Health status and portfolio choice: Is their relationship economically relevant?

https://doi.org/10.1016/j.irfa.2014.01.008Get rights and content

Highlights

  • This paper questions the empirical relevance of the link between health and portfolio choice.

  • Portfolio choice is measured as stockownership and overall fraction of risky securities held.

  • We handle with caution the findings from previous papers.

  • We use data from the first wave of the Survey of Health, Aging and Retirement in Europe (SHARE).

  • We find that only poor self-reported health negatively impacts the portfolio choice.

  • Other health measures are irrelevant for investment decisions.

Abstract

Recent empirical work on individual portfolio choice focuses on the role of the individual's health in making financial decisions. The key idea is that, through precautionary saving or reducing investors' time horizon, health issues make people choose safer financial portfolios. This paper questions the empirical relevance of the link between health and portfolio choice, measured as stockownership and overall fraction of risky securities held. We handle with caution the findings from previous papers and ask whether data from the first wave of the Survey of Health, Aging and Retirement in Europe (SHARE) are able to clarify some of our doubts. We find that only poor self-reported health negatively impacts the portfolio choice, while other health measures (chronic conditions, limitations in daily activities of life, mental health) are irrelevant for investment decisions.

Introduction

The Modern Portfolio Theory (MPT) postulates that risk averse investors choose their investment portfolios in order to maximize their expected return for a predetermined level of risk. The optimal portfolio chosen by the investor will depend on the shape of her utility function. In recent years, the basic assumptions of MPT have been widely challenged by the behavioral finance approach, which has thrown new insight onto investment decision theory. Behavioral finance evaluates risk mostly based on laboratory experiments and surveys or questionnaire instruments, and concentrates on beliefs, attitudes and risk perception in particular circumstances. Among other factors, an individual's health status has recently gained attention as a potential determinant of risk perception and security holdings. Recent literature has developed a portfolio choice theory which includes the presence of “background” risk, defined as an uninsurable component of individuals' income risk which decreases additional financial risks (Gollier, 2001, Guiso et al., 1996, Guiso and Paiella, 2001). Some researchers have put effort into demonstrating how health can be regarded as a form of background risk, and provide empirical results on how financial investment choices change with health conditions. The most frequent finding is that, a sudden health issue or poor health conditions tend to shift resources towards safer types of financial investments, with disaffection from the participation to the equity market and from the holding of risky securities. Papers support with various arguments the way in which health affects portfolio choice. Some sustain that there is a precautionary saving purpose behind the health effect, to the extent that the expected future medical expenditures subtract resources from financial investments (among others, see Atella et al., 2012, Berkowitz and Qiu, 2006, Goldman and Maestas, 2007, Gollier and Pratt, 1996, Heaton and Lucas, 2000, Pang and Warshawsky, 2010, Pratt and Zeckhauser, 1987). Others consider the interplay between health and other circumstances which affect the life span horizon, such as aging (Coile & Milligan, 2009) or bequest motives (Feinstein & Lin, 2006). Finally, some papers propose models where health enters as a direct argument of the investor utility function, and the marginal utility of consumption is found to vary with health (among others, see Cardak and Wilkins, 2009, Edwards, 2008, Finkelstein et al., 2008, Love and Smith, 2010, Rosen and Wu, 2004). Evidently, the relationship occurring between health status and portfolio choice is actively discussed in the literature, and the standpoint of researchers in the field is extremely heterogeneous.

