The Effect of Dependency Burden on Household Entrepreneurial Exit Behavior: Empirical Evidence from Chinese Households

Abstract

The elderly dependency ratio in China has increased from 10.7% in 2003 to 17.8% in 2019, and there may be a link between the widespread withdrawal of family entrepreneurship in society and the increased dependency burden. Based on China Family Panel Studies, we empirically examine the impact of dependency burden on household entrepreneurial exit behavior. We find that the likelihood of household exit from entrepreneurial activity increases as the population ages. Compared to active exit from entrepreneurship, dependency burden mainly induces failure exit from household entrepreneurship. In addition to reducing physical capital accumulation and labor supply, dependency burden also reduces the household workforce’s mental and physical health, leading to the entrepreneurial exit. The application of digital technology can mitigate the negative effects of dependency burden in household entrepreneurship. The mitigating effect of digital technology will be enhanced with the deepening of the application. Therefore, government and academia need to pay attention to the negative effects of dependency burden on the sustainability of household entrepreneurship.

1. Introduction

Innovation and entrepreneurship are the basic driving force for developing countries to defuse the pressure of economic growth and promote full and high-quality employment. Under the policy guidance, social subjects with passion and dreams have set off a wave of entrepreneurship. Family-based subjects occupy an important position in the wave of entrepreneurship [1]. However, in contrast to the “entrepreneurship wave” phenomenon, the proportion of households exiting entrepreneurship has been increasing yearly (The China Household Finance Survey shows that the proportion of household entrepreneurial exits in China was 33% and 41% in 2015 and 2017, respectively, data obtained by the authors). The 2019 Global Entrepreneurship Monitor report also shows that the exit ratio of Chinese entrepreneurs is close to one fourth of new ventures. Theoretically, both entry and exit of entrepreneurship are in line with the market law and a normal phenomenon under the entrepreneurship trend. However, it is important to note that frequent entry and exit costs can be costly and disruptive to the market, especially due to poor performance [2,3]. For example, start-ups provide relatively limited employment and social wealth creation, and firms with long continuation cycles tend to have a more dynamic role in socioeconomic development [4]. Therefore, as China enters a phase of high-quality economic development, it is essential to pay attention to the exit of entrepreneurship hidden under the “entrepreneurial boom” phenomenon.

Entrepreneurial exit is a common global phenomenon. By searching for topics such as “entrepreneurship”, “founder”, “exit”, or “entrepreneurial exit”, we found that previous literature focused on the start-up and growth phases of entrepreneurship and discussed more entrepreneurial opportunity identification and utilization, resource acquisition and integration, but less on the exit of entrepreneurship. However, the entrepreneurial process is not only limited to the entrepreneurial start-up and growth phases but also includes the entrepreneurial exit phase [5]. Current entrepreneurial exit research can be divided into three orientations, process-oriented, feature-oriented, and event-oriented [6]. The process orientation emphasizes how time evolves to trigger outcomes. For example, DeTienne [7] defines entrepreneurial exit as the process by which a private firm’s founders or entrepreneurial team leave the company and exit from the primary ownership and decision-making structure. As defined in this paper, entrepreneurial exit is based on a feature-oriented approach that focuses on the covariance between different concepts. This paper considers entrepreneurial exit as the act of entrepreneurial exit of micro-entities that make decisions on a household basis. Why do people exit entrepreneurship? Scholars have given explanations based on different theories. For example, the firm lifecycle theory suggests that different stages of entrepreneurship face differentiated opportunities and challenges and that entrepreneurs exit entrepreneurial activities when their capabilities cannot meet the firm’s growth and development needs [8]. Wennberg et al. [9] constructed a general framework in which they argue that factors influencing entrepreneurial exit behavior include macro, organizational and individual factors. The macro-level includes institutions such as economic development, industrial structure characteristics, culture and norms, and the organizational level includes the firm’s financial performance and growth potential. The individual level is the entrepreneur’s family life and social network. Some other scholars analyze the factors influencing entrepreneurial exit behavior in terms of human capital [10], personality traits [11] and family factors [12].

Overall, the process of entrepreneurial exit is very complex, and scholars’ explanations of the entrepreneurial exit phenomenon remain limited. Currently, scholars focus on the performance problems of firms leading to entrepreneurial exit at the macro and organizational levels and the micro-levels on the characteristics of entrepreneurs. The existing literature mostly considers the antecedents of entrepreneurial exit in terms of individual and family characteristics. The demographic structure is an important manifestation of household characteristics, and demographic structure is inextricably linked to the amount of economic activity and economic behavior of households [13]. While previous literature has analyzed the impact of the number of children on entrepreneurial exit [12], fewer studies have focused on the impact of the number of older members of the household on entrepreneurial exit, especially in the context of dependency burden that human societies need to face together. China’s old-age dependency ratio rose from 10.7 percent in 2003 to 17.8 percent in 2019, according to China’s National Bureau of Statistics [14]. In addition, we observed the China Household Finance Survey and found that the percentage of those who ended their entrepreneurial operations due to family factors was 37.4% and 36.7% in 2015 and 2017, respectively. Therefore, there may be a link between the increased dependency burden and the phenomenon of household entrepreneurial exit.

