To Your Health: Health Insurance, Self-Employment Exit, and Older Entrepreneurs
ABSTRACT
Drawing on a sample of 5,238 individuals in the Health and Retirement Study (HRS), we examined the relationship between having one’s own health insurance and exit from self-employment to employment. Our results indicate that individuals who have health insurance are less likely to exit self-employment. When self-employed individuals have their own health insurance, males, relative to females, are more likely to exit self-employment. Additionally, when self-employed individuals have their own health insurance, those who do not have health problems that limit work are more likely to exit. We discuss the implications that these findings have on the literature and policymakers.
Keywords: self-employment, health insurance, gender, age, policy
INTRODUCTION
The decision to exit a venture is an important event in the entrepreneurial process. While a growing body of research has investigated the factors that might influence or relate to venture exit (DeTienne and Cardon, 2012; DeTienne et al., 2008; Wennberg et al., 2010), there is still much that needs to be understood with regard to what elements could impact venture exit. Both internal factors and external contexts can serve to influence whether or not one persists in, or exits from, self-employment (Gimeno et al., 1997). Although evidence has indicated that individual factors, such as personal values (Holland and Shepherd, 2013) and socioemotional wealth (Chirico et al., 2015; Parastuty et al., 2016), as well as expected financial and emotional costs (Shepherd et al., 2009) can all effect whether or not individuals exit self-employment, it is also important to consider the impact that external contexts and the life-cycle stage of an individual can have on the decision to persist in self-employment (Gimeno et al., 1997).
Following this logic, we ask whether having one’s own health insurance, an important determinant of cost of living in older age, influences older individuals to either continue with or exit from self-employment, which has become a particularly salient topic given recent policies and legislation related to health care regulation (Blumberg et al., 2014). Evidence has shown that health insurance has important effects on both labor force participation as well as job choice (Currie and Madrian, 1999), but the findings with regard to the specific relationship between health insurance and self-employment have been less conclusive. While some evidence has indicated that these policies could increase individuals’ probability of self-employment (Heim and Lurie, 2010) as well as the rate of entry into self-employment (Gumus and Regan, 2009), recent investigations have not found a significant relationship between health insurance and entry into self-employment (Gumus and Regan, 2015). In fact, some evidence even suggesting that having health insurance can actually increase the overall financial risk of the household by encouraging individuals to seek care more frequently and from more expensive, high-level providers (Wagstaff and Lindelow, 2008).
Related to our focus on elderly self-employed individuals, age is an important factor in the continuation of self-employment (Shane, 2000). With age, risk aversion increases (Dohmen et al., 2011; Ekelund et al., 2005; Pålsson, 1996), which could reveal one of the potential mechanisms via which age may moderate the relationship between having one’s own health insurance and exiting self-employment. The uncertainty of medical costs stemming from transitioning out of self-employment if individual health insurance has been obtained, coupled with the increased risk aversion among older individuals, could further lower the chances of exit from self-employment among older individuals with health insurance. While the influence of health care availability on “job lock” has been extensively studied within organizational contexts (Madrian, 1994), with few exceptions (Gumus and Regan, 2015), relatively less attention has been given to how this phenomenon relates to likelihood to transition from self-employment. Furthermore, while evidence has suggested that policies related to establishing continuance of coverage have increased job mobility, thereby reducing job lock (Gruber and Madrian, 1994), these policies are explicitly aimed at individuals who are organizationally employed, and as such are not applicable with regards to the potential exit from self-employment. Additionally, while recent evidence suggests that job lock is not a major concern for young adults (Bailey and Chorniy, 2016), less is known about what effect health insurance might have on older individuals with regards to their likelihood of exiting self-employment.
In addition to the association between health insurance and exit from self-employment, we also examine two additional moderating factors—in the form of gender and whether health conditions interfere with work—and how they could influence the relationship between health insurance and self-employment exit. Research has suggested that there could be demographic influences on venture exit (DeTienne and Cardon, 2012), with growing evidence indicating that gender can have a substantial influence on a number of factors related to self-employment (Leoni and Falk, 2010; Leung, 2006). As such, it is critical to investigate whether having one’s own health insurance and gender might relate to exit. Controlling for the availability of health insurance, it is possible that differences in certain factors, such as risk preferences (Byrnes et al., 1999; Charness and Gneezy, 2012; Sapienza et al., 2009), as well as the difficulties faced by women to enter self-employment in general (Leoni and Falk, 2010; Leung, 2006) could contribute to important variance in self-employment exit between males and females. Furthermore, having existing health conditions that interfere with work may limit individuals’ ability to transition out of self-employment if they obtain their own health insurance, as insurance firms are generally unlikely to cover individuals with health conditions that would interfere with future work. Even with the passage of the Affordable Care Act, while insurers may not deny coverage for pre-existing conditions, recent studies have found that insurance companies are delaying payments or only partially covering such conditions[1]. Thus, it is likely that those with conditions that are serious enough to interfere with work, and who have adequate health insurance, would be hesitant to exit self-employment so as to avoid the risk of being uninsured or under-insured by potential employers.
