Longitudinal analysis of income-related health inequality

https://doi.org/10.1016/j.jhealeco.2009.10.005Get rights and content

Abstract

This paper considers the characterisation and measurement of income-related health inequality using longitudinal data. The paper elucidates the nature of the Jones and López Nicolás (2004) index of “health-related income mobility” and explains the negative values of the index that have been reported in all the empirical applications to date. The paper further presents an alternative approach to the analysis of longitudinal data that brings out complementary aspects of the evolution of income-related health inequalities over time. In particular, we propose a new index of “income-related health mobility” that measures whether the pattern of health changes is biased in favour of those with initially high or low incomes. We illustrate our work by investigating mobility in the General Health Questionnaire measure of psychological well-being over the first nine waves of the British Household Panel Survey from 1991 to 1999.

Introduction

A strong relationship between socioeconomic status and health has been documented in numerous studies: for example, individuals with high income are healthier than those with low income (Benzeval and Judge, 2001, Deaton, 2003, Gerdtham and Johannesson, 2004). While there has been an increasing amount of literature on health inequality, income-related health inequality and its determinants (Wagstaff and Van Doorslaer, 2000), little attention has focused on measuring health mobility or whether the health of the poor is improving relative to the rich over time. This is an important issue since significant income-related inequalities in health have persisted, and even increased, in some western countries over the last decade in spite of considerable improvements in average health status (Van Doorslaer and Koolman, 2004). As a result, most European governments have recognised the need to tackle income-related health inequalities. For example, England is committed to reduce socioeconomic inequalities in infant mortality and life expectancy at birth by 10% from 1997 to 1999 baseline levels by 2010 (Department of Health, 2008).

The main measure of income-related health inequality within the health economics literature is the concentration index (Wagstaff and Van Doorslaer, 2000). This captures the extent to which good health in any period is concentrated among the rich compared to the poor and is equal to twice the covariance between health and income rank normalised by average health. Changes in the concentration index over time can be analysed in the manner of Wagstaff et al. (2003) or Gravelle and Sutton (2003) using repeated cross-sections, but it is not possible thereby to track the experience of individuals but only of groups, such as the poor, whose composition may change over time. To better understand the dynamics of health and income, longitudinal or panel data on individuals are required. For example, longitudinal data are required to distinguish between income-related health inequalities arising from chronic or persistent social disadvantage as opposed to those due to transitory episodes of both poverty and sickness, where the former would call for policies to tackle the structural problems that trap some individuals in deprivation and ill-health while the latter might demand measures such as improvements in acute health services or temporary welfare assistance.

In a pioneering paper on the use of longitudinal data to analyse income-related health inequalities, Jones and López Nicolás (2004; hereafter JLN) aim to explain the relationship between a set of T period-specific or short-run concentration indices CIt (t = 1, …, T) and the long-run concentration index CIT obtained from income and health data averaged over all T periods. In particular, they propose an index of “health-related income mobility”, modelled on Shorrocks’ (1978) income mobility index, that measures the extent to which income-related health inequality is greater or smaller in the short-run than in the long-run. JLN illustrate the use of the index with data for the UK and it has since been employed in a number of other empirical studies (Lecluyse, 2006, Hernández-Quevedo et al., 2006, Brandrup and Kortt, 2007) using health data sets for Belgium, European Union member states and Australia respectively. All these studies report negative values for the health-related income mobility index, suggesting that income-related health inequality, unlike income inequality (see Wodon and Yitzhaki, 2003), is typically greater the longer the time span over which the measurements are taken. JLN use a numerical example to illustrate that negative values will arise if “individuals who are downwardly (income) mobile, in the sense that, in the long run, their income rank is lower than in the short run … have a lower than average level of health in the short run, compared to individuals who are upwardly mobile” (JLN, p. 1019). However, it is not entirely obvious what this means in terms of the relationship between health and income as the mathematical properties of the measure are not self-evident.

The current paper makes two main contributions. First, the paper further elucidates the nature of the JLN index of health-related income mobility. In particular we show that the negative values of the index reported in the literature may be explained by the common unimodal shape of the income distribution (Chotikapanich, 2008) in conjunction with the strength of the positive association between income and health in the long-run compared to the short-run. Secondly, the paper develops a complementary approach to the analysis of longitudinal data that brings out other policy-relevant aspects of the evolution of health-related inequalities over time. In particular, we propose a new index of “income-related health mobility”, based on a decomposition of the change in the short-run concentration index CIt over time, which measures whether the pattern of relative health changes between two periods is biased in favour of those with initially low or high incomes. We illustrate our work, like JLN, by investigating mobility in the General Health Questionnaire (GHQ) measure of psychological well-being (Goldberg and Williams, 1988) over the first nine waves of the British Household Panel Survey (BHPS) from 1991 to 1999. The structure of the paper is as follows: Section 2 presents a critical review of the JLN mobility index. Section 3 presents our alternative mobility analysis and explores the properties of the proposed new indices. Section 4 summarises and concludes the paper.