Influenced by the long-standing Chinese culture of “filial piety”, intergenerational transfer with the core of old age and filial piety is a traditional family function. When household demographics tend to age, upward intergenerational transfers from children’s generation may inhibit household entrepreneurial activities [15]. For example, intergenerational household transfers reduce households’ social interaction expenditures and reduce risk tolerance [1]. Dependency burden can facilitate entrepreneurial exit by crowding labor supply and material capital accumulation through intergenerational transfers [16]. Women members of households tend to be more affected by dependency burden because they usually need to take on the role of caring for the elderly, which inhibits entrepreneurial activity [17]. Thus, dependency burden may be important in household exit from entrepreneurship. In terms of mechanism research, previous literature has focused on exploring risk tolerance and social interaction ability, ignoring human capital, which plays an essential role in economic activity. As defined by Schultz [18], human capital includes individuals’ knowledge, skills, abilities, and health. Among them, health is an essential component of human capital measurement and can be called health capital [19]. The impaired physical and mental health of caregivers resulting from family caregiving activities may be an important influence mechanism in the impact of dependency burden on household entrepreneurial exit behavior.

Our study explores whether there is some connection between the phenomenon of household entrepreneurial exit, widespread in society, and the phenomenon of dependency burden. In other words, we will explore the relationship between dependency burden and household entrepreneurial exit behavior from the perspective of entrepreneurial household characteristics. Based on this, we will expand the health capital transmission mechanism of dependency burden, affecting household entrepreneurial exit behavior. We want to discover more adverse effects of dependency burden on economic and social development to provide some suggestions for advancing quality entrepreneurship in developing countries. Possible innovations of this paper are listed as follows. First, we focus on the impact of dependency burden on household entrepreneurial exit behavior hidden under the phenomenon of entrepreneurial boom from a micro perspective. Second, we distinguish between entrepreneurial exit motives, i.e., active exit or failure exit, and observe whether dependency burden induces failure exit or active exit from household entrepreneurship. Third, we introduce health capital from human capital theory in the mechanism study, and we focus on the impact of health problems induced by dependency burden on entrepreneurial exit behavior. Fourth, the moderating role of digital technology use in the context of digital economy development in affecting household entrepreneurial exit by dependency burden is discussed.

2. Theoretical Analysis

Influenced by traditional Confucianism, the long-standing “culture of filial piety” is deeply engraved in the bones of Chinese people. The Book of Filial Piety mentions this: “Use the ways of heaven, divide the benefits of the earth, be careful with your body and use it sparingly to support your parents, and this is the filial piety of the common people”. Family caregiving activities are a bottom-up transfer of resources based on intergenerational transmission [20]. In addition, household old age dependency ratio increase symbolizes the “aging” and “decline” of families in terms of human capital and physical capital accumulation. Household entrepreneurship is continuously acquiring resources, developing them and integrating them [21]. Among the many resources, households’ physical and human capital are important prerequisites for entrepreneurial survival and development. Therefore, this paper analyzes how dependency burden affects households’ entrepreneurial activities through entrepreneurial resources and thus from the perspective of family caregiving.

Within families, wealth transfer is a common phenomenon in society. Along with the law of life, frailty and illness become the norm for most elderly family members. Children of the family consume physical capital stock to support the daily living expenses and medical expenses of the elderly [22]. When older members become ill, medical-related expenditures such as hospitalization costs and health insurance product purchases become very “rigid” in household consumption [23]. Undoubtedly, a dependency burden may deplete the stock of material capital available to households for entrepreneurial activities. In addition, when households face support needs, some members may avoid high-wage-high-risk jobs to ensure household livelihood stability. The tendency toward stable work can lead to potential opportunity costs for households and reduce physical capital accumulation. In the aging trend, physical capital is gradually shifted to older people, making household entrepreneurial survival and development face limitations or even decline problems, which leads households to exit from entrepreneurial activities.

Due to the problem of physical decline, aging household members usually find it difficult to continue to engage in productive business activities. Therefore, dependency burden directly reduces the effective labor supply of households. In addition, the deepening of dependency burden causes an increase in the burden of family caregiving, indirectly reducing total family labor time. Caring for the elderly is a time-intensive activity because daily activities such as walking around indoors, putting on clothes, and going to the bathroom occur often. Caring for the elderly causes the labor time of other family members in entrepreneurial operations to be shifted to the elderly, and labor time becomes fine-tuned by caregiving shocks. It has been found that the female labor force participation rate decreases when households are in caregiving status due to dependency burden [24]. As entrepreneurial households move into an aging population, the large time commitment to older members can reduce the labor supply allocated to entrepreneurial business activities. The reduction in labor time may be detrimental to the sustainable growth of household entrepreneurship, thus forcing households to exit from entrepreneurial activities.

In different social relationships, a single individual may focus on multiple roles. However, it is difficult for an individual to play all roles within a limited time constraint, leading to a state of role overload. Role overload leads to role conflict, resulting in individuals choosing different roles [25]. Many studies have confirmed that role overload and role conflict are detrimental to an individual’s level of physical and mental health [26]. In families with aging populations, family members are required to play the role of caregivers for the elderly and entrepreneurial operators. The role conflict leads to negative mental health states among family members, and prolonged mental health imbalance also induces physiological health abnormalities. Reduced physiological health affects the labor supply in entrepreneurial activities, but more importantly, it reduces learning efficiency and cognitive ability, which is detrimental to the labor productivity of entrepreneurial activities [19]. At the same time, mental health status affects entrepreneurs’ evaluation of entrepreneurial activities and their perception and evaluation of their abilities. Unhealthy psychological levels can lead to negative evaluations of entrepreneurship and oneself. Mental health imbalance can also affect family learning about entrepreneurship, leading to entrepreneurial exit [27].

