Targeting of social transfers: Are India’s poor older people left behind?
Introduction
Accelerating demographic change, a persistently large informal sector and weakening family support for the older people have important implications for old-age poverty in developing countries. Multi-generational household models that traditionally provided support to older people are diminishing due to migration and declining fertility (James, 2011). In contrast to the small minority of formal sector workers that benefit from comprehensive social protection and old-age income security, the vast majority of informal sector workers is predicted to face increased risks of old-age poverty in the near future given their lack of social protection coverage (e.g. Lloyd-Sherlock, 2000). Implemented as cash transfers, social pensions aim to mitigate old-age poverty faced by individuals who lack social protection coverage (Holzmann & Hinz, 2005). To improve the old-age income security of the poor, in 1995, the Indian government introduced the National Old Age Pension Scheme (Government of India, 1995).1
The National Old Age Pension Scheme targeted towards poor older people belongs to the National Social Assistance Programme which supports vulnerable groups through various social assistance schemes (Government of India, 1995). Motivated by the increasing risks of old-age poverty in the presence of demographic change and widespread informal sector employment, I focus on evaluating the targeting performance of the social pension scheme for poor older people.2
The effectiveness of social pensions in terms of old-age poverty reduction depends essentially on whether social pensions reach the poor older people or not. However, the targeting performance remains an under-researched topic in India. Existing studies do not focus on the targeting performance and suffer from different limitations. Dutta et al., 2010, Gupta, 2013 analyzed the implementation of social pensions in a descriptive manner for a few selected states. Chopra and Puddussery (2014) assess the implementation of social pensions but focus in their analysis only on benefciairies and hence cannot assess the targeting performance. The latest study by Kaushal (2014) used repeated cross-sectional data for all of India but lacked data on social pension receipt and needed to approximate beneficiary status. Research on social pensions in other countries including Brazil and South Africa has made the importance of social pensions for poverty reduction evident. The impact of social pensions is not restricted to the well-being of direct beneficiaries; other household members and especially grandchildren seem to benefit as well from the cash transfer (e.g. Duflo, 2000, Edmonds et al., 2005, Lloyd-Sherlock, 2006).
This paper contributes to the existing literature in two ways: First, improving our understanding of whether social pensions reach the poor older people is an important prerequisite for analyzing the effectiveness of social pensions in India and other developing countries with similar institutions and similar targeting challenges. Second, methodologically I contribute to existing targeting studies by comparing indicators of the targeting performance to a hypothetical random allocation of social pensions. Moreover, in examining the relevant factors affecting access to social pension benefits, the availability of panel data allows me to minimize potential omitted variable bias and a placebo test shows that the identified factors are indeed relevant and not just driven by spurious correlations.
To address limitations in targeting and coverage of social pensions, the central government introduced social pension reforms in 2007. The results of this study indicate that from 2004–05 to 2011–12, these national level reforms contributed to a reduction of the exclusion and inclusion error but both targeting errors continue to be very high. Comparing the actual allocation of social pensions to a hypothetical random allocation, the results suggest that the benefits from targeting are relatively small for the exclusion error and relevant but decreasing over time for the inclusion error. Even though the allocation of social pensions has moved towards the Below Poverty Line (BPL) card as a more observable criterion, this criterion itself is too weakly implemented to achieve effective targeting of the poor. BPL card holding is relevant for individuals from both asset poor and asset non-poor households to access social pensions and individuals who have direct connections with the local government have higher chances to receive the benefits.
The remainder of the paper is structured as follows: Section 2 provides background information on the implementation of social pensions in India and summarizes existing literature in this field. Section 3 presents the theoretical framework by describing the targeting challenge and how social pension reforms in the Indian context are related to it. Section 4 describes the data and explains the methodology. In Section 5, I present the results from the empirical analysis before concluding in Section 6.
Section snippets
Background: The need for social pensions in India and national reforms
The need for an effective social pension scheme for poor polder people in India has been reinforced by progressing demographic change. As presented in Fig. 1, for the time period of 2010 to 2050, India’s population aged 60 years and older is expected to triple (from 96 million to 316 million) while India's population of individuals aged younger than 60 years is expected to grow only by 18 percent in the same time period (from 1.134 billion to 1.342 billion). After 2050, the United Nations World
Theoretical framework
After briefly summarizing the theoretical literature on the targeting challenge in general, I describe the theoretical expectations on the effect of the social pension reforms on the targeting performance of social pension benefits in the Indian context.
The India Human Development Survey
The IHDS was conducted by the National Council of Applied Economic Research and the University of Maryland (Desai et al., 2007, Desai et al., 2015). This nationally representative individual-level panel dataset surveyed 41,554 households (215,753 individuals) in 1503 villages and 971 urban neighborhoods across India using a stratified, multistage sampling procedure in 2004–05 and re-interviewed households in 2011–12. The survey is spread over all the states and union territories of India except
Descriptive statistics
The sample used for the analysis is only restricted by the age of the individuals and therefore includes beneficiaries and non-beneficiaries of social pensions. It consists of all older people who are at maximum 10 years below the eligibility age and surveyed twice by IHDS (balanced panel). The summary statistics are shown in Table 2 separately for 2004–05 and 2011–12.
Conclusion
This study aimed to examine the targeting performance of social pensions in India and to answer the question of who receives social pension benefits. The descriptive statistics show that from 2004–05 to 2011–12, a time period encompassing important national social pension reforms, the targeting of social pensions improved but both targeting errors continue to be very high. The exclusion error reduced substantially from 92 percent to 74 percent and the inclusion error from 57 percent to 41
Acknowledgements
I thank Ankush Asri for providing insights and expertise that greatly assisted the research and I am very grateful to Katharina Michaelowa for valuable discussions and suggestions. This paper has also benefitted from numerous helpful comments by Andy McKay, Krisztina Kis-Katos, Kai Gehring, Michael Schleicher, Sourabh B. Paul, Sitakanta Panda, Simon Bornschier and Saskia Ruth and by participants at the ADBI-AGI Workshop on Ageing in Asia 2016, India Human Development Survey User Conference
References (54)
- et al.
Interaction terms in logit and probit models
Economics Letters
(2003) - et al.
Using an asset index to assess trends in poverty in seven sub-Saharan African countries
World Development
(2008) - et al.
Recent social security initiatives in India
World Development
(2017) How public pension affects elderly labor supply and well-being: Evidence from India
World Development
(2014)Old age and poverty in developing countries: New policy challenges
World Development
(2000)- et al.
Influence of literacy on India’s tendency for age misreporting: Evidence from census 2011
Journal of Population and Social Studies
(2015) Performance of social safety net programs in Uttar Pradesh
(2007)- et al.
Selecting a targeting method to identify BPL households in India
Social Indicators Research
(2013) - et al.
Manufacturing confusion: BPL Lists in Uttar Pradesh
Economic & Political Weekly
(2014) - et al.
Mostly harmless econometrics
(2009)