Section snippets

A critical exposition of the JLN health-related income mobility index

We start by providing a brief outline of JLN's measurement framework in order to establish the basis for our subsequent analysis. JLN investigate the relationship between the set of T short-run CIs:CIt=2h¯tcov(hit,Rit)=2Nh¯ti(hith¯t)Rit12;t=1,,Tand the long-run concentration index defined over all T periods,CIT=2h¯¯Tcov(hiT,RiT)=2Nh¯¯Ti(hiTh¯¯T)RiT12where hit is a cardinal measure of health for individual i(i = 1, …, N) in period t, h¯t is the average health of the population in period t,

A new approach to the longitudinal analysis of income-related health inequalities

The particular value of the JLN index to health policy-makers is that it shows the persistence of income-related health inequalities. In particular the negative estimates that have been reported to date imply that long-term or chronic problems of income-related health inequalities are typically more severe than would be inferred from cross-sectional estimates. By implication, policies designed to tackle income-related health inequalities need to address structural problems that trap some

Conclusions

The characterisation and measurement of the evolution of income-related health inequalities in a population over time is important for the design and evaluation of policies to reduce such inequality. This study has extended a key paper in the literature by demonstrating that the common negative value of the JLN health-related income mobility index is likely to be explained by the stronger positive association between permanent, income and health disparities across individuals than between

Acknowledgements

The authors bear sole responsibility for the further analysis and interpretation of the British Household Panel Survey data employed in this study. Financial support (Gerdtham) from the Swedish Council for Working Life and Social Research (dnr 2006-1660 and dnr 2007-0318) is gratefully acknowledged.

References (33)

  • P. Contoyannis et al.

    The dynamics of health in the British Household Panel Survey

    Journal of Applied Econometrics

    (2004)
  • A. Deaton

    Health, inequality, and economic development

    Journal of Economic Literature

    (2003)
  • Department of Health, 2008. Health Inequalities: Progress and Next Steps....
  • U.-G. Gerdtham et al.

    Absolute income, relative income, income inequality and mortality?

    Journal of Human Resources

    (2004)
  • D. Goldberg et al.

    A User's Guide to the General Health Questionnaire

    (1988)
  • Cited by (31)

    • Regional innovation ability and its inequality: Measurements and dynamic decomposition

      2022, Technological Forecasting and Social Change
      Citation Excerpt :

      Second, based on the essential characteristics of innovation ability and the internal attributes of measurement methods, the traditional inequality measurement method is extended to match the regional economic level with innovation ability, and the concentration index (CI) is used to measure the inequality of innovation ability related to economic level. Third, on the basis of considering inequality measurement method characteristics and policy oriented needs, we use the dynamic decomposition method proposed by Allanson et al. (2010) to classify the change of regional innovation ability inequality into “regional economic growth effect” and “innovation ability cumulative effect,” further decomposing the latter into “innovation ability distribution effect” and “innovation ability scale factor.” In addition, this research endeavors to explore four primary issues in the time series of this paper from 1978 to 2020.

    • Do socioeconomic health gradients persist over time and beyond income? A distributional analysis using UK biomarker data

      2021, Economics and Human Biology
      Citation Excerpt :

      In context of the association of socioeconomic status (SES) with health, these measures does not establish how these different dimensions of deprivation may be distributed across individuals and over time or how they are related to health outcomes (Blázquez et al., 2014). Consequently, the inequalities in health (Allanson and Petrie, 2013; Allanson et al., 2010) vary with the measures of socioeconomic status and are difficult to reconcile (Lindelow, 2006). There is thus a need for “... more careful research on how different dimensions of SES are related, and on the pathways by which the respective dimensions impact on health related variables” (Costa-Font and Hernández-Quevedo, 2012).

    • Thank goodness for stickiness: Unravelling the evolution of income-related health inequalities before and after the Great Recession in Europe

      2020, Journal of Health Economics
      Citation Excerpt :

      This occurred due to the fact that the austerity measures reduced the redistributive effect of transfers, especially pensions. While this extension of the Allanson et al. (2010) decomposition is informative about how the separate income sources influence IRHI evolution via re-ranking changes, it is silent about the distinct role played by factors such as age, gender, region or income that are associated with health. Since – to take one example – transfer and market incomes are distributed differently across various age groups, such a distinction is likely to be important in tracing out their distinct roles.

    View all citing articles on Scopus
    View full text