With this work we wish to take part in this debate, and our focus is on deeply understanding whether there is a relevant link between health status and portfolio choice. More specifically, we contribute to the previous literature surveying the individual health condition in a more extensive way. Since the health dimension is characterized by various different facets, our analysis aims at establishing the relationship occurring with portfolio decisions from different proxies of health, where each proxy should reflect a specific health dimension (subjective, objective, and mental health). We use data contained in the first wave of the Survey of Health, Ageing and Retirement in Europe (SHARE), which paints a picture of the lives of Europeans aged 50 and over from 11 countries. We concentrate on the decision of households on their stockholding, and on the relative share of their portfolio which is invested into risky assets. Our empirical models relate these two choices on the household health condition, while controlling for a large number of individual and household characteristics, as the household composition and various socio-economic features. We mainly test the most common channels that the literature so far has suggested as drivers for the impact from health on investments, and that we have mentioned few rows above. In particular, we check whether the implication from health on portfolio choice is imputable to precautionary saving and/or risk aversion changes. Despite we cannot rule out if one channel is more effective than the other, our results suggest that, a significant path from health on portfolio choice is only associable to a measure of subjective health. More specifically, people self-reporting a negative health condition are found to be less likely in owning some stocks and are holding fewer risky assets. If we instead measure the physical health status of the same respondents (counting the number of limitations in daily life activities or the number of chronic diseases), or measure their psychological feelings (through a depression scale), than it does not hold any more that being worse in health has significant affections on investments. Our results show that, the prediction of a negative health risk on financial choices highly depends on the notion of health we refer to, since we showed that different health measures might yield inconsistent results. The rest of the paper is organized as follows: Section 2 reviews the most relevant contributions to the literature which has surveyed the effect of health status on portfolio choice; 3 Data, 4 Measuring health status present the data and describe in more detail the variables of interest; Section 5 reports the econometric results; Section 6 provides the conclusions and some discussion of the results, comparing them with previous contributions in the literature.

Section snippets

The effect of health status on portfolio choice

There is abundant evidence of an association between economic measures, such as income and wealth, and a variety of variables which are linked to the state of health of individuals. Indeed, several studies confirm the existence of a link between the two spheres. However, there is still no consensus on the nature and the direction of this relationship. On the one hand, changes in socio-economic status may lead to worse health status. On the other hand, also changes of the health status may have

Data

In this study, we use data contained in the first wave of SHARE, the Survey of Health, Ageing and Retirement in Europe; its main focus is to provide information on the lives of Europeans aged 50 and over, collecting data in 11 countries from Scandinavia to the Mediterranean. The survey covers 19,548 households and 28,517 individuals on a wide range of topics, encompassing health, socioeconomic status, financial transfers, and intensity of social interaction. For our research, particular

Measuring health status

SHARE devotes a section of the questionnaire to measuring health status. Empirically, health status is an intrinsically unobservable variable. A formal definition of health status unavoidably involves a number of facets suggesting an exploration of different variables available in SHARE to take into account the multiple dimensions among which the notion of health develops. As a first approximation, we can divide the information of health status into objective and subjective. Concerning the

Stockownership and risky investment

During the investment process, an individual makes the crucial decision whether or not to spend some of his resources in financial securities traded in organized exchange markets. SHARE respondents are asked about the amount of money they hold in the following financial products: bank accounts, bonds, stocks, individual retirement accounts, contractual savings for housing, whole and term life insurance. As in Christelis, Jappelli, and Padula (2010), we distinguish between direct stockownership

Unconditional relationship between stock market participation and health status

Our purpose is to investigate the relationship linking stockholding and risky investment to health conditions, identified by four variables (disabilities in daily life activities (ADL), chronic conditions, self-perceived health status (SPHEU) and mental health). To get a first glance of the type of connection, in Fig. 2, Fig. 3 we plot stockholding and risky investments (respectively) over the four proxies of health conditions. In general, objective health seems to have a negative relationship

Conditional relationship between stock market participation and health status

An unconditional analysis may hide the role of other explanatory variables affecting stockholding. For this reason, we estimate a regression model for each of our three dependent variables (direct stockholding, total stockholding, share of risky investment), in which we consider stockholding or the share of risky investment on the left hand side and health status plus a set of control variables on the right hand side. In particular we include the following control variables: age and its square,

Econometric results

This section presents, first of all, the baseline results. We then discuss the results obtained from a deeper analysis conducted to disentangle the channels of health risk transmission.

