Debt and depression

https://doi.org/10.1016/j.jhealeco.2010.02.003Get rights and content

Abstract

We examine the effect of household financial indebtedness on psychological well-being using a large household survey of families with children in Britain. Existing studies that find a link between debt and depression tend to utilise small and highly selective samples of people and only self-reported measures of financial stress, responses to which are likely to correlate with subjective measures of health. From additional household data, we can construct a variety of ‘objective’ quantitative measures of financial stress in order to validate self-reported measures. We show that, although there is a positive association between subjective measures of financial well-being and psychological well-being, individuals differ in their psychological response to objective household financial situations. We also examine how the potential simultaneity of financial and psychological health might be handled.

Introduction

Problems with financial indebtedness and the impact that financial stress has on family well-being have loomed large in media coverage of the consumer credit market in recent years, heightened by the impact of the ‘credit crunch’. This paper investigates the impact of financial indebtedness on the psychological well-being of a large random sample of women with children.

The relationship between financial and psychological well-being has been studied before in a variety of settings. Reading and Reynolds (2001) investigate maternal depression among women with young children. Using self-reported worries about debt on a categorical scale to proxy for the financial position of the household, they obtained a positive link between the extent of self-reported debt concerns and a scaled depression score, controlling for other variables. Hatcher (1994), using hospital records to study patients who had deliberately poisoned themselves, found that 37% had reported problems with financial debts. Maciejewski et al. (2000) show that self-reported financial stress is a significant factor undermining self-efficacy (i.e. appropriate behavioural responses) of individuals with a prior history of depression. Ferrie et al. (2003) examine gradients of morbidity, including measures of depression, across socio-economic groups (SEGs). They found that differences in self-reported financial insecurity across SEGs were a major determinant of differences in the incidence of depression.

More generally, Wildman (2003), using the British Household Panel Survey, shows that scaled self-reported financial status and year-on-year past and expected changes in financial status, are associated with differences in self-reported health, incidence of longstanding illness and depression among survey respondents. Using the same data set, Brown et al. (2005) argue that household psychological well-being is adversely affected by large values of unsecured debt (although not secured debt such as mortgages on homes). They conclude that ‘our results highlight the psychological cost associated with the consumer credit culture in Britain’ (Brown et al., 2005, p. 642). Drentea (2000), using US credit card data, concludes that high credit card debt-to-income ratios and arrears and defaults are associated with greater anxiety, especially among older respondents. We can also note the symmetric result: that positive ‘shocks’ to the financial well-being of the household tend to be associated with improvements in psychological well-being (Gardner and Oswald, 2007) and in health more generally (Fritjers et al., 2005).

There are two potential methodological problems with these studies, although they may not ultimately undermine their basic findings. First, existing studies on the association between financial difficulties and the psychological well-being of individuals or families typically utilise questions that invite respondents to provide rather general subjective perceptions of the household's financial insecurity or distress (though, as we have seen, a few utilise measures of aggregate spending on unsecured debt). These general responses are often treated as proxies for actual indebtedness and for underlying household budgetary problems. However, it is likely that respondents who are depressed or anxious may perceive a given set of financial circumstances in a different light from individuals who are not depressed or anxious. Moreover, in an interview setting, respondents may see self-reported financial difficulties as a socially acceptable rationalisation for psychological difficulties, especially where they lead to traumatic outcomes, as in the study by Hatcher (1994). We should thus perhaps err on the side of caution in making inferences from subjective questions on financial indebtedness.

Ideally, therefore, if we are to fully understand the wider consequences of household indebtedness, the issue of bias in subjective perceptions of financial insecurity needs to be addressed. One general strategy to deal with reporting bias is to search for additional evidence which allows us to bound potential reporting error and to examine its sensitivity to model assumptions (as in Kreider and Pepper, 2008). A complementary strategy is to replace or augment self-reported responses to general questions on financial well-being with more ‘objective’ and comprehensive (albeit, still likely to be self-reported) measures of the financial circumstances of the household by which we, the researchers, can judge whether there is indeed a risk of financial difficulties or debt troubles. Moreover, with quantitative measures of indebtedness, it is possible to be more precise as to the link between the degree of financial difficulties and psychological well-being – for example, the point at which the level of arrears becomes a significant predictor of depression.