Overall, the physical and human capital possessed by the household is an important prerequisite for entrepreneurial survival and development. When entrepreneurial resources cannot meet the production demands of entrepreneurship, it leads to the withdrawal of households from entrepreneurial activities. Dependency burden consumes the entrepreneurial resources households possess through intergenerational transfer, including physical capital, labor supply, and health capital. The imbalance between the supply of entrepreneurial resources and resource demand leads to the household entrepreneurial exit.

Hypothesis 1.

Dependency burden induces entrepreneurial exit behavior in households.

Hypothesis 2.

Dependency burden leads to household entrepreneurial exit mainly by reducing the household’s physical capital stock, labor supply, and health capital stock.

3. Data, Variables and Models

3.1. Data Sources

This paper uses data from the China Family Panel Studies (CFPS) project conducted by the China Social Science Survey Center (ISSS) of Peking University in 2016 and 2018. The survey covered 25 provinces/municipalities/autonomous regions in mainland China (Except for Hong Kong, Macau, Taiwan, Xinjiang, Tibet, Qinghai, Inner Mongolia, Ningxia, Hainan). This paper uses data from the adult, child, and household pools of the 2016 survey and the 2018 household pool.

According to the study, the sample selected for this paper met two conditions. First, the tracking sample households were interviewed for the survey in 2016 and 2018. Second, the sample households need to be engaged in home business activities in 2016. After processing missing data and outliers, 1201 sample households were obtained. To ensure the robustness of causal identification, dependency burden and a series of control variables are used for 2016.

3.2. Variable Definition

The dependent variable is household entrepreneurial exit behavior. DeTienne [7] defines entrepreneurial exit as when founders leave the business they have created and exit from its primary ownership and decision-making structure. However, most household entrepreneurial units are small organizational structures such as self-employment and micro and small enterprises. The main participating members are family laborers, showing strong family dependence. Unlike large corporations, family business exit is manifested by the overall exit from entrepreneurial business activities with the family as the decision-making unit. First, this paper defines household entrepreneurship in China as industrial and commercial operations in the non-farm sector, including private enterprises or self-employment, i.e., nonfarm entrepreneurship. Second, a combination of previous research and data availability [28]. This paper identifies entrepreneurial exit by identifying a household as having exited entrepreneurial activity in 2018 if the respondent household was engaged in entrepreneurial activity at the time of the 2016 visit but ended the entrepreneurial business project at the time of the 2018 visit. In this paper, entrepreneurial exit households are assigned a value of 1, and entrepreneurial survival households are assigned a value of 0.

The core independent variable is old-age dependency burden. For households, dependency burden is reflected in the number of older persons in the household as a proportion of the total population. When the proportion of older people within a household increases, the dependency burden is deeper. By the World Health Organization’s definition of older adults, this paper considers household members aged 60 years and above as the elderly population.

Control variables: Many other factors influence household entrepreneurial exit behavior in addition to demographics. The control variables selected in this paper include individual-level and household-level exit. At the individual level, the head of the household often holds the primary decision-making power. Therefore, the authors controlled for the household head’s age, gender, health status, marital status, and years of education. Household-level control variables include labor-force size, number of children, government contributions, human expenditures, the value of household durables, risk preferences, bank and private loan status, and household savings. There are differences between urban and rural market and institutional environments, and the authors control for urban-rural dummy variables to avoid unwanted effects. The authors control province characteristics to prevent the effect of unobservable factors at the regional level. Descriptive statistics results refer to

Table 1. Definition of variables and descriptive statistical analysis.

Variable Name	Variable Definition	Mean	S.D.
Entrepreneurial exit	Household entrepreneurial exit: exit = 1, no exit = 0	0.418	0.493
Dependency burden	The proportion of the population aged 60 and over to household size	0.127	0.214
Age	The actual age of the head of household	45.66	12.72
Gender	male = 1, female = 0	0.510	0.500
Marriage	Marriage of head of household: marriage = 1, other = 0	0.899	0.301
Health status	Head of household health: healthy = 1, unhealthy = 0	0.679	0.467
Education	Years of education of the head of household (years)	7.769	4.742
Number of laborers	Number of working-age people in the household	2.486	1.289
Number of children	Number of juveniles under 16 years of age in households	0.837	0.971
Government donation	Have received government donations: Yes = 1, No = 0	0.412	0.492
Interaction expenses	Logarithmic value of transfer expenditures to nonhousehold members	7.777	2.138
Consumer durables	The logarithm of the value of consumer durables owned by households	9.784	2.584
Risk appetite	Household ownership financial assets indicates objective risk appetite: Yes = 1, No = 0	0.066	0.025
Bank loans	Household loan from bank: Yes = 1, No = 0	0.133	0.340
Private borrowing	Household borrowing money from friends and relatives: Yes = 1, No = 0	0.178	0.383
Value of savings	Logarithmic value of cash and savings in the household	7.725	4.717
urban or rural	Urban households = 1, rural households = 0	0.596	0.491

. The results show that the proportion of households exiting entrepreneurship between 2016 and 2018 is 41.8%, reflecting that the current proportion of households exiting entrepreneurship is not low and deserves the attention of the government and society.