Baseline results

As mentioned above, we build a regression model for each of our three dependent variables (direct stockholding, total stockholding, fraction of risky investment). Table 4, Table 5 provide the results for the effect of health status on direct stockholding and total stockholding, respectively. Table 6 contains the outcome of the regression estimated for the fraction of risky investment. The whole set of health variables has a negative effect on both stock market participation decision as well as

Disentangling health risk transmission

In this section we try to disentangle the different channels through which health risk may have an effect on investment decisions, thus we wish to provide information on the reasons which bring people in bad health to disaffect from the equity market and to decrease the level of financial risk undertaken. A deeper understanding of the channels of health risk transmission is important because it might involve relevant policy implications. Indeed, it would help to introduce incentives

Risk aversion and health conditions

On the basis of the analysis presented so far, while we can exclude the hypothesis of the shortening of the life time horizon, it is quite difficult to disentangle the other two potential channels of health risk transmission, namely the precautionary saving motive and the direct effect on the marginal utility of consumption (pure risk aversion). Both channels of risk transmission are likely to affect jointly portfolio choices. As a final test of our analysis, we explore whether the health

Conclusion and discussion

Using data from the Survey of Health, Aging and Retirement in Europe, we investigated the relationship between health conditions and portfolio choice for a large sample of people aged 50 and more. We inspected various profiles of human health using several proxies for the health condition: self-perceived health, objective health measured with the number of chronic conditions, and mental health proxied by the number of symptoms of depression. We find only a negative statistically relationship

References (28)

  • J.S. Feinstein et al.

    Elderly asset management

  • A. Finkelstein et al.

    What good is wealth without health?

  • D.P. Goldman et al.

    Medical expenditure risk and household portfolio choice

  • C. Gollier

    The economics of risk and time

    (2001)
  • Cited by (21)

    • What drives stock market participation? The role of institutional, traditional, and behavioral factors

      2023, Journal of Banking and Finance
      Citation Excerpt :

      Elliott and Zhao (2009) and Love and Smith (2010) find evidence consistent with a causal effect, although the latter paper only for married households. Atella et al. (2012) show that the negative effect of bad self-assessed health on the probability to hold stocks is only present in countries with no national healthcare system, while Bressan et al. (2014) show that health measures other than self-assessments are irrelevant for portfolio choice. The relationship between religiosity and stockholding has received a fair amount of attention, but the results are mixed.

    • The effects of health shocks on risk preferences: Do personality traits matter?

      2022, Journal of Economic Behavior and Organization
      Citation Excerpt :

      This paper combines the above two strands of literature to investigate whether health shocks have heterogeneous effects on financial risk preferences, according to personality traits. We focus on acute health conditions, since these conditions (such as stroke, health conditions and cancer) have been found to be more relevant for financial risk preference than chronic conditions (such as arthritis, diabetes, psychological issues, etc.) (Fan and Zhao, 2009; Bressan et al., 2014; Love and Smith, 2010). Further, acute health shocks, at least in terms of their timing, can be thought of as exogenously determined (see for example, Jones et al., 2020).

    • Early health, risk aversion and stock market participation

      2021, Journal of Behavioral and Experimental Finance
    • Changing attitudes to risk at older ages: The role of health and other life events

      2020, Journal of Economic Psychology
      Citation Excerpt :

      Health conditions can increase precautionary savings in the face of (uncertain) increases in future health expenditure (Hubbard, Skinner, & Zeldes, 1995; Palumbo, 1999) and health deterioration can trigger the demand for early retirement (Disney, Emmerson, & Wakefield, 2006), as well as disability benefits (Blundell, Meghir, & Smith, 2002), or reduction in working hours (Trevisan & Zantomio, 2016). Finally, a stream of work has explored the effect of health deterioration on portfolio choice (Atella, Brunetti, & Maestas, 2012; Bressan, Pace, & Pelizzon, 2014; Edwards, 2008; Rosen & Wu, 2004) and finds that declining health conditions increase the ownership of less risky assets, in line with an increasingly important precautionary savings motive.2 Within this stream of literature, one particularly relevant paper is that of Bressan et al. (2014), since it also uses the SHARE data that we use for our study.

    View all citing articles on Scopus
    View full text