In completing our study, we make several important contributions. First, our study uncovers important associations between having one’s own health insurance and exiting from self-employment. We find a significant negative relationship between having one’s own health insurance and exiting self-employment. This result extends previous findings, which showed that “the sum of the estimated marginal effects (of having health insurance on self-employment exit) . . . is negative, but lacks statistical significance” (Gumus & Regan, 2015: 369). We believe that our findings add empirical support for previous interesting, but empirically non-significant, findings by Gumus and Regan (2015), thereby providing additional evidence to help better understand previously reported findings.
Second, our findings indirectly suggest that females and those with health problems limiting work who have their own health insurance are less likely to exit self-employment. As mentioned, gender has been shown to have a substantial influence on self-employment (Leoni and Falk, 2010). Evidence suggests that females face increased difficulties in becoming self-employed. From difficulties obtaining bank loans (Carter et al., 2007) or venture capital funding (Greene et al., 2001) to the persistent earnings gap between males and females who are self-employed (Leung, 2006), research has shown that women face unique challenges with regard to self-employment. However, while substantial research has examined the influence of gender on self-employment, there remains a relative paucity of research into the relationship between gender and remaining self-employed. Our findings suggest that compared to females, males who have their own health insurance are more likely to exit self-employment, providing an important insight into factors that could impact the relationship between gender and self-employment.
Additionally, while previous research has examined the relationship between both age (Zissimopoulos et al., 2009) and overall health (Lewin-Epstein and Yuchtman-Yaar, 1991) on self-employment and there has been growing interest in how external conditions may influence new venture persistence or exit (Gimeno et al., 1997; Holland and Shepherd, 2013), this study represents an initial attempt at combining these two streams of research. Our results indicate that, when having one’s own health insurance, relative to those who have health problems limiting work, those without health problems limiting work are more likely to exit from self-employment.
Finally, our results offer substantial insight regarding government and public policy. While recent studies have examined what, if any, effects government incentives might have on self-employment rates (Gumus and Regan, 2015), they did not investigate the role that gender, age, and health conditions interfering work might play in these relationships. Our findings show that gender and the presence of health conditions interfering work are indeed important factors to consider in terms of how policy-related influences, such as access to adequate health insurance, may influence exit from self-employment. This could be a potentially beneficial finding for policymakers and scholars alike when considering actions that could be taken to increase (or decrease) levels of self-employment through exit. In the pages that follow, we first describe the theoretical foundation of our model and present our hypotheses. We then follow up with a description of our research method, our results, and a discussion of our findings.
THEORY
Having One’s Own Health Insurance and Exiting Self-Employment
Self-employment exit, and the motivations behind this decision, is an important component in the entrepreneurial process (Wennberg et al., 2010). It is important to note that the decision to exit self-employment is complex in nature, and in fact can encompass multiple possible forms with regards to where individuals transition to after exit (e.g. full or part-time employment, unemployment, retirement, etc.). While we acknowledge the complexity of this decision, and detail the need for further research into the nuances and complexity of this decision in our section regarding future research, for the purposes of this study we focus specifically on individuals who exit self-employment for wage labor, either full or part-time in nature.
It has been established that some factors that influence self-employment exit are external in nature, with the most notable of those being related to financial underperformance that leads to self-employment exit out of necessity. However, while financial performance can influence exit, a substantial amount of previous research focuses on factors other than performance as key determinants in the exit decision (Lin et al., 2000). In fact, it has been postulated that new venture exit is less linked to firm performance and more tied to the perceptions of individual entrepreneurs (DeTienne, 2010). From this perspective, because individuals have varying degrees of acceptable performance thresholds, the decision to exit self-employment often hinges on factors other than financial values or returns (DeTienne et al., 2008). Indeed, research has shown that many times individuals decide to persist with failing businesses in direct opposition to the financial logic of reducing losses that would result from a more immediate termination of the venture (DeTienne et al., 2008; Holland and Shepherd, 2013). Furthermore, evidence regarding self-employment exit suggests that a substantial proportion of self-employment periods are not terminated through bankruptcy, but rather as a result of a move to alternative, more attractive, employment opportunities (Taylor, 1999).