3.3. Estimation Strategy

Household entrepreneurial exit behavior is a binary dummy variable of 0 and 1. Theoretically, either Probit or Logit models can be chosen as the empirical strategy. These two models do not differ much in terms of estimation results. However, the Logit model is superior to the Probit model. It is more sensitive when dealing with rare events or predicting probabilities close to endpoint values, which is beneficial to the accuracy of the estimation results. The following logit model is designed to test the relationship between dependency burden and household entrepreneurial exit behavior, as shown in Equation (1).

$$Y_i={\beta }_0+{\beta }_1\sum Agin{g}_i+\sum {\beta }_j{X}_i+{\epsilon }_i$$

(1)

In Equation (1), Y is a dichotomous variable for household exit from entrepreneurship, when Y = 1 means that the household exits from entrepreneurship and Y = 0 mean that the household does not exit from entrepreneurship. $Agin{g}_i$ denotes the degree of household dependency burden and is a continuous type variable in the interval (0, 1). $X_i$ containing control variables such as individual-level and household-level characteristics. ${\epsilon }_i$ are systematic error terms. The Probit model and LPM model were also estimated simultaneously to ensure robustness.

4. Empirical Results and Analysis

4.1. The Impact of Dependency Burden on Household Entrepreneurial Exit Behavior

Table 2. Regression results of the effect of dependency burden on household entrepreneurial exit behavior.

Variables	Entrepreneurial Exit
	(1) Logit	(2) Logit	(3) Probit	(4) Probit	(5) Lpm	(6) Lpm
Dependency burden	0.153 ** (0.073)	0.167 ** (0.082)	0.154 ** (0.073)	0.167 ** (0.081)	0.156 ** (0.074)	0.170 ** (0.083)
Age	0.000 (0.001)	−0.001 (0.001)	0.000 (0.001)	−0.001 (0.001)	0.000 (0.001)	−0.001 (0.001)
Gender	0.017 (0.028)	0.013 (0.028)	0.017 (0.028)	0.013 (0.028)	0.017 (0.028)	0.013 (0.029)
Marriage	−0.179 *** (0.046)	−0.166 *** (0.048)	−0.179 *** (0.046)	−0.164 *** (0.048)	−0.183 *** (0.048)	−0.169 *** (0.050)
Health status	−0.055 * (0.031)	−0.047 (0.031)	−0.056 * (0.031)	−0.047 (0.031)	−0.056 * (0.032)	−0.048 (0.032)
Education	−0.013 *** (0.003)	−0.010 *** (0.003)	−0.013 *** (0.003)	−0.010 *** (0.003)	−0.013 *** (0.003)	−0.010 *** (0.003)
Number of laborers		0.008 (0.013)		0.008 (0.013)		0.008 (0.013)
Number of children		−0.007 (0.015)		−0.009 (0.015)		−0.007 (0.015)
Government donation		0.043 (0.029)		0.043 (0.029)		0.044 (0.030)
Interaction expenses		−0.002 (0.007)		−0.002 (0.007)		−0.002 (0.007)
Consumer durables		−0.013 ** (0.006)		−0.013 ** (0.006)		−0.013 ** (0.006)
Risk appetite		0.028 (0.061)		0.028 (0.060)		0.032 (0.058)
Bank loans		−0.030 (0.042)		−0.031 (0.042)		−0.030 (0.042)
Private borrowing		−0.000 (0.038)		−0.000 (0.038)		−0.001 (0.039)
Value of savings		−0.006 * (0.003)		−0.006 * (0.003)		−0.006 * (0.003)
Urban orrural		−0.078 *** (0.029)		−0.078 *** (0.029)		−0.080 *** (0.030)
Province		Control		Control		Control
Intercept term	-	-	-	-	0.680 *** (0.084)	0.919 *** (0.114)
Sample size	1201	1201	1201	1201	1201	1201
Wlad chi2	46.35 ***	66.09 ***	47.99 ***	68.72 ***	8.96 ***	4.84 ***
Pseudo R2	0.0318	0.0471	0.0316	0.0468	0.0429	0.0626

Note: (1) This table reports marginal effects of explanatory variables; (2) Figures in parentheses are robust standard errors of coefficients; (3) ***, **, and * indicate significance at 1%, 5%, and 10% statistical levels, respectively; (4) The mean value of VIF for multicollinearity test is 1.19, and there is no serious multicollinearity problem. The following tables are interpreted in the same way as Table 2.

reports the baseline results of regression estimation using the Logit model, to which the Probit model and the linear probability model LPM were added for regression. All three regression models contain two subregression results, one with the addition of core explanatory variables and controlling for individual-level characteristic variables, and the other with the gradual addition of more control variables such as household-level characteristic variables, and urban–rural classification, and province characteristics. Comparing the results of the three regression models reveals that the choice of the econometric model and the addition or subtraction of control variables only slightly affect the magnitude of the estimated coefficients of dependency burden without affecting the direction of influence or significance.