This paper uses data from a household panel data set for the United Kingdom (UK): the Families and Children Survey (FACS), over the period 1999–2005. Unlike other UK panel data sets, this data set provides very detailed information on the family's financial circumstances as well as asking more general questions of respondents concerning perceived financial difficulties and debt problems, as well as a range of detailed health questions. Moreover, the typical respondent in this data set is a woman, and there is evidence elsewhere that women are twice as likely to suffer (or at least report) depressive symptoms as men.1 We utilise responses to the precise questions concerning the financial circumstances of the household to form indicators of ‘objective’ financial well-being and in doing so are able to report two findings.

First, we find evidence of a positive association between subjective measures of financial well-being and self-reported depression, controlling for other variables. Respondents reporting ‘financial stress’ or ‘debt problems’ do indeed also tend to report a greater incidence of depression and other indicators of adverse psychological well-being. Second, the probability of reporting financial stress or debt problems can be linked to specific or ‘objective’ household financial circumstances. However, we find that without the inclusion of subjective indicators of financial stress in the statistical analysis, the direct link between the underlying ‘objective’ indicators of the financial position of the household and psychological stress is quite weak. In other words, the link between psychological and financial well-being is mediated through household-specific perceptions, suggesting that households prone to stress and depression respond quite differently to given household financial circumstances than those that are less prone. Therefore only a weak link exists between ‘objective’ measures of financial well-being and psychological stress.2

The second potential problem with the literature is that there may be a two-way causation between psychological well-being on the one hand and economic status on the other (although this simultaneity is difficult to test, as we discuss later). For example, individuals with a history of depression, mental illness and stress may have erratic employment and earnings histories, thereby generating financial difficulties, including indebtedness. Past psychological problems may also cause such individuals to be less able to cope with adverse financial shocks.

It is noteworthy that one of the first contributions by economists to the field of economic and psychological well-being drew the link wholly from mental illness to economic circumstances, and not the other way round (Bartel and Taubman, 1986). Although some studies (such as Theodossiou, 1998) continue to presume a unidirectional link from economic events to psychological events, other studies have attempted to explore the simultaneous link between economic and psychological circumstances, either by assuming that individual responses to economic shocks are mediated through unmeasured individual characteristics (Björklund, 1985) or by estimation of an explicit simultaneous model of the relationship between the two, as in the study of unemployment and mental health by Hamilton et al. (1997).

We attempt to address this potential endogeneity by estimating the relationship between subjective indicators of financial well-being and self-report depression simultaneously as a recursive bivariate probit (Greene, 1997). We also exploit the panel aspect of the data to control for person-specific effects. Taking these results together, the following ‘story’ emerges. Adverse financial circumstances cumulatively lead the respondent to an increasing probability of reporting financial difficulties. In turn, self-reported ‘debt problems’ and ‘financial stress’ have adverse effects on self-reported psychological well-being (depression). So although ‘objective’ measures of a household's financial circumstances have a limited direct effect on psychological well-being, they have an indirect effect mediated through the individual's propensity to perceive such factors as resulting in financial difficulties. This is similar to the ‘self-efficacy’ argument described earlier. It appears that it is this interaction of the respondent's perception of financial difficulties with ‘objective’ measures of indebtedness that lies at the heart of the relationship between psychological and subjective financial well-being.

The remainder of this paper is structured as follows. In Section 2 we describe our data, and provide some descriptive statistics. Section 3 examines the degree of overlap between psychological and financial well-being, Section 4 provides our main empirical results while Section 5 concludes.

