Estimation of rates of return (ror) on social protection investments in Lesotho

Type Working Paper
Title Estimation of rates of return (ror) on social protection investments in Lesotho
Author(s)
Publication (Day/Month/Year) 2016
URL http://www.merit.unu.edu/publications/uploads/1461167192.pdf
Abstract
Compared to other low-income countries, Lesotho is one of the leaders in social protection. It is at the
forefront of moving towards a social protection systems approach in Sub-Sahara Africa and beyond.
The National Social Protection Strategy (NSPS) 2014/15-2018/19 (Government of the Kingdom of
Lesotho, 2015) represents the Government’s vision and ambitions for the coming years to address the
risks and challenges over the life-course to protect its citizens and particularly the poor and vulnerable
Basotho. The implementation of the envisaged core social protection programs for children, the
elderly and poor adults, supplemented by complementary programs in other sectors has the potential
to significantly reduce the extent and depth of poverty and provide citizens with the means to improve
their livelihoods in the short and long term.
It is estimated that the implementation of the core programs of the NSPS will cost four percent of GDP
at full coverage (Government of the Kingdom of Lesotho, 2015). Although the analysis has shown that
the strategy is affordable, four percent of GDP represents a considerable amount of national
resources. In order to garner continued political and financial support for the implementation of the
NSPS, it is essential to build strong economic arguments, proving that the investment is worthwhile in
terms of expected benefits in the future.
The aim of this study is to estimate the Rate of Return (RoR) on Social Protection Investments (SPI) in
Lesotho, thereby generating evidence to support the advocacy for social protection in Lesotho and
assisting relevant ministries in planning the allocations for SP instruments. The primary focus of the
study is the Child Grant Program (CGP). The CGP targets extremely poor households with children
aged between 3 and 17 years. The net benefits of the CGP are compared with the Old Age Pension
program (OAP), the school feeding program (SFP) and with a combined package of CGP and OAP.
Non-contributory social transfers directly affect household disposable income, and as such household
consumption. However, social transfers also affect household behavior through income and nonincome
effects. Additional and/or secure income encourages households to invest in health,
education, livelihoods and productive activities. The study thus builds on a framework assuming both
direct and indirect benefits through increased consumption due to social protection investments (i.e.
poverty reduction and human capital accumulation).
The methodology applied in this study consists of three main elements. First, a static simulation is
implemented, revealing the direct effects of the increase in household consumption on poverty and
inequality. Secondly, different empirical models are used to estimate the relationship between
household consumption and school attendance, school attainment and household consumption, and
vii
household consumption and labor market participation. Finally, a dynamic simulation model was
constructed in order to predict the effects of social transfers over a period of 15 years. The simulation
procedure remains the same in each period. Eligible beneficiaries of the respective SP instrument
receive the benefits, which increases their consumption levels with 80% of the transfer values. Based
on the new consumption level, the likelihood of school age children to attend school is predicted.
Subsequently, the educational attainment is updated depending on whether children attended
school. The new consumption levels are calculated as the sum of the previous consumption level plus
the direct effect (transfers) and the behavioral benefit. Fertility and mortality rates are integrated into
the simulation model in order to reflect demographic changes over time. The dynamic simulation
compares the outcomes of the programs to a scenario without SPI. Therefore, the focus is not on
predictions of outcome variables in future periods, but rather on the relative development in
outcomes compared to the control scenario. The Rate of Return compares the net present value of
benefits of an intervention to the net present value of the costs of this intervention.
The analysis is based on the nationally representative Household Budget Survey 2002/2003. Further,
data on demographic projections was obtained from the Bureau of Statistics of Lesotho and from the
World Health Organization to simulate demographic developments over the time period of the
analysis.
With respect to the direct effects on poverty and inequality, the largest reduction in poverty
headcount can be observed for OAP followed by SFP and CGP. For the CGP no immediate effect on
the poverty headcount can be measured. This is due to its focus on extremely poor households that
do not graduate out of poverty with the transfer as the gap is too large. However, the potential effects
of the CGP become evident when considering extreme poverty outcomes. The simulation suggests
that CGP reduces the extreme poverty headcount by 20% compared to a scenario without SPI. The
results of the static simulation also suggest that the CGP is the most cost-effective program as it would
generate the largest reductions in outcomes for each percent of GDP invested. According to the static
simulation, investing 1% of GDP in the CGP would generate a reduction of 15.3% in extreme poverty,
which is about five times the size of the OAP program effect at the same cost.
Based on the empirical models, which are based on the situation in 2002, the estimates imply positive
returns to education of an additional year of schooling of 9%, which is close to the international
standard of approximately 10% (Psacharopoulos and Patrinos 2004). The effect of household
consumption on school attendance is also positive. At the national level, a 10 percent increase in
household consumption is associated with a 1.1 percentage point higher probability of a child
attending school. The findings suggest that household consumption positively affects school