The results show that the dependency burden variable positively affects the entrepreneurial exit behavior of households at the 1% statistical level. An increase in the elderly population of household-based entrepreneurial subjects increases the likelihood of household exit from entrepreneurial activities, consistent with Hypothesis 1. Regarding economic implications, when the dependency burden level increases by 1%, the likelihood of households exiting entrepreneurial business projects increases by about 17%. Changes in household demographics are an important factor in the entrepreneurial exit. In 2020, China’s population aged 60 and above is 18.7%, and it is expected that in 2026 it will be a “moderately aging” society. By 2026, the population aged 60 and above will grow to 20.8%. If the external environment remains unchanged, the proportion of household entrepreneurial exits in China will increase by 34%.

The control variables show that the marital status of the household head has a significant negative effect on household entrepreneurial exit behavior. Marriage is a very important life node. After completing this node event, the household is more stable, which helps continue operation of entrepreneurial activities. The education level of the household head reflects human capital. When human capital is abundant, entrepreneurial performance is better, thus discouraging households from withdrawing from entrepreneurial activities, consistent with the consensus that human capital is important for entrepreneurship. The value of consumer durables and savings significantly negatively affect entrepreneurial exit activity. These two factors symbolize the stock and flow of physical capital of households. The richer the stock and flow of physical capital, the more factors of production households can invest in entrepreneurial activities, which affects entrepreneurial performance and exit behavior. The entrepreneurial environment, such as the marketization process and institutional environment, is better in urban than rural areas, making the exit possibility of entrepreneurial households significantly higher in rural areas than in urban areas.

4.2. “Active Exit” or “Failure Exit“?

Entrepreneurial exit is usually bundled with entrepreneurial failure, and scholars mostly see entrepreneurial exit as eclipsing the entrepreneurial arena due to failure [29]. However, some studies have shown that this is not entirely the case. Other non-economic reasons (e.g., good educational opportunities, employment opportunities, etc.) for entrepreneurial exit and poor entrepreneurial performance lead to exit [30]. Thus, the firm’s status at the exit, exit strategies, etc. can be divided into failure exits due to entrepreneurial economic distress and active exits due to non-economic factors [31]. Is the exit of Chinese households from entrepreneurship motivated by economic circumstances or other hardships, or is it a proactive gesture to exit by finding better employment opportunities? Suppose the exit from entrepreneurship is due to economic hardship or other reasons. In that case, there is an urgent need for policy attention and improvement in the hardship faced by micro-entrepreneurial agents. This paper provides a heterogeneous discussion of exit motivation. The dependency burden leads to a downward and upward flow of family resources. Family caregiving activities reduce the allocation of family resources in entrepreneurial business activities, leaving household entrepreneurship in relative distress and thus passive withdrawal from entrepreneurial activities. Therefore, this study concludes that dependency burden may lead to failure exit from household-based entrepreneurship and has no significant effect on active exit from household-based entrepreneurship. It is difficult to measure entrepreneurial exit motivation directly, and there are no relevant questions in the Chinese Household Tracking Survey database. This study intends to use the opportunity cost comparison method according to economic profit [27]. The average value of the household’s operating income is used as the benchmark in the current period. If the operating income of the exiting entrepreneurial household is lower than the benchmark, it is regarded as a failure exit. If the operating income of the exiting entrepreneurial household is higher than the benchmark line, it is considered an active exit. By processing the exit motives, we find that the proportion of households’ failure to exit entrepreneurship is about 90% of the total exit, reflecting that most current micro-entrepreneurial subjects exit entrepreneurship passively due to economic factors and other difficulties.

Table 3. Results of the test of heterogeneity of entrepreneurial exit motivation.

Variables	Failure Exit		Active Exit
	(1)	(2)	(3)	(4)
Dependency burden	0.143 ** (0.074)	0.175 ** (0.082)	0.077 (0.065)	0.044 (0.067)
Individual characteristics	Control	Control	Control	Control
Family characteristics		Control		Control
Province		Control		Control
Sample size	1100	1100	800	800
Wlad chi2	56.49 ***	98.06 ***	18.41 ***	47.57 ***
Pseudo R2	0.0417	0.0741	0.0282	0.0851

*** and ** indicate significance at 1% and 5% statistical levels.

demonstrates the effect of dependency burden on household motivation to exit entrepreneurship. (2) and (4) add household characteristics and province characteristics variables to (1) and (3), respectively, and the results show that the magnitude and significance of the coefficients of dependency burden do not fluctuate much, reflecting the results with some credibility. The results of (2) and (4) show that dependency burden significantly positively affects household entrepreneurial failure exit at the 5% statistical level. There is no significant effect of dependency burden on active exit from household entrepreneurship. Reflecting the structural changes within households due to dependency burden, which induce households to exit from entrepreneurial activities due to economic hardship, etc., is consistent with the theoretical analysis.