Individual factors ranging from age and experience (DeTienne and Cardon, 2012) to socioemotional wealth (DeTienne and Chirico, 2013) have been shown to influence venture exit. Conversely, environmental munificence and previous organizational success (DeTienne et al., 2008) as well as personal values (Holland and Shepherd, 2013) and considerations of balancing the financial and emotional costs of exit (Shepherd et al., 2009) have all been pointed to as factors that might motivate individuals to persist with ventures even when those ventures are failing. Extending these streams of research, attention has recently turned to how policy-related factors in general, and the availability of health insurance in particular, may also be related to the likelihood that individuals will be self-employed.
The individual ability to obtain and maintain adequate levels of health insurance coverage has been shown to influence both labor force participation as well as the specific job choices that individuals consider for employment (Currie and Madrian, 1999). However, the results regarding the relationship between health care coverage and self-employment to this point have provided less conclusive findings. While some studies have suggested that increases in the deductibility of health insurance premiums for those who are self-employed could increase the probability of self-employment (Heim and Lurie, 2010) as well as the rate of entry into self-employment (Gumus and Regan, 2009), recent evidence has called these findings into question. In a recent study, Gumus and colleagues (2015) failed to find a significant relationship between health insurance and entry into self-employment, and some data suggests that having health insurance can actually increase overall financial risk by increasing the likelihood that individuals who have health insurance will incur higher levels of spending on healthcare as a result of seeking more frequent, higher-cost care (Wagstaff and Lindelow, 2008), resulting in negative consequences for those who are self-employed.
Unpaid medical bills are estimated to be the top reason for individual bankruptcy filings in the United States, surpassing even credit card bills and mortgage defaults (Mangan, 2013). While the majority of individuals rely on their employer to provide them with health insurance (Blumberg et al., 2014), this is not an option for those who are self-employed. Self-employed individuals must purchase health insurance themselves, and although it is tax deductible, individual and family coverage can be quite costly. In fact, a majority of individual health insurance purchases are from self-employed individuals (KFF, 2011). Indeed, recent research has indicated that health insurance influences entry into self-employment in the United States, and the inability to obtain adequate health insurance has been suggested to be one possible explanation for lower levels of self-employment (Blumberg et al., 2014). However, just as the inability to obtain health insurance could negatively influence self-employment entry, for those who are self-employed and have obtained adequate health insurance, it is possible that this could have the opposite relationship with regards to reducing the likelihood of exiting self-employment. Although self-employment accounted for only 6.6% of overall employment levels in 2014 (EMSI, 2014), the self-employed and those who work for very small businesses constitute more than one-third of uninsured working adults in the United States (Fronstin, 2013). As such, for self-employed individuals who are able to acquire and maintain adequate health insurance, this could provide a strong, non-financial incentive to motivate them to persist with their self-employment endeavors, even in the face of lower than expected financial performance (DeTienne et al., 2008; Holland and Shepherd, 2013). Based upon this reasoning, we propose the following:
Hypothesis 1: Older individuals having one’s own health insurance will be less likely to exit from self-employment.
Gender, Having One’s Own Health Insurance, and Exit
While it is important to consider the effects that health insurance has on exit from self-employment, it is possible that this relationship is moderated by a number of factors. A summary of the moderation effects that we present and test can be found in the table in Appendix A. With a growing amount of research focusing on the differences between men and women with regard to both self-employment (Clain, 2000) and health care (Bertakis et al., 2000), it is evident that gender could potentially moderate the relationship between having one’s own health insurance and exiting self-employment. Past research suggests that women are more risk averse than men (Booth and Nolen, 2012), particularly in financial situations (Dwyer et al., 2002; Sapienza et al., 2009). While this risk aversion might prevent women from becoming self-employed (Ekelund et al., 2005), for those who have already started a new venture, it could further prevent them from exiting self-employment, especially if they have been able to obtain individual health insurance. Previous findings indicate that females are less likely to receive employer-sponsored health insurance than males (Buchmueller 1996) but that they are also more likely to utilize health care services (Bertakis et al., 2000), thus increasing uncertainty in health care costs in transitioning out of self-employment if health insurance has been obtained. Additionally, social perceptions of masculinity and femininity further widen the gap between what is perceived as acceptable with regard to health care–service utilization for women versus men (Courtenay, 2000). When these factors are taken into account, it is possible that for women who are self-employed and have obtained adequate health insurance of their own, exiting self-employment could be perceived as a higher-risk proposition, as the opportunity to obtain adequate insurance if they were to transition out of self-employment could potentially be reduced as a result of gender influences. If this is the case, it is likely that they will be less inclined to exit self-employment when such contexts exist. Based upon this reasoning, we propose the following:
Hypothesis 2: Older females who have their own health insurance will be less likely to exit self-employment than older males who have their own health insurance.
