigned to the number greater than 99%, so as to test the robustness of the main conclusions of this paper. As the results are shown in Table 13, the regression results are basically consistent with the previous one. There is a significant negative impact of commercial insurance on household financial vulnerability, with a coefficient of 0.0460. After winsorizing, the results are stable.
Table 13 Winsorizing of 1% for commercial insurance, the number of participants incommercial insurance and the participation rate of commercial insurance |
| | (1) | (2) | (3) |
| Commercial insurance | -0.0460*** | - | - |
| (0.0066) |
| Number of participants in commercial insurance | - | -0.0249*** | - |
| (0.0037) |
| Participation rate of commercial insurance | - | - | -0.0567*** |
| (0.0111) |
| Housing | -0.0651*** | -0.0649*** | -0.0661*** |
| (0.0078) | (0.0078) | (0.0078) |
| Age | -0.0023* | -0.0023* | -0.0024* |
| (0.0013) | (0.0013) | (0.0013) |
| Square of age | -0.0013 | -0.0012 | -0.0012 |
| (0.0013) | (0.0013) | (0.0013) |
| Note: ***, **, * represent significance level of 1%, 5% and 10% respectively. |
Table 14 Excluding the Elderly over 65 |
| | (1) | (2) | (3) |
| Commercial insurance | -0.0430*** | - | - |
| (0.0071) |
| Number of participants in commercial insurance | - | -0.0207*** | - |
| (0.0036) |
| Participation rate of commercial insurance | - | - | -0.0553*** |
| (0.0117) |
| Housing | -0.0755*** | -0.0754*** | -0.0765*** |
| (0.0095) | (0.0095) | (0.0095) |
| Age | -0.0006 | -0.0006 | -0.0009 |
| (0.0022) | (0.0022) | (0.0022) |
| Square of age | -0.0032 | -0.0031 | -0.0028 |
| (0.0023) | (0.0023) | (0.0023) |
| Education level | -0.0200*** | -0.0200*** | -0.0200*** |
| (0.0008) | (0.0008) | (0.0008) |
| Gender | -0.0068 | -0.0065 | -0.0068 |
| (0.0072) | (0.0072) | (0.0072) |
| Marital status | -0.0609*** | -0.0598*** | -0.0626*** |
| (0.0090) | (0.0090) | (0.0090) |
| Number of the unhealthy | 0.0898*** | 0.0896*** | 0.0897*** |
| (0.0055) | (0.0055) | (0.0055) |
| Financial market participation | -0.1362*** | -0.1367*** | -0.1369*** |
| (0.0101) | (0.0102) | (0.0101) |
| Social security | -0.0625*** | -0.0625*** | -0.0624*** |
| (0.0076) | (0.0076) | (0.0076) |
| Sample size | 29002 | 29002 | 29002 |
| R2 | 0.0567 | 0.0567 | 0.0567 |
| Note: *** is significant at a level of 1%. |
4) Excluding the older respondents over 65 years
Supporting the older is a responsibility that falls on the family. Its population structure's complexity will have an effect on how well its function is realized. The higher the family's labor force participation rate, which frequently indicates the diversity and stability of its economic sources, the stronger the family's resistance to various risks such as disasters, diseases, unemployment, etc. On the contrary, as the number of dependents in the family gradually increases, the family's security function gradually weakens. The number of elderly people over 65 accounts for about 25% of the family population. With the increase of age, the risk coefficient will increase and the risk tolerance will decrease
[16]. Commercial insurance is an important supplement to family security, so there is an increase in the demand for commercial insurance. Benefiting from the insurance, the pressure of providing for the elderly will be relieved, and the risks families may face will be transferred from the inside to the outside. Excluding the elderly over 65 years old, the impact of commercial insurance on household financial vulnerability is significantly negative, showing that the results are still robust.
6 Conclusions and Policy Recommendations
6.1 Main Conclusions
This study examines how commercial insurance affects household financial vulnerability using data from the China Household Finance Survey (CHFS). The instrument variable method has been applied to better deal with the endogenous problem caused by missing variables to analyze social capital heterogeneity, urban-rural heterogeneity, income heterogeneity, households' years of education, and regional heterogeneity. The main conclusions of this study are as follows:
1) The purchase of commercial insurance has a significant positive effect on reducing the household financial vulnerability, suggesting that commercial insurance should be considered as a key strategy for preventing household financial vulnerability.
2) Education level, financial market participation, social security-social endowment insurance, and housing, etc., have a significant negative inhibitory effect on household financial vulnerability.
3) By participating in commercial insurance, households in areas with low human capital, low income, low consumption in social interactions, rural areas, and the central and western regions experience a large reduction in their financial risk.
6.2 Policy Recommendations
The following policy recommendations are made in light of the findings of this study.
1) To reduce the cost of participation and promote commercial insurance to families.
To design a tax policy to encourage families to participate in commercial insurance. For instance, commercial medical insurance offers a greater payment ratio and covers more diseases than social security, which is a good approach to shift health risks. The findings showed that commercial insurance has a positive significant role in improving household financial vulnerability. The higher price is an important factor that restricts the demand for family commercial insurance. Therefore, the policy of individual tax extension can be implemented in commercial insurance because it allows people to pay for all of their family's medical expenses before tax and lowers the high cost of commercial insurance, which encourages family participation.
2) To meet the family's diverse and personalized insurance needs and increase insurance penetration.
Actively develop local commercial insurance and commercial insurance products that meet the needs of the general public. In order to improve the service level of insurance companies, they should take the initiative to conduct market research to understand consumer needs, customize different insurance products according to the heterogeneity of the regions where families live and the income levels of those families, and improve the quality of their services in order to provide better and quicker protection for the low and middle income groups as well as the central and western regions.
3) To increase the understanding of commercial insurance knowledge in rural areas and investigate the rural commercial insurance market for consumption.
The financial and insurance systems in rural areas are not yet perfect. In order to assist rural development, it is crucial to establish and strengthen the financial and insurance systems, insurance outlets, and insurance specialists in rural areas. Additionally, emphasis must be paid to the differences in financial literacy between urban and rural residents; insurance education and training must be strengthened; and to actively promote financial and insurance knowledge. We can only do this if we want to protect farmers' legitimate financial rights and interests while also doing our best to prevent economic losses brought on by illness and accidents. In addition, as China's current commercial insurance coverage is not sufficiently comprehensive. Therefore, a "low yield" policy should be adopted as well, with insurance companies receiving preferential rates and tax incentives as well as government guarantees, purchases, and subsidies to lessen the purchasing pressure on rural low-income families.
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
PDF(161 KB)
Table 1 Descriptive statistics