4.3. Testing the Mechanisms of Dependency Burden on Household Entrepreneurship Exitdependency Burden

4.3.1. The Impact of Dependency Burden on Physical Capital Accumulation

The theoretical analysis reflects that dependency burden may affect household entrepreneurial exit by reducing physical capital accumulation. In this study, the intrahousehold transfer of physical capital is captured by “whether children give financial help” and “household medical expenditure”, respectively, with the former set as 0 and 1 dummy variables and the latter as a continuous variable. The regressions are conducted in the form of logarithms, taking into account the magnitude of the dependent variable.

Table 4. Test results of physical capital mechanism.

Variables	Financial Help		Medical Expenses
	(1)	(2)	(3)	(4)
Dependency burden	0.443 *** (0.050)	0.480 *** (0.052)	0.879 ** (0.406)	1.680 ** (0.445)
Individual characteristics	Control	Control	Control	Control
Family characteristics		Control		Control
Province		Control		Control
Sample size	1201	1201	1201	1201
Wald chi2	138.00 ***	157.85 ***	16.89 ***	8.63 ***
R-squared	0.2272	0.2444	0.0775	0.1135

*** and ** indicate significance at 1% and 5% statistical levels.

demonstrates the results of the test for the physical capital mechanism. The results of (2) and (4) show that the degree of dependency burden has a positive effect on children providing financial help to parents at the 1% statistical level, and the degree of dependency burden has a positive effect on household medical expenditures at the 5% statistical level. The more severe the aging of the household population, the higher the intergenerational transfer of physical capital to the household due to aging. Dependency burden reduces physical capital accumulation and thus affects household entrepreneurial exit.

4.3.2. The Impact of Dependency Burden on Labor Supply

The time spent by households on caregiving reflects the loss of labor supply. This paper selects “whether household members help to take care of the elderly’s daily household chores or meals” and “whether they spend time on household chores or parental care” from the CFPS database to reflect the time spent on caregiving.

Table 5. Test results of labor supply mechanism.

Variables	Spending Time Caring for the Elderly		Housekeeping or Parental Care
	(1)	(2)	(3)	(4)
Dependency burden	0.411 *** (0.046)	0.488 *** (0.050)	0.153 ** (0.075)	0.160 ** (0.078)
Individual characteristics	Control	Control	Control	Control
Family characteristics		Control		Control
Province		Control		Control
Sample size	1201	1201	1201	1201
Wlad chi2	121.23 ***	150.06 ***	92.42 ***	85.97 ***
Pseudo R2	0.1742	0.1936	0.0691	0.0658

*** and ** indicate significance at 1% and 5% statistical levels.

shows the results of the labor supply mechanism. The results of (2) and (4) show that for every 1% increase in dependency burden in the entrepreneurial sample, the likelihood that the entrepreneurial household spends time caring for the elderly increases by 48.8%, and the likelihood that the entrepreneurial household spends time taking care of household chores or parents increases by 16%. Along with the increase in dependency burden, households need to spend time caring for their elderly members daily, and dependency burden reduces labor supply, affecting entrepreneurial exit.

4.3.3. Impact of Dependency Burden on Health Capital Accumulation

Dependency burden may affect home business exit by reducing the health of family caregivers. Health accommodates both physical and mental health, and to verify the mechanism, this study conducts an empirical analysis in terms of physical health and mental health levels, respectively. Regarding the level of physical fitness, this study proposes that we use the average self-rated health level of other members of the household workforce, with answers ranging from unhealthy to very healthy, assigned a score of 1–5, respectively. Regarding mental health, some studies have measured the level of mental health by using mental health scales such as the CES-D scale [32]. However, the subjects studied in this paper are entrepreneurial families. Using general scales to measure directly may not exclude the interference of other factors on mental health. The mental health level of family members should be related to entrepreneurial activities. The subjects studied in this paper are entrepreneurial families, measured directly using general scales that may not exclude other factors’ interference in mental health. The level of psychological health of family members should be related to entrepreneurial activities, and the satisfaction of family members with entrepreneurial activities reflects the psychological health status at work. Referring to existing studies [33], this paper uses the satisfaction of family workforce members with their jobs to measure mental health level. The answers were given on five levels, from very dissatisfied to very satisfied, with a 1–5 respectively.

Table 6. Test results of health capital mechanism.

Variables	Self-Assessment of Health Status		Job Satisfaction
	(1)	(2)	(3)	(4)
Dependency burden	−7.241 *** (0.286)	−7.195 *** (0.311)	−4.055 *** (0.103)	−4.002 *** (0.237)
Individual characteristics	Control	Control	Control	Control
Family characteristics		Control		Control
Province		Control		Control
Sample size	1201	1201	1201	1201
Wald chi2	829.64 ***	897.71 ***	406.42 ***	469.60 ***
R-squared	0.2052	0.2144	0.0953	0.1032

*** indicate significance at 1% statistical levels.

shows the results of the health capital mechanism test. The results of (2) and (4) show that the degree of dependency burden hurts family self-rated health at the 1% statistical level, and the degree of dependency burden harms job satisfaction at the 1% statistical level. As household dependency burden deepens, both self-rated health status and job satisfaction of household members decline, and dependency burden reduces healthy human capital, affecting household entrepreneurial exit.