Age, Insurance, and Exit
While health insurance is an important factor to consider in terms of individuals’ exit from self-employment, the significance of this relationship likely varies depending upon individuals’ age. Age has been shown to substantially influence a number of serious medical conditions and also affect risk aversion. From cancer (Doll and Peto, 1981) to coronary heart disease (Jousilahti et al., 1999) to the potential for stroke (van Walraven et al., 2009), evidence suggests that as individuals age, they are much more susceptible to experiencing serious health issues. Additionally, chronic conditions, such as macular degeneration (Friedman et al., 2004) and Alzheimer’s disease (Craft, 2005), which require substantial, long-term, and costly treatment, are much more prevalent in older individuals. These factors suggest that having one’s own health insurance becomes increasingly important as individuals age, and previous research has suggested that the costs associated with treating such serious conditions can be substantial (Yabroff et al., 2008). Therefore, as individuals age, the relationship between having one’s own health insurance and exiting self-employment will likely become even more pronounced, and older individuals will be less likely to exit from self-employment if they have their own health insurance, since it is likely that they will have less opportunities to obtain employer-sponsored health insurance based upon their age. Based upon this reasoning, we propose the following:
Hypothesis 3: Age moderates the relationship between older individuals and exit from self-employment such that those of relatively higher age groups are more likely to exit self-employment when they have their own health insurance.
Health Problems Limiting Work, Having One’s Own Health Insurance, and Exit
Just as age might influence the relationship between health insurance and exit from self-employment, it is likely that having health problems limiting work will effect this relationship. Having a health condition that limits work can increase the future cost of health insurance when self-employed individuals seek health insurance through their potential new employer (Collins et al., 2006). As a result, leaving self-employment when health conditions are severe enough to limit work, could limit the odds of getting future health insurance. Moreover, the cost of treating future health conditions, particularly if they are chronic in nature, can have a severe impact on individuals’ financial situations and have been shown to be a substantial cause of bankruptcy (Himmelstein et al., 2005). Additionally, evidence suggests that having work-limiting health conditions can exacerbate the potential negative effects of subsequent health conditions (Pittet et al., 1993), which further increases the costs associated with health care. As a result, having health problems that limit work can cause individuals to remain in an employment role that provides adequate insurance rather than risk leaving such a position. An analogical condition, termed “job-lock”, has been extensively studied within organizational contexts (Kapur, 1998). For individuals who have health problems limiting work and are self-employed but have their own health insurance, it is possible that they experience a similar job-lock situation such that they are hesitant to leave self-employment. This lack of job mobility can in part be influenced by the fact that they are reluctant to risk having to find health insurance from a different source, as the opportunity to obtain employer-sponsored health insurance could be limited based upon their prior medical conditions. Based upon this reasoning, we propose the following:
Hypothesis 4: Older individuals who have their own health insurance will be less likely to exit self-employment if they have a health problem limiting work than those who do not have health problem limiting work.
METHOD
Data
We draw on the Health and Retirement Study (HRS) to test our proposed hypotheses. The HRS is a panel (1992–2012) of 20,000 adults aged 50 and older in the United States who were born between 1924 and 1959. Supported by the National Institute on Aging (NIA) (U01AG009740) and the Social Security Administration, the HRS is geared toward assessing labor force and health transitions in the later stages of participants’ lives. Since the first wave in 1992, the study has collected a variety of data on income, assets, work, retirement planning, health, and cognitive functioning. The HRS data have been used in a wide range of studies in a diverse set of academic fields[2]. Through its in-depth interviews, the HRS provides an invaluable and growing body of multidisciplinary data that researchers can use to address important questions about the challenges and opportunities of aging.
The HRS is one of the most widely used studies, and the detailed discussion of this expansive study is beyond the scope of this work. We refer interested readers to additional information available at: http://hrsonline.isr.umich.edu/. The first wave of the HRS started in 1992 with a random sample of households with at least one member born between 1931 and 1941, or between the ages of 51 and 61 years old. Later on, additional cohorts for “War Baby” (WB) (born 1942 to 1947), “Early Baby Boomer” (1948 to 1953), and “Mid-Boomer (MBB)” (born between 1954 to 1959) were added. Each cohort was surveyed biennially, and cohorts born before 1931 were also added later.
More details regarding the specifics of this survey are available on the HRS site as well as the RAND HRS Data Documentation site.[3] We use the RAND HRS data file provided by RAND, a clean and easy-to-use version of 1992, 1993, 1994, 1995, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010 final, and 2012 early release waves. Based on pairwise deletion for all the variables in the analysis, the final sample includes 5,238 individuals representing 18,628 individual-year observations.
