Rates of Return of Social Protection The case for non-contributory social transfers in Cambodia Franziska Gassmann Arusha, Tanzania – 17 December 2014.

Rates of Return of Social Protection

The case for non-contributory social transfers in CambodiaFranziska Gassmann

Arusha, Tanzania – 17 December 2014

Maastricht Graduate School of Governance

The usual arguments for extending social protection rely on…

• Human rights • Empirical evidence on impacts• Pilot projects• Affordability studies


Rationale for investment case

• Additional arguments are needed to move SP up the national development agendas– Demonstrate value for money– Analyze fiscal sustainability– Prove cost-effectiveness, capture multidimensional effects– Compare with alternative investments

• Develop economic argument for social protection– Costs AND benefits– Short term AND long term– Direct AND indirect


Non-contributory social protection as economic investment

• Paradigm shift: SP not just as a cost for the economy – Source of resilience in tough times– Support for growth and productivity in good times– Mechanism for social inclusion

• SP and economic growth – transmission channels– Building and protecting human capital, child wellbeing– Fostering productive investments, protect assets– Reducing liquidity constrains– Enhancing community assets, infrastructure– Stabilizer of aggregate demand, improving social cohesion,

making reforms feasible


Background• Cambodia has achieved rapid economic development and

poverty reduction.• Annual GDP growth of 7.7% on average (1995-2011).• GNI per capita US$): $830 (2011) (low income country)• Human Development Index: 0.523 in 2011 (#139)

• NSPS launched at the end of 2011.• Double objective: economic stability and human development

• Objective of our study:• Contribute to the evidence on the links between social protection

investments and socio-economic development in Cambodia