Successes in social protection: what lessons can be learned?

Abstract

The recent food, fuel and financial crises have triggered innovation in social protection programmes around the world. Existing evidence suggests that such programmes both protect people from shocks and enable them to participate in economic growth. This paper explores lessons learned in social protection initiatives across countries and contexts, and identifies a set of ‘success conditions’ for social protection programmes, focusing mainly on sub-Saharan Africa. Ghana's national health insurance illustrates well three key features of an enabling environment: political commitment at the top; high administrative capacity; and financial sustainability and commitment to programme spending controls. To influence policy, these lessons need specification to the given context.

Résumé La crise récente de nourriture, carburant et finance a déclenché l'innovation des programmes de protection sociale autour du monde. La preuve actuelle démontre que ce genre de programmes protège les gens des chocs et les permet de contribuer à la croissance économique. Cet article examine les leçons apprises par des initiatives de protection sociale dans divers pays et contextes. Nous identifions quelques conditions de succès pour les programmes de protection sociale, tout en mettant l'emphase sur l'Afrique subsaharienne. Le programme d'assurance de santé nationale du Ghana illustre bien trois éléments clé d'un environnement favorable : l'engagement politique; une grande capacité administrative; le soutien financier et la régulation continue des dépenses. Pour avoir une influence sur la politique, ces leçons doivent être adaptées au contexte particulier en question.

Keywords:

• social protection
• poverty reduction
• social insurance
• lessons learning
• developing countries

Acknowledgements

We thank Francois Bourguignon, Charlotte Bué, Stefan Dercon, Stephan Klasen and Stefania Innocenti for discussion and help, and for letting us draw on common work done in the context of the European report on development 2010: social protection for inclusive growth (EUI 2010). Remaining mistakes are of course ours.

Notes

See Dercon (2011); DFID (2011), EUI (2010), Fiszbein and Schady (2009).

Grosh et al. (2008) provides a comprehensive survey of design issues. There has been a booming industry of assessing impacts of programmes (for example, the Mind-the-Gap Forum, www.impactevaluation2011.org/forum/), which has greatly enhanced the empirical knowledge of specific programmes, but – arguably – at the cost of narrowing the field of analysis. See Hagen-Zanker et al. (2011) for a recent assessment.

In Mexico, the estimated increase in educational attainment is between 0.7 and 1.0 per cent per year; in Brazil, children are now less likely to drop out of school and have lower rates of absence than children living in similar households that are not beneficiaries of cash transfers (Bastagli 2010). For recent analyses of the impact of Bolsa Familia on school enrolment and drop out, see Glewwe and Kassuf (2011) and Holmes et al. (2011).

A notable exceptional case of independent evaluation is that of Oportunidades. Here there was a sequence of evaluations concerning different aspects of the program: education, health, nutrition, fertility and labour supply with the aim of providing hard evidence to be used for the survival of the program in case of a change in government. Indeed, when the change took place, the program was not discontinued (Grosh et al. 2008).

The categorisation of these conditions was the product of the review conducted by the team that wrote the European report on development 2010, and the idea of ‘lessons for success’ was inspired by the ILO working paper by Michael Lipton (1996) on success for anti-poverty programmes.

Reallocation can be politically difficult, depending on which expenditures are substituted. Based on efficiency and equity motivations for instance, Indonesia moved out of fuel subsidies into a comprehensive social protection plan, including health, education and nutrition (Sudarto and Bazzi 2011). Also, Egypt is shifting resources from subsidies to social protection on the basis of higher efficiency (IMF 2011).

A very large proportion of health spending in developing countries is financed privately, which is likely to be a main reason for slow progress in health indicators.

Coverage of the programmes varies but tends to be broad. At the same time, spending in the Latin American examples is kept low with total benefits per recipient being relatively low: between four per cent of beneficiary consumption in Honduras and 20 per cent in Mexico (Bastagli 2010).

Data from the IDLA dataset: http://www.wider.unu.edu/research/current-programme/en_GB/Impact-of-Economic-Crisis/.

See DFID (2011). Explicit design features help to reduce spending as time passes. In South Africa, where social grants ‘consume’ about 3 per cent of GDP, amounts are varied to manage the fiscal costs. Demographic changes in South Africa will increase spending on pensions, but this may be moderated by raising the age of eligibility, while spending on child grants may decline because of reduced number of children.

The exercise has been carried out as a pilot for a limited number of low income countries from Asia and Africa, where data were available. The seven African countries were Burkina Faso, Cameroon, Ethiopia, Guinea, Kenya, Senegal, and Tanzania.

Namely, Ghana, Equatorial Guinea, Mali, Republic of Congo and Senegal, whose fiscal budgets have been analysed by ODI and UNICEF (2009).

During the first phase of implementation, however, targeting was badly conducted, resulting in many ‘extremely poor’ households being excluded (Devereux and Ndejuro 2009). See also Hartwig (2010).

As described by Chopra's contribution in the European report on development 2010 (EUI 2010, pp. 60–61).

In Lesotho the old-age grant is administrated by the Ministry of Finance and Developing Planning, and a special independent unit has sole responsibility of ruling it. Generally, as in Namibia, social pensions are managed by social ministries (Pelham 2007).

Source: the government information website, http://www.nhis.gov.gh/?CategoryID=309 [Accessed 11 November 2011].

This result is confirmed by Asante and Aikins (2009), who also find that limited information about the programme is one of the main factors affecting uptake, especially in rural areas.

Data on poverty are from the 2006 Ghana living standard survey. Another selection bias emerging from empirical analysis is the significant disparity between educated and non-educated people, with the former more likely to enrol (USAID 2009, Mensah et al. 2010).

Premiums have been computed by dividing national annual user fees by total population (a figure of about $4 per capita per year). In addition, given that the number of exempt to non-exempt was in the order of 1:1 the figure was doubled to achieve $8 (Rajkotia 2007).

The Inter-American Social Protection Network is one example of initiatives that promote lesson learning (www.socialprotectionnet.org). The International Policy Centre for Inclusive Growth (www.ipc-undp.org) has a focus on articulating lessons from social protection programmes. In 2008 the British government proposed a similar initiative (Lloyd-Sherlock and Barrientos 2009).

Gough (2008) suggests a framework to analyse social policy with five pillars: industrialisation (or structural change), interests, institutions, ideas, and international influence.

The classic study on the welfare state by Mommsen (1981) was an explicit attempt to compare German and British experience.