Section snippets

The Families and Children Survey

This paper uses data from the UK's Families and Children Survey (FACS) for the period 1999–2005. The FACS was first established in 1999 as the Survey of Low-Income Families (SOLIF) and was originally designed to elicit information on household characteristics, health status and the economic and financial position of a sample of low-income families with children.

The original sampling frame of the FACS comprised all sampled lone parents with children, as well as couples with children where the

Psychological and financial well-being: correlations

Given the existing literature, and its limitations in terms of population representativeness, our first task is to confirm for this large random sample that there is indeed a relationship between self-reported depression on the one hand and self-reported measures of financial difficulties and over-indebtedness on the other.

To approach this task, Table 2 examines the degree of overlap between self-reported financial difficulties and depression. It shows that the incidence of depression is higher

Modelling the impact of indebtedness on psychological well-being: univariate probit

Our first model is broadly designed to replicate existing studies, while using the FACS data set. In doing so we test whether the correlation between self-reported depression and financial difficulties arises primarily from individual-specific correlated responses to subjective questions concerning financial well-being, or whether ‘objective’ measures of a household's financial circumstances are more powerful factors in explaining psychological well-being.

To examine this, we pool our data

Conclusions

This paper examines the relationship between depression and debt among families, using data from a large representative sample of families with children, where the mother is the normal respondent. We delineate several problems with existing studies in this area: small and selective samples, the use of rather general qualitative questions on a household's level of indebtedness, and the lack of simultaneous modelling of the probability of reporting depression and financial difficulties.

We find

Acknowledgements

Our thanks to Experian Ltd. and the University of Nottingham for funding this work, the ESRC Data Archive for providing the data set, and the useful comments of participants at a seminar at the Centre for Economic Performance, LSE, and two referees of this journal.

References (35)

  • M. Baker et al.

    What do self-reported, objective, measures of health measure?

    Journal of Human Resource

    (2004)
  • A. Bartel et al.

    Some economic and demographic consequences of mental illness

    Journal of Labor Economics

    (1986)
  • H. Benitez-Silva et al.

    How large is the bias in self-reported disability?

    Journal of Applied Econometrics

    (2004)
  • A. Björklund

    Unemployment and mental health: some evidence from panel data

    Journal of Human Resources

    (1985)
  • J. Bound

    Self-reported versus objective measures of health in retirement models

    Journal of Human Resource

    (1991)
  • J. Bound et al.

    The dynamic effects of health on the labor force transitions of older workers

    Labour Economics

    (1999)
  • S. Bridges et al.

    Use of credit and arrears on debt among low income families in the United Kingdom

    Fiscal Studies

    (2004)
  • Bridges, S., Disney, R., 2006. Debt and depression. Centre for Finance and Credit Markets Working Paper 06/02, School...
  • S. Brown et al.

    Debt and distress: evaluating the psychological cost of credit

    Journal of Economic Psychology

    (2005)
  • T.C. Buchmueller et al.

    Access to physician services: does supplemental insurance matter, evidence from France

    Health Economics

    (2004)
  • G. Chamberlain

    Analysis of covariance with qualitative data

    Review of Economic Studies

    (1980)
  • Charles, K.K., 2002. Is retirement depressing? Labor force inactivity and psychological well-being in later life....
  • D. Deadman et al.

    Offenders as victims of crime?: an investigation into the relationship between criminal behaviour and victimization

    Journal of the Royal Statistical Society (Series A)

    (2004)
  • R. Disney et al.

    Ill-health and retirement in Britain: a panel data-based analysis

    Journal of Health Economics

    (2006)
  • P. Drentea

    Age, debt and anxiety

    Journal of Health and Social Behavior

    (2000)
  • R.W. Fairlie

    The effects of home computers on school enrolment

    Economics of Education Review

    (2005)
  • J. Ferrie et al.

    Future uncertainty and socio-economic inequalities in health: the Whitehall II study

    Social Science and Medicine

    (2003)
  • Cited by (0)

    View full text