Health and financial strain: evidence from the Survey of Consumer Finances.

• Author: Lyons, Angela C.

1. Introduction

   A large body of literature has examined the relationship between health and various socioeconomic factors such as education, income, and wealth. (1) These studies find a strong positive correlation between health and socioeconomic status (SES). Yet, there is little consensus on the direction of causality, and strong evidence exists that poor health can be both a cause and consequence of SES just as SES can be a cause and consequence of poor health.

   On the one hand, some studies find that poor health affects socioeconomic status such that individuals who are in poor health work fewer hours or are unemployed, limiting their ability to accumulate income and wealth (e.g., Ettner 1996, Smith and Kington 1997a, b; Zagorsky 1999; Wu 2003). These studies typically find that serious health conditions have a larger affect on SES than less serious conditions. Other studies show that lower SES results in poor health (e.g., Caplovitz 1974;

   Parker 1990; Havlik, Vukasin, and Ariyan 1992; Weyerer and Wiedenmann 1995; Smith 1998, 1999; Roberts et al. 1999; Drentea and Lavrakas 2000). (2) They argue that an individual's health can be affected by SES in primarily two ways--either by the physical stress that it creates or because of limited access to quality health care services. These studies find that psychosocial stress brought about by financial strain is associated with physical and mental illness including diseases, depression, and even suicide. These studies also show that lower SES can lead to poor health by affecting individuals' access to preventative health care and their ability to make health-maximizing decisions (see Jacoby 2002 and Meer, Miller, and Rosen 2003 for additional discussion).

   The previous studies provide significant insight into the relationship between health and SES. However, few adequately control for the endogeneity between health and SES. Smith and Kington (1997a, b) and Adams et al. (2003) have been among the first to investigate causal pathways between health and SES; however, none of these studies actually control for dual endogeneity. Regardless, Smith and Kington (1997a, b) find evidence using the Health and Retirement Survey (HRS) and the Asset and Health Dynamics among the Oldest Old (AHEAD) that the direction of causation is primarily from health to SES rather than vice versa. Adams et al. (2003) use panel data from the AHEAD to distinguish between acute, chronic, and mental health conditions and to control for existing health conditions. They find no evidence that wealth increases the incidence of mortality or acute, sudden onset health conditions. However, there is evidence that wealth increases the incidence of some mental, chronic, and degenerative conditions. In general, they reject the hypothesis that socioeconomic status results in health problems. (3)

   A recent study by Meer, Miller, and Rosen (2003) is one of the first to control for the endogeneity of SES using an instrumental variable for changes in wealth (receipt of an inheritance). Meer, Miller, and Rosen (2003) use the Panel Study of Income Dynamics (PSID) to examine changes in wealth and health. They find that, without taking endogeneity into account, changes in wealth result in a positive and significant effect on changes in health. Yet the effect is very small. When the endogeneity of wealth is taken into account, the effect of wealth on health becomes insignificant and essentially disappears.

   While the results are mixed on the direction of causality, there is general consensus that the relationship between health and financial strain goes deeper than the mere presence or absence of income, wealth, and debt. The literature argues that what matters more is the magnitude of debt a household has in comparison with their income and wealth (e.g., Drentea and Lavrakas 2000; Jacoby 2002). Yet most studies continue to focus on traditional socioeconomic factors such as income and wealth. Most studies are also limited in that the samples used are not representative of the U.S. population as a whole. For example, Drentea and Lavrakas (2000) use a sample of Ohio residents, and Roberts et al. (1999) use a small sample of British college students. Smith and Kington (1997a, b) and Adams et al. (2003) use data from health surveys on the elderly. Thus, with almost all of these studies, sample selection is a concern.

   This study moves beyond measuring the effects of traditional socioeconomic factors and takes a more comprehensive approach by examining the relationship between the overall level of household financial strain and the self-reported health status of the head. At the same time, this study also controls for the possible endogeneity between health and financial strain using instrumental variable techniques and a representative sample of the U.S. population. By examining a household's overall level of financial strain (i.e., asset-to-debt ratio, debt-to-income ratio), a more accurate accounting of the relationship between socioeconomic status and health can be established. Measures such as the household's asset-to-debt ratio and debt-to-income ratio take into consideration the magnitude to which the household is having financial difficulties.

   Using cross-sectional data from the Survey of Consumer Finances (SCF), two-stage probit models for the health status of the household head and for three separate measures of financial strain are estimated, controlling for the fact that financial strain can be both a cause and a consequence of poor health. Households in the SCF are identified as being under financial strain if they have one or more of the following three characteristics: delinquent on any type of loan payment by two months and more, total assets/total debts

   The financial guidelines used in this study were constructed based on previous research related to liquidity and household finances (e.g., Zeldes 1989; DeVaney 1994; DeVaney and Lytton 1995; Black and Morgan 1999; Stavins 2000; Lyons 2002). (5) Each measure captures a slightly different aspect of financial strain. The first measure looks at whether households are financially strained to the point that they are having difficulty paying their bills. The second measure identifies households that are highly leveraged since their liabilities are more than their wealth. The final measure looks specifically at households who have very little cash on hand to meet unexpected expenses such as those related to health. Households that are classified as being financially strained may have one or more of these characteristics. The empirical research shows that households with one or more of these characteristics are significantly more likely to be insolvent and thus under a substantial amount of financial strain.

   Using these three measures, the results show that poor health significantly increases the likelihood of financial strain. There is little evidence that financial strain contributes to poor health. The results are robust for all three measures. The findings suggest that health status may play a more important role in explaining why some households are under financial strain than vice versa. The findings have important policy implications, especially for low-income individuals who are in poor health. Poor health is likely to affect an individual's capacity to earn income or accumulate assets due to a limited ability to work or rising medical expenditures or both. The end result may be even greater financial strain or more severe health problems or both. Thus, the widening disparity in health is likely contributing to household financial problems and, in turn, to the growing gap in income and wealth inequality.

   The remainder of this article is structured as follows. The next section discusses the methodology and constructs the empirical framework. The third section describes the data, and the fourth section presents the empirical findings. The final section summarizes the results and their implications for policy and future research.

2. The Model

   The interaction between socioeconomic status and health can be illustrated using the standard economic model of health (Grossman 1972). This model consists of an intertemporal utility function, a health production function, and a budget constraint. The intertemporal utility function represents preferences for an individual's level of health and consumption of another good at each time period. An individual's level of health at any point in time depends on (i) medical care and other factors that influence an individual's health such as diet, smoking, physical activities, and psychological stress; (ii) personal attributes such as age, family background, genetics, and education; and (iii) health in the previous period. In this model, financial strain may lead to poor health for at least two reasons. As previously mentioned, it may increase an individual's psychological stress and/or result in fewer economic resources being devoted to health.

   Similarly, health status may also affect the financial position of the household. As discussed in Smith (1999), individuals with poor health often work fewer hours or are perhaps unemployed. A reduction in labor supply decreases an individual's future earnings capacity and limits the household's ability to accumulate assets and, in turn, savings. While poor health reduces current income and savings, it also increases out-of-pocket medical expenses. In such cases, savings may be depleted fairly quickly, placing financial strain on the household, which can threaten its overall financial well-being.

   In the standard model, an individual's level of health depends on both current and past health behaviors, personal attributes, income and wealth holdings, and prices. A past negative shock on the financial well-being of a family (e.g., economic recession) can be harmful to an individual's current health status, especially if it results in a decrease in the household's economic...