——————————————
Insert Tables 1 and 2 about here
——————————————-
Measures
Table 1 presents the list of variables, how they are operationalized, and descriptive statistics. To code the exit variable, if an individual was self-employed in the previous wave but not in the current wave, and was employed full-time or part time in the current wave and if the current wave health insurance status was not missing in the current wave we code it as 1 (otherwise coded 0). By this definition, those who retired, were disabled, or were not in the labor force are coded as missing to indicate a distinct type of exit from self-employment. Note, that as Medicare is available for people age 65 or older, and as this qualification could drive the choice of exit, the inclusion of age as a predictor controls for these effects.
Unobserved heterogeneity[4]. A majority of studies in entrepreneurship have focused on economic drivers (e.g. firm not performing well or bankruptcy) as possible explanations for exit. In the main analysis although we control for household income and assets, however, it is possible that the magnitude of change in household income or assets from the previous wave could also affect our inferences[5]. This change in assets and income, in turn, could also interact with having one’s own health insurance and influence the likelihood of self-employment exit. To control for this type of unobserved heterogeneity, we control for the change in household income (natural log of household income (‘000) in the current wave minus the natural log of household income (‘000) in the past wave, and this difference is lagged in all analyses. This lagged difference allows for us to control for potential changes in incomes and assets between waves. To further control for unobserved heterogeneity, we include ‘Own Health insurance × Change in income since past wave’ (lagged) and ‘Own Health insurance × Change in assets since past wave’ (lagged). These controls allow for assessing the joint consideration that changes in income and household assets and having one’s own health insurance could have regarding the decision to exit self-employment.
It is also possible that exit decisions could in part be based on future health insurance prospects. In other words, spousal access to health insurance or the ability to obtain future health insurance from the government may confound the inferences we draw, as the decision to exit may not be based on the current health insurance situation but also on expectations of garnering health insurance in the future through spouse, personal funds, or the ability to get a better job that provides health insurance, among other factors. To include unobserved effects from future prospects for receiving health coverage, we control for whether the participant is covered by their own health insurance, through a spouse, through government programs, or have any other type of health insurance.
Finally, related to the moderator for Hypothesis 4, we use the ‘health problem limiting work’ variable and control for existing health conditions. Health conditions that could be driven by a variety of biological, physiological, and genetic factors and could be managed by medication and may not interfere with work. To limit such unobserved heterogeneity, we control for health conditions and use ‘health problem limiting work’ as a predictor.
Predictors. Our main predictor is whether the individual had their own health insurance (lagged and controlling for insurance status [own, through spouse, government, or other]). The moderators are age in years (lagged), gender, and whether health problem limits work (lagged). We also include several alternate control variables to limit the potential effects of alternate explanations.
Controls. As household assets (Fairlie and Krashinsky, 2012) and household income (Hurst and Lusardi, 2004) could affect the propensity to exit from self-employment, we include them as controls (lagged). We also include education (Iyigun and Owen, 1998), race (Fairlie and Robb, 2008), and health condition (lagged).
The personality of an entrepreneur is related to both entry into (Zhao and Seibert, 2006) and exit from (Caliendo et al., 2014) self-employment, and so we control for individual personality aspects. Furthermore, both positive and negative affect are related to self-employment (Baron, 2008), and therefore are included as controls as well. Finally, life satisfaction is related to continuance in self-employment (Binder and Coad, 2013; Blanchflower and Oswald, 1998), and is included as a control in all analyses. Table 2 lists the correlations of all variables used in our analyses.
———————————————————–
Insert Tables 1-3 and Figure 1 about here
———————————————————–
Results
We use generalized estimating equations xtgee regression (Stata 15, with options family binomial and link logit) to test the proposed hypotheses. The fixed effects specification is not feasible for these analyses, as several of the personality and affect variables are measured only once. Hypothesis 1 proposed that individuals with own their health insurance were less likely to exit from self-employment. We find a negative significant relationship between having health insurance and exit from self-employment (Table 3, Model 1: β = -1.450, p < 0.001); therefore, Hypothesis 1 is supported. As both the predictor and the outcome variables are dichotomous, the interpretation is that those with their own health insurance are 1.45 times less likely to exit than those who do not.
Hypothesis 2 proposed that females with health insurance are less likely to exit from self-employment, and our results indicate that there is indeed a significant interactive effect between gender and health insurance on exit from self-employment (Table 3, Model 2: β = 0.658, p < 0.10). Figure 1(a) shows that relative to females, males with own health insurance are more likely to exit. Hypothesis 3 proposed that those of relatively higher age groups are more likely to exit self-employment when they have their own health insurance. Our results do not support this hypothesis (Table 3, Model 3: β = -0.00528, p > 0.10). Finally, Hypothesis 4 proposed that having a health problem limiting work would reduce the likelihood of exit from self-employment for individuals who have their own health insurance, and our results support this hypothesis (Table 3, Model 4: β = 1.308, p < 0.05). As presented in Figure 1(b), individuals with health problems not limiting work are more likely to exit self-employment than those with health problems limiting work; therefore, Hypothesis 4 is supported.
DISCUSSION
The importance of adequate health care coverage has become an issue of ever-increasing importance, particularly in light of the recent Affordable Care Act. By creating specific incentives and benefits for those who purchase individual health insurance, these policies also aim to drive self-employment (Gruber and Poterba, 1994). However, while health insurance’s influence on entering self-employment has been a topic of increasing scholarly interest (Gumus and Regan, 2009; Gumus and Regan, 2013), the potential effects that health insurance may have on exiting self-employment remain relatively less well studied. In this study, in order to help bridge this gap, we investigated a framework for the effects of having one’s own health insurance on the likelihood of exiting self-employment.
The results demonstrate that individuals who have their own health insurance are less likely to exit self-employment. Among those having their own health insurance, males, relative to females, are more likely to exit self-employment. Additionally, our results suggest that for older individuals, having one’s own health insurance does not have a significant relationship with self-employment exit. When having one’s own health insurance, individuals for whom health does not limit work are more likely to exit self-employment relative to those for whom health does limit work. The results show that our understanding of the relationship between having one’s own health insurance and exiting self-employment is far from complete, and that there are important boundary conditions to this relationship that need to be further investigated to help better understand what distinctions there may be.
Implications for Self-Employment Research
Exit from self-employment is an important issue in the entrepreneurial process (DeTienne, 2010), and while this topic is gaining growing scholarly interest (Wennberg et al., 2010), controlling for economic endowments, we aimed to better understand health insurance related factors among older individuals that might influence such individuals to either persist with or ultimately exit from self-employment. While research has uncovered both internal (Holland and Shepherd, 2013) and external (DeTienne et al., 2008) factors that might influence the decision to persist with self-employment, relatively little attention has been paid to how health insurance might influence self-employment exit. We believe that our findings help shed additional light on this issue and provide an important contribution to the discussion of the antecedents to self-employment exit.
Our results suggest that older individuals with their own health insurance who are self-employed actually have a lower likelihood of exiting self-employment than those who do not have their own health insurance. This result is consistent with recent evidence suggesting that having health insurance actually reduces self-employment exit (Gumus and Regan, 2015). We suggest that our results complement prior research and indicate that the relationship between having one’s own health insurance and exiting self-employment is consistent between the two samples. Gumus and Regan (2015) draw on the cross-sections of Current Population Survey that includes individuals from all age groups, whereas our work draws on individuals over 50. It is possible that for individuals in the HRS, the salience and relevance of health insurance could provide a stronger influence than for younger populations. As such, whereas research that included a wider range of age groups found a negative, but non-significant, association between health insurance and self-employment exit, our findings indicate a similar, but significant, relationship amongst elderly individuals. This suggests that while health insurance might have similar effects for all who are self-employed, the magnitude of this effect might not be homogeneous across all demographics, and provides an important insight into our understanding of the nuances of these relationships.
Our results provide further evidence for the potential boundary conditions that might exist with regard to the relationship between having one’s own health insurance and exiting self-employment in that we found that for females (males) and those with health conditions that (did not) interfered with work, having one’s own health insurance actually reduces (increases) the likelihood of exit. Building on previous findings that have shown that gender can influence nascent self-employment efforts (Greene et al., 2001), individuals’ engagement in self-employment (Leoni and Falk, 2010), and even the relative earnings realized for those who are self-employed (Leung, 2006), our results indicate that gender can also effect other stages of the self-employment process, specifically the likelihood of exit. Since females are generally more risk averse (Ekelund et al., 2005; Watson and McNaughton, 2007) and likely have greater difficulties in obtaining more traditional employer-sponsored health care coverage (Buchmueller, 1996), it is possible that if they are self-employed and have been able to obtain individual health insurance, they are more reluctant to risk the possibility of losing those benefits and will thus be less likely to decide to exit self-employment. This reluctance for women to exit self-employment if they have been able to obtain health insurance suggests that there could be a discrepancy with regards to the opportunities that are available in terms of obtaining employee-sponsored health insurance based upon gender influences. Essentially, our results suggest that it is possible that because women have fewer, or less attractive, opportunities with regards to the availability and access to employee-sponsored health insurance, then they are less likely to exit from self-employment if and when they are able to obtain their own health insurance. These results highlight the need for additional research regarding how gender could play an important role in several key processes with regards to self-employment.
Interestingly, with regard to the influence that age has on the relationship between having one’s own health insurance and exiting self-employment, our results do support the proposition that older individuals are actually less likely to exit self-employment if they do not have their own health insurance. While this finding is not supported, future studies could provide important insight into how age, health insurance, and self-employment are related. Age has been shown to increase the likelihood of a number of serious medical conditions (van Walraven et al., 2009), and the costs associated with these diseases are quite substantial (Yabroff et al., 2008). These factors not only increase the importance of having health insurance of one’s own, but they also make obtaining and maintaining such coverage more difficult, particularly outside more traditional employer-sponsored programs. Therefore, while movement into self-employment can be an attractive option for older individuals who are in good health (Zissimopoulos and Karoly, 2007), remaining self-employed might be more feasible for those who do not have the burden of trying to acquire and maintain adequate health insurance in older age, particularly in light of government sponsored programs available to older individuals such as Medicare. Essentially, because there are no alternative options available for older individuals to obtain similar, or superior, health care coverage outside of what they are afforded while self-employed, it is possible that they could choose to continue with their self-employment despite the lack of formal, work-related health insurance. Without the burdens that come from obtaining and maintaining self-insurance, it is possible that older individuals who are self-employed could better allocate key resources (e.g., financial, human, and social capital) to help enhance the overall performance of their ventures, thereby increasing the likelihood that they will meet or exceed the acceptable performance thresholds they deem necessary to persist in their entrepreneurial endeavors. Lack of support for the influence of age on our proposed relationship calls for additional research.
Finally, with regards to the influence that having health conditions which interfere with work might have on the relationship between having one’s own health insurance and exiting self-employment, it is likely that the presence of such work-interfering health conditions would serve to strengthen the relationship between having one’s own health insurance and exiting self-employment. Health conditions severe enough to limit work often, after exiting from self-employment with own health insurance, limit future access to many health insurance providers, and increase future premiums (van de Ven et al., 2000). As a result, the options that are available for individuals who have such pre-existing conditions with regards to obtaining employer-sponsored health insurance could be severely limited, or even non-existent. Therefore, for individuals who have such conditions, having their own health insurance could reduce the likelihood of self-employment exit for fear of not being able to obtain adequate and affordable coverage from alternative sources.
Implications for Policy and Practice
While the results of our study have important theoretical implications, it is crucial that research also be relevant to practice (Frank and Landström, 2016). To that end, there are several important implications that our results have with regards to both policy as well as practice. A large amount of research has focused on the influence of government policies on self-employment and entrepreneurial activities (Parker, 2004). While recent efforts have uncovered valuable insights into how policy might influence the relationship between how health insurance might influence individuals to become self-employed (Gumus and Regan, 2013; Gumus and Regan, 2015), relatively less is known about how policy might influence older individuals to remain self-employed, specifically via effects on health insurance. Some have suggested that policies like the Tax Reform Act of 1986, which included increased deductibility for insurance premiums for those who are self-employed, can actually result in an overall increase in the rate (Gumus and Regan, 2009) and probability (Heim and Lurie, 2010) of self-employment. However, recent evidence has suggested that policies like these have actually had little effect on the ability for self-employed individuals to obtain health insurance (Gumus and Regan, 2013), which brings to light the complex relationship between policy and individual behaviors.
Our results suggest that having one’s own health insurance could actually decrease the likelihood of self-employment exit. From a practical perspective, this suggests that while initiatives such as the Affordable Care Act might have been successful at providing the impetus needed to increase the number of individuals who have health care coverage, there could be unforeseen ramifications of forcing individuals to obtain health insurance, one of which could be a decreased likelihood of exit from self-employment. If indeed entrepreneurship and self-employment are the engines that drive economic growth and development (Acs, 2006; Acs and Szerb, 2007), then our results suggest that additional attention must be paid to how policies regarding health care and insurance can have consequences not only on the likelihood of entry into self-employment, but on the likelihood of exit from self-employment as well.
Moreover, the results of our study indicate that there are important boundary conditions that could influence the relationship between health insurance and self-employment exit. If indeed having one’s own health insurance provides a positive influence to remain self-employed for both women as well as those with health problems limiting work, then policies aimed at further increasing the ease and availability of obtaining health coverage for those groups could promote future entrepreneurial endeavors for individuals within those groups. However, if the motivation for members of these groups to remain self-employed once they have obtained health insurance is rather a function of a reduced availability of employer-sponsored options if they were to exit self-employment, then this points to a larger, more systemic issue that needs to be addressed. A lack of willingness for organizations to offer affordable and acceptable health insurance options to certain groups (e.g. women or those with pre-existing conditions), could represent a subtle form of institutionalized discrimination that would need to be addressed both from an organizational as well as legislative perspective. Additional research will be needed in order to determine the extent to which these relationships hold consistent over diverse organizational, industry, and national environmental settings.
Limitations and Future Research
As with all studies, the limitations of this research must be noted in interpreting the reported results. First, conclusions regarding the causal direction of the proposed relationships must not be inferred and are not implied, and additional investigations with longitudinal data will be necessary to empirically confirm the direction of these relationships. Second, HRS as a biennial study, measures exit between waves, so the precise timing of exit is not known. This “censoring” between waves could bias the estimates. Third, the participants’ motives for exit are not known, nor is information available on the nature of their health insurance. In other words, information related to business performance, family conditions, or other unobserved factors that may drive exit is not known. While we cannot assign a causal relationship between having one’s own health insurance and exiting self-employment, we would like to highlight that our findings are robust to controlling for unobserved heterogeneity and other controls. These controls partially assuage the influence that business performance (on household assets and income) or other personality factors and life satisfaction might have on motivating individuals to exit self-employment.
Future research could focus on different areas specific to entrepreneurship and policymaking. If the availability of self-obtained health insurance lowers the odds of exit, should policymakers increase tax deductions to make health insurance costlier for the self-employed? Accounting for the social cost of entrepreneurship (i.e., higher failure rates consume more resources that could be used for other purposes), policymakers should consider whether lower health insurance deductions could increase exit. Seeing that exit is considered “bad” from a policy perspective, managing levers of accessibility to health insurance among the employed and self-employed would be worth considering. Given the results of the moderation effects of gender and health conditions, tax deductions based on these characteristics could also help promote and maintain self-employment among women and older individuals.
Recently, Cumming et al. (2016) studied exit choices of high-tech entrepreneurs, and found that such entrepreneurs are likely to be serial entrepreneurs or angel investors. While our sample focuses on individuals over 50 who may be considering retirement in later waves, future studies could consider the exit choices of elderly entrepreneurs. With a greater likelihood of more endowed wealth among older high-tech entrepreneurs, it is possible that older high-tech entrepreneurs would increasingly substitute angel investing over serial entrepreneurship. Furthermore, it is possible that elderly, self-employed individuals may transit from self-employment to retirement, semi-retirement or some other states. To maintain theoretical tractability, we focused only on the transition from self-employment to employment. Focusing on a variety of alternate outcomes would require a variety to additional factors – social, cognitive, psychological – that would drive such complex choices. We call on future research to focus on these factors to develop a more nuanced understanding of such exit choices.
For entrepreneurship research, while the undertone of inferences from exit have ranged from negative (i.e., reduced prevalence of self-employment at the economic level) to neutral (i.e., exiting self-employment due to personal priorities) to positive (i.e., exit enables newer firms to enter), policy factors seem to play an important role in driving exit. Future research could merge these personal and policy aspects to further understand how individuals respond differently to policies. Given the recent implementation of the Affordable Care Act, it would be interesting to assess if self-employed individuals continued with self-employment to a greater extent before or after the passage of the law. Such research could have both positive and negative implications—continuation of self-employment could lower “churn” rates and increase efficiency, or it may maintain the continuity of entrepreneurial activity. It may also be interesting to assess whether, after the passage of the law, more individuals chose self-employment and whether more “freelancing” individuals (e.g., artists) chose full-time self-employment.
Conclusion
Our findings shed light on the important question of what factors influence exit from self-employment for individuals over 50. The results presented imply that having one’s own health insurance lowers the odds of exit. While our results indicate that in general, individuals who have their own health insurance are less likely to exit self-employment, this relationship has important boundary conditions. Notably, when having their own health insurance, males (relative to females) or individuals with health problems that do not limit work (relative to individuals with health problems that do limit work) are more likely to exit. Taken as a whole, the results of this study provide valuable insight into the influence that having one’s own health insurance can have on the decision to exit self-employment as well as the potential nuances and boundary conditions of this relationship.
[1] Source: http://www.webmd.com/health-insurance/20150128/have-insurers-found-way-around-obamacare-pre-existing-conditions-rule
[2] A complete list of studies from HRS is available at http://hrsonline.isr.umich.edu/index.php?p=biblio
[3] http://hrsonline.isr.umich.edu/index.php?p=shoavail&iyear=X7
[4] We thank an anonymous reviewer for suggestions for this test.
[5] We thank an anonymous reviewer for this comment
