Paper short abstract:

This paper argues that two types of biases may lead to the dampening of the economic gradient in self-perceived health in developing country contexts, a) inconsistency and b) positional objectivity. It proposes an alternative measure of SRH after adjusting for variations in age, education, and cognitive ability. This measure has a strong economic gradient for aged individuals in India.

Paper long abstract:

Research Content

Research has shown that indicators of population health are worse when economic inequality is higher. While the existence of a socioeconomic gradient in the capability to live a healthy life for the elderly in developed countries is well-documented, there has been relatively less research on the economic differential in the health of the elderly in developing countries. Analyzing data from health and retirement surveys conducted in Brazil, India and China, scholars have shown that while the socioeconomic gradient in frailty (the likelihood of poor recovery from a physiological stressor) of the elderly was strong when education was used as the indicator of socioeconomic status, evidence was mixed and inconsistent when economic indicators such as income or wealth were used. Studies have noted the income gradient to be weaker than the education gradient even when the indicator of self-rated health (SRH) was used. Counter-intuitive results on the linkages between socioeconomic status and self-perceived health led to a complete dismissal of these indicators by a section of important scholars (Sen (1993; 2002; 2009).

Though recent studies vouch for the validity of SRH in developing countries such as India (Cullati et al. 2018), surveys still show that the absence of a strong economic differential in SRH. The Longitudinal Ageing Study in India (LASI) Wave-1 (2017-18) shows that among the elderly population in India, the proportion reporting poor health varies between 24.4 for the richest economic quintile and 26.4 for the poorest. Contrastingly, it varies between 26.5 for people with no education and 15.7 with those who have completed 10 or more years of schooling.

Methodology and Analysis

This paper argues that two types of biases may lead to the dampening of the economic gradient in self-perceived health in developing country contexts, namely a) inconsistency and b) positional objectivity (Sen 1993; 2002; 2009). Inconsistent responses would mean that an individual provides substantively different ratings of health if they are asked the same question a second time (in the same survey) using a different rating scale. Consider a person who says that she has ‘very good’ health when asked to rate her health in a scale of 1-5 where 1 is poor, 2 is fair, 3 is good, 4 is very good and 5 is excellent. The same person says that she has ‘poor’ health when asked to rate her health in a scale of 1-5 where 1 is very poor, 2 is poor, 3 is fair, 4 is good and 5 is very good. We would consider her responses to be inconsistent.

In developing countries like India, a second type of bias may further confound indicators of self-perceived health, namely positional objectivity a la Amartya Sen (1993; 2002; 2009). Sen (2002) dismissed these indicators, since they ‘can thoroughly mislead public policy on health care and medical strategy’ (p. 860). Sen (2009) again wrote, ‘[t]he comparative data on self-reporting of illness and the seeking of medical attention call for critical scrutiny, taking serious note of positional parameters’ (p.165). He provided the contrasting examples of certain Indian states, namely, Kerala, Bihar, and Uttar Pradesh on the other. While Kerala is the best-performing Indian state in terms of life expectancy at birth (even higher than that of China) and a successful health transition, it also has the highest figures for self-perceived morbidity. Bihar and Uttar Pradesh are typical less developed states not undergoing any health transition and having low life expectancies at birth but ‘astonishingly low rates of self-assessed morbidity.’ Sen understood such responses for self-perceived health to be positionally objective. Responses vary, but not subjectively, they are instead conditional on positional parameters such as education, awareness, and access to healthcare. Thus, in Bihar, where these parameters are poor, a person’s understanding of their actual health condition may be systematically limited.

We are unaware of scholarly work that has comprehensively dealt with both types of biases. This paper is novel in its approach since it proposes a simple method that would purge the self-rated health responses of the two types of biases.

We use (LASI) Wave 1 data to estimate the association between economic status We find that age, cognitive ability (measured by word recall and verbal fluency) and educational attainment significantly determine the likelihood of providing inconsistent responses to the two questions on SRH. We then use age, cognitive ability, and educational attainment to predict SRH ratings. We argue that this also controls for biases arising from positional objectivity since education has been shown to be the most important positional parameter for SRH (Subramanian et al. 2009; Mukhopadhyay et al. 2022). Running state-fixed effects regression to obtain adjusted SRH controls for all unobservable positional parameters. This new measure of SRH is seen to have a strong economic gradient for aged individuals in India. The substantive finding that the economic gradient is dampened by the aged poor systematically reporting better health because of certain biases merits attention of ageing research and health policy in India.

Conclusion

India is currently undergoing a unique wave of demographic changes. The much-discussed demographic dividend is accompanied by a dramatic increase in dependency ratio and the population above 60 years (from 8.6% to 19.5%, at 319 billion in 2050, as projected by UN estimates). This paper concludes that a weak economic gradient, which is obtained when one uses the unadjusted SRH scores, may misguide public health policy. Our results show that the aged poor in India indeed have worse health and deserve policy focus.

References

Cullati S, Mukhopadhyay S, Sieber S et al. 2018. Is the single self-rated health item reliable in India? A construct validity study. BMJ GlobalHealth 3(6): 1-12.

Sen A.K. 1993. Positional objectivity. Philosophy and Public Affairs 22:126–145.

-- 1998. Mortality as an indicator of economic success and failure. Economic Journal 108:1­25.

--2002. Health: perception versus observation. British Medical Journal 324:860–861.

-- 2009. The Idea of Justice. Cambridge, Mass: Belknap Press of Harvard University Press.

Keywords: Self-rated health, positional objectivity, ageing, socioeconomic gradient, inequality