The role of medical care in explaining the link between health and economic resources.

Publication: ASEAN Economic Bulletin

Publication Date: 01-APR-06

Bookmark this article Print this article Link to this article Email this article Digg It! Add to del.icio.us RSS

COPYRIGHT 2006 Institute of Southeast Asian Studies (ISEAS)

I. Introduction

The positive association between economic resources and health has been reported extensively in the previous literature. It has also been shown that this association has antecedents in childhood (Case, Lutotsky, and Paxson 2002; Currie and Stabile 2003). However, the direction of causation remains unclear, particularly as to whether economic resources affect health or vice versa (Smith 2005).

Pritchett and Summers (1996) used instrumental variables estimation with cross-country, time series data to find that the positive income effect on health is causal and structural. Ettner (1996) also used instrumental variable estimation and found that an increase in income significantly improved mental and physical health.

However, Smith (1998, 1999) eloquently argued that the direct causal link runs from health to income and wealth, not the other way round, especially for the elderly American population. Moreover, Meer, Miller, and Rosen (2003) showed that the causal link running from wealth to health is in fact illusory, using individual-level data from the Panel Study of Income Dynamics and inheritance as an instrument for the change in wealth. (1) Finally, Adams et al. (2003) directly tested for the absence of causal links between socio-economic status (SES) and health innovations and mortality, using data from the Asset and Health Dynamics of the Oldest Old Panel, and found no supporting evidence for a direct causal link between SES and mortality and the incidence of most sudden onset health conditions. (2)

Our research takes a different step to explore the link between health and economic resources by studying the intermediary role of medical care utilization. While Newhouse (1993) and Smith (2005) cast doubt on the role of medical care utilization, Bindmand, Keane, and Lurie (1990) and Currie and Gruber (1996) find supporting evidence for it.

The present study uses data from the third Indonesian Family Life Survey (IFLS3) to examine the association between economic resources and individual health status while controlling for other individual health risk factors. This study makes additional use of the data, where variables on medical care utilization are available, to examine the mechanism whereby economic resources may affect individual health.

The remainder of this paper is organized as follows. Section II describes in detail the data that is used in this study. Section III presents the empirical results that show a positive association of economic resources with health and examines the role of medical care utilization in explaining this association. Section IV concludes the paper.

II. Data and Descriptive Statistics

Data from the third wave of the IFLS3 (collected from June to November in 2000) is used in the empirical analysis. Over 30,000 individuals in 7,224 households, representing about 83 per cent of the Indonesian population living in thirteen of the nation's twenty-six provinces, were originally sampled and followed up in a large-scale socioeconomic and health survey (Strauss et al. 2004a; Strauss et al. 2004b). (3)

One of the key variables for the empirical analysis is household expenditure as a measure of economic status. Household expenditure is less prone to measurement error than wealth or income and more likely to provide an accurate picture of economic well-being (Frankenberg, Thomas, and Beegle 1999; Strauss et al. 2004a). Household expenditure for thirty-seven food and nineteen non-food items including housing expenses was cumulated and converted to a monthly equivalent, as in Frankenberg, Thomas, and Beegle (1999). It was then divided by the square root of the total number of people in the household to equivalize household expenditure. This equivalized expenditure captures the amount of economic resources available to each household member, taking into account the fact that people in a larger household are worse off than those in a smaller household conditional on the same level of household economic resources (Deaton 2001). Figure 1 shows that non-parametric kernel density estimates of log equivalized expenditure appears to support the normality assumption.

[FIGURE 1 OMITTED]

Another key variable in the analysis is one used to measure individual health status. Both self-rated individual health status, measured on a four-point scale from "unhealthy (1)", "somewhat unhealthy (2)", and "somewhat healthy (3)" to "healthy (4)", and nurse-assessed health status, measured on a nine-point scale from "the most unhealthy (1)" to "the most healthy (9)", are used in the empirical analysis. While the drawbacks of the former measure have been well explained (Strauss and Thomas 1998), Idler and Benyamini (1997) found that the binary indicator of poor health derived from it is a powerful predictor of subsequent mortality, even after controlling for more objective indicators of individual health. (4) Moreover, the use of a more objective health measure such as nurse-assessed individual health status will provide assured robust empirical evidence. (5)

Table 1 presents the...

Read the full article for free courtesy of your local library.