Policy implications of the multidimensional nature of poverty

Despite the consistent effort to reduce hunger and poverty, a sizeable proportion of the population in India is living below the poverty line and 36% women and 34% men were underweight (IIPS & Macro International, 2007). Malnutrition and poverty form a vicious circle of poverty which needs to be removed through government intervention. In the context, using data from the national representative “level and pattern of consumer expenditure survey” (66th Round 2009-10) of the National Sample Survey Organisation (NSSO) we have tried to establish the association between nutrition status and household income. In order to take into account the heterogeneous effect of the calorie intake across the income distribution and endogeneity of the calorie consumption we have applied instrumental variable quantile regression. Regression result confirms the heterogeneous impact of per-capita calorie intake across household income distribution. Result shows that the marginal effect of per consumer unit calorie intake on income decreases with the increase in income. Calorie intake elasticity of income gain is 2.077 at the lowest 10th quantile than the 1.755 at the highest 90th quantile of income distribution. Study clearly shows that public intervention at the low level of nutrition intake will have relatively higher gain in income.

This paper explores behavioural changes resulting from the presence of a background risk. Due to markets incompleteness, not all risks are insurable. The literature suggests that, according to the structure of preferences, agents bearing a background uninsurable risk are less willing to bear other insurable risks and increase their demand for insurance. The empirical evidence of this effect is limited and, despite the relevance of this question, unexplored in developing countries. This paper fills this gap. It explores the effect of a background risk on the decision to buy health insurance using household data from the Palestinian Territories. We consider the risk of food insecurity as a background uninsurable risk. Using a bivariate probit model, we find that the propensity to buy health insurance is positively affected by the presence of a background risk of food insecurity. When allowing the background risk to vary in intensity, we find that the propensity to insure is higher as the background risk becomes more intense. These results are robust to alternative indicators of background risk. The study shows that, in presence of background risks, there might be incentive changes towards the desirability of insurance that have implications for policy design.

This paper evaluates the health card subsidy, issued by Indonesian government after the financial crisis in 1997, benefits the elderly health care demand. The health card subsidy provided free access to health services; and was accepted at all health center or ‘puskesmas’. Using the Indonesia Socio-economic survey data (Susenas) for 2003, our results show health cards issued to the poor consistently caused increase the demand on health care facilities among the 55 year old population. Unfortunately, it is also found that the health card benefited wealthier individuals in their access of health centers or ‘puskesmas’.

Comparisons of levels of poverty across contexts and circumstances overlook the composition and characteristics of the domain of non-poor. But implications of poverty in terms of compromised welfare as well as its potential for alleviation do depend on the structure and composition of the non-poor. On this premise, this note empirically illustrates a contrast between the levels and burden of poverty for Indian states. It finds that burden of poverty is not linear with levels of poverty implying that similar levels of poverty have differential potential for its alleviation.

Building on the capability approach, this paper discusses the multidimensionality of deprivation in urban Brazil between 2003 and 2008. Based on a temptative four dimensions index (e.g., living conditions, health standards, education levels and labor market participation) relating thirteen indicators, we find that the majority of individuals lived in households with no deprivation or deprived in only one indicator. Then, we explicitly addressed the inter-indicator covariance structure of the thirteen items using factorial analyses. Deprivation tended to come in pairs, as water and sewage, or child labor and school attendance, or in triples, as assets, functional illiterate and non-working adult. After this, we compared this index with poverty measured by income standards for states in Brazil. Evidences confirmed the regional inequality in both types of measurement. However, comparisons showed that correlations between them, although positive and significant, were not strongly. Finally, we discussed the dynamics of deprivation using data for synthetic cohorts and OLS and logistic models. We verified the association between attributes, such as sex, race, cohort and educational level and the propensity of a household to fall and continue in poverty or deprivation.