4.4. Robustness Test

4.4.1. Heckman Model

The previous paper examines the impact of dependency burden on entrepreneurial exit behavior by directly restricting the study to households with entrepreneurial behavior in 2016. However, the observed entrepreneurial exit behavior presupposes that households are engaged in entrepreneurial activity, and the group engaged in entrepreneurial activity may have certain characteristics. Excluding households without entrepreneurial activity from the sample would ignore the problem of sample selection bias due to many unobservable factors. This paper deals with the sample selection bias problem from a full-sample perspective using the Heckman model, also known as a two-step approach. The first step is the selection equation, and the second step is the exit equation. One is the proportion of the community’s population over 60 years old to the total population, and the other is the number of community nursing homes or places for senior activities.

Table 7. Robustness test results: Heckman model.

Variables	Heckman Model
	Select Equation (to Enter or Not)	Exit Equation (Whether to Exit)
Dependency burden	-	0.184 ** (0.087)
Individual characteristics	Control	Control
Family characteristics	Control	Control
Province	Control	Control
Community Features	Control	-
Observations	11109	1201
Insigma	−0.742
Sigma	0.476
Prob > chi2	37.54 ***

*** and ** indicate significance at 1% and 5% statistical levels.

shows the results of the Heckman model treatment. The likelihood ratio test rejects the original hypothesis. This result indicates that the sample selection model chosen in this paper is reasonable and the sample has some selection bias problems. The unbiased estimation of the exit equation shows that dependency burden positively affects the exit of household entrepreneurship.

4.4.2. Replacement of Independent Variables

This paper uses the replacement of core explanatory variables for robustness testing. First, aging is directly reflected in the growth of the elderly population, so the number of people aged 60 and above in households is directly used to measure the degree of aging. Second, with the improvement of medical facilities and the social security system, the physical strength of older adults declines slowly. Based on the caregiving perspective, this paper raises the defined age of the elderly from 60 to 65 years old and observes whether the aging of the population under this threshold plays the same effect on entrepreneurial exit behavior.

Table 8. Robustness test results: replacement of independent variables.

Variables	(1)	(2)	(3)	(4)
Number of people over 60 years old	0.043 ** (0.020)	0.046 ** (0.020)	-	-
Percentage of population over 65 years old	-	-	0.154 ** (0.075)	0.145 * (0.083)
Individual characteristics	Control	Control	Control	Control
Family characteristics	-	Control	-	Control
Province	-	Control	-	Control
Observations	1201	1201	1201	1201
Wald chi2	37.41 ***	72.70 ***	36.27 ***	61.07 ***
Pseudo R2	0.0254	0.0474	0.0252	0.0410

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

shows the estimation results of replacing the independent variables. The independent variables in columns (1) and (2) are the number of households aged 60 and over, and in columns (3) and (4) they are the proportion of people aged 65 and over in the family. First, the increase in the number of persons aged 60 and above in the household significantly affects entrepreneurial exit behavior. Second, the proportion of older adults aged 65 and above in the household positively affects entrepreneurial exit behavior, reflecting that dependency burden under stricter bounds induces entrepreneurial exit behavior in households likewise holds.

5. Discussion

Integrating digital technology with the aging industry, digital promotion of multi-channel flexible employment, and digital social activities in the digital economy may be a long-term mechanism to promote active aging [34,35]. However, the level of access to active aging benefits for aging households in the new digital economy is closely related to digital technology use. Therefore, this paper explores the moderating role of digital technology on dependency burden from two aspects: digital technology accessibility and usage.

The network platform functions and services gradually present diversification, wisdom and refinement. The digital economy has given rise to many new business forms to provide opportunities for solving various aspects of dependency burden, such as the smart elderly industry [36]. This paper explores whether the negative effects of digital technology applications on dependency burden can be mitigated. Micro-subjects’ connectivity to the external digital environment through digital technologies may help households achieve a partial shift in the burden of old age. In terms of physical capital accumulation, digital technologies have enriched flexible forms of employment and expanded employment age boundaries, thus helping households increase their physical capital stock. The development of digital platforms has greatly contributed to the boom of the digital economy, which has created many flexible employment opportunities. Relying on social economy employment on internet platforms, etc., older people can flexibly participate in employment positions in the labor market and increase household income. In terms of labor supply, older adults flexibly participate in the entrepreneurial sales segment through live-streaming with goods and other forms [37]. In terms of caregiving time, smart health services on digital platforms, especially internet-based online medical services’ built-in hospitals, help save caregivers’ commuting time to accompany the elderly population to hospitals for medical treatment to a certain extent. The iteration and application of smart digital products such as smart homes and smart devices that can broadcast and reflect the situation of the elderly for the first time are helpful to save caregivers’ fine-grained care time [38]. In terms of health capital stock, digital technology has significantly improved family members’ psychological well-being, self-rated health, and self-approval through enriching recreational activities or expanding social engagement methods [39]. The reduced caregiving time also contributed to increased mental and physical health. Overall, digital technology use helped mitigate three negative effects of dependency burden in the entrepreneurial exit behavior of households.

Digital technology use relies on internet port access and the sharing of digital information through the platform, including the content of activities such as work, learning, social, recreational or business activities. Therefore, digital technology use is defined in two aspects—one is accessibility, and the other is the degree of use. Digital technology accessibility is reflected in the exposure to digital technology. Digital technology usage reflects the extent and ability of households to use digital technology for activities. The more frequently households use digital technology for learning, work, entertainment, and social activities, the stronger the digital technology usability. This paper classifies digital technology usage as low and high and assigns 0 and 1, respectively.

The results of the subgroup regressions are referred to in

Table 9. Results of the test for moderating effects of digital technology use.

	(1)	(2)	(3)	(4)
Variables	Use of Digital Technology	No Digital Technology	High Usage	Low Usage
Dependency burden	0.063 (0.128)	0.196 * (0.115)	0.165 (0.158)	0.170 ** (0.087)
Householder characteristics	Control	Control	Control	Control
Family characteristics	Control	Control	Control	Control
Province	Control	Control	Control	Control
Sample size	650	551	498	703
Wald chi2	28.37 ***	44.30 ***	20.49 ***	57.50 ***
Pseudo R2	0.0333	0.0678	0.0443	0.0657

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

. The results show that the coefficient of the effect of dependency burden on household entrepreneurial exit behavior is insignificant and significant among households with and without access to digital technology, respectively. They reflect that the positive effect of dependency burden on entrepreneurial exit is greater among households without digital technology than those with access to digital technology. The subsample regression of digital technology usage shows that the positive effect of dependency burden on entrepreneurial exit behavior of households shifts from significant to insignificant during the change from low to high usage. This result reflects that digital technology use does weaken the facilitating effect of dependency burden on entrepreneurial exit.

6. Conclusions

Many household-based subjects have joined the entrepreneurial trend driven by China’s entrepreneurship policy. Still, the higher entrepreneurial exit behavior hidden under the entrepreneurial trend phenomenon is often overlooked. This paper explores the relationship between dependency burden and entrepreneurial exit behavior of residents’ households from the perspective of dependency burden and household entrepreneurial exit phenomenon.

The findings of the study are as follows. First, the likelihood of resident households exiting entrepreneurial activities increases along with the increasing dependency burden. Second, the heterogeneity analysis of entrepreneurial exit motives results show that, compared to active exit, dependency burden mainly induces failure exit of resident households from entrepreneurship. Third, in addition to reducing households’ physical capital accumulation and labor supply, dependency burden also reduces caregivers’ mental and physical health, thus inducing failure exit from entrepreneurship. Fourth, the use of digital technology by households mitigates the negative effects of dependency burden in household entrepreneurship to some extent. This mitigating effect on entrepreneurial exit will be enhanced with increasing use.

The findings of this paper are important for further deepening the understanding of the relationship between family factors and entrepreneurial exit behavior. The theoretical contributions of the study are mainly as follows. First, unlike other scholars who take ordinary entrepreneurs, start-ups or industries, and the macro-environment as the objects of analysis, this paper, for the first time, aims the research perspective at family entrepreneurship. This paper is the first to focus on family entrepreneurship, a socially marginalized group. Based on the aging trend of the population, this paper constructs a model on the motivation of entrepreneurial exit from the perspective of the aging of the family, which not only explains the causes of entrepreneurial exit at the family level but also enriches the research results on the antecedents of entrepreneurial exit to a certain extent. Second, Wennberg et al. [9] and Rybczynski [12] find that child-rearing reduces the likelihood of exit for male entrepreneurs and that the higher the number of children raised reduces the probability of exit for female entrepreneurs. However, unlike their study, this paper finds that supporting the elderly causes household entrepreneurial exit, and intergenerational upward and downward shifts have very different effects on entrepreneurial exit. The possible reason is that child rearing is an investment behavior with self-motivating effects for entrepreneurs, but supporting the elderly is a depleting process of family resources. Finally, this paper also extends the transmission mechanism by which dependency burden affects entrepreneurial exit behavior; dependency burden not only reduces households’ physical capital accumulation and labor supply but also leads to lower health capital due to other household laborers playing the role of caregivers.

The relevance of this paper to reality is that our findings are significant for government efforts. The increasing aging of society’s population requires attention to the relationship between dependency burden and the entrepreneurial exit phenomenon if relatively high-quality entrepreneurship is to be maintained. The government needs to alleviate the burden that care and retirement place on families through a robust social security system such as pension support. The government can also intervene in the financial burden of an aging population by providing some policy support (e.g., financial incentives) to provide entrepreneurial opportunities to the middle of the workforce to make entrepreneurship sustainable. In addition, the government needs to promote the development of the “Internet + Aging” and smart aging industry, strengthen the digital literacy of entrepreneurial subjects, and achieve a balance between social aging and family aging.

This study has room for expansion in the future. First, the process of entrepreneurial exit is very complex. We do not discuss this issue based on process theory but rather on a feature-oriented approach to entrepreneurial exit as a decision-making behavior, making it difficult to discuss the household entrepreneurial exit process. In addition, although we used a sample correction model to address the selectivity bias, we did not find a suitable instrumental variable to identify the causal relationship more precisely. We can try to use new data sources to test the findings in the future further. Again, the matching problem between the aging population and entrepreneurial activities should also be a direction for deeper research because there is a labor supply shortage for entrepreneurship due to the increasing aging of the society’s population. In the future, there is a need to gain insight into how new technological tools, such as digital technology, can help the aging population achieve sustainable entrepreneurship. Finally, the applicability of the research findings to the aging population and the evaluation of the effectiveness of related support measures also need to be further explored in the future.