viii
attendance rates in Lesotho. This implies that SP instruments that increase household consumption
levels likely improve education outcomes and therefore contribute to human capital development in
Lesotho. The results further show that household consumption has a positive effect on labor market
participation. At national level, a 10 percent increase in household consumption level is associated
with a 1.6 percentage point increased probability of labor participation for individuals aged 18 to 69.
Overall, the findings suggest that SPI that increase household consumption levels (income) potentially
raise participation in the labor markets in Lesotho.
The dynamic simulation model is applied to examine the effects over time including the behavioral
effects through increased school attendance and higher school attainments. Thus, the SP effects are
simulated over a 15 years’ time range. All three programs affect school attendance and educational
attainments positively. School attendance rates of individuals between 6 and 24 years increased
strongest for the CGP scenario and the combination of CGP and OAP. The CGP school attendance rate
increased by 5% in the first period, which grows up to an annual increase of more than 12% in period
15 compared to the control scenario. As the SP effects sum up over time, an exponential growth in
school attendance rates can be observed. As a consequence, after 15 periods the working-age adults
dispose of a 2% higher school attainment in the CGP scenario as compared to the control scenario.
The OAP effect on school attendance is markedly smaller. Despite the larger transfer values of OAP
the effect is lower as it is not specifically targeted at children. The combination of CGP and OAP further
increases school attendance rates, however, adding only little to the CGP effect.
The effect of the SFP on school attendance is smaller than the CGP effect and increases up to around
7% at the end of the simulation period. Yet, the annual growth rates are smaller than for the other
programs. This is due to the SFP assignment to children that are already enrolled in school with thus
little scope to further increase attendance rates. However, it has to be noted that the potential effects
of SFP on aspects such as school performance or health cannot be regarded in the model. These effect
pathways could have important impacts on school attainments and may result in underestimation of
the educational effects of SFP in the simulation model.
The analysis of the returns to education suggests that an additional year of schooling increased
consumption levels on average by 9% in 2003. At the same time results of the dynamic simulation
model show that CGP increased the number of years of schooling on average by 2% per year. This
highlights the potential of SPI to generate large returns in future periods. However, the results also
showed that school attainments tend to be low in Lesotho and that the education effects especially
on the extremely poor need more time to unfold their full returns.
ix
This results in initially negative RoR, which slowly start to improve and turn positive after 10 periods
for the CGP. The simulation results suggest that from period 10 onwards the net benefits of the CGP
exceed the costs. The RoR of OAP and SFP remain negative throughout the simulation, but show a
positive trend. This finding is related to the fact that both programs have universal targeting
mechanisms and do not particularly benefit the extremely poor. Secondly, beneficiaries either already
attend school (SFP) or left school age long ago (OAP). Thus, their scope to generate returns through
school attendance is much lower compared to CGP resulting in lower behavioral benefits.
The findings suggest large program effects on poverty and inequality outcomes. Simulating the CGP
on the national level reduced extreme poverty by more than 20% per year and reduced inequality by
up to 7%. This indicates the potential of CGP for poverty reductions in Lesotho. Taking all future
returnsinto account, the educational benefits exceed all cost including transfers and operational costs
after 10 periods. This underpins the power of SPI for educational but also welfare developments in
Lesotho. On top of that, additional returns through health and agricultural investments and increasing
tax revenues are not considered in this study. Therefore, the results might only reflect a lower bound
estimate of the full potential of social cash transfers.
As a model can never cover the entire set of SPI linkages, it needs to be born in mind that simulation
models are always a simplification of reality. The study has a number of particular limitations that
need to be born in mind when interpreting the results. Due to the limitations of the HBS 2002/2003
data, not all potential indirect benefits of social transfers could be incorporated in the model. Effects
through improved health or investments in productive activities are not considered, which may be
particularly important for the OAP. Therefore, the resulting rates of return are likely an
underestimation of the actual achievements. Furthermore, the economic and social situation in
Lesotho has changed considerably since 2002/2003. For example, school attendance and highest
education achievements have increased considerably over the last decade. Nonetheless the models
show how specific aspects of SPI pathways generate monetary returns over the long term.
It is recommend to repeat the present analysis once more comprehensive and more recent household
survey data are available. Particularly the inclusion of other transmission channels next to education
would add value and provide additional insights in the potential benefits and the respective RoR in
the long term. Furthermore, information such as access to services and infrastructure would allow a
more detailed analysis of the returns of SPI which goes beyond the national average and provide
insights into policy areas that need to be strengthened in order to maximize the impact of SPI. The
BOS is keen to improve their data collection and adjust the survey instruments such that they better
serve the overall needs for regular analysis and evaluation of social protection policies.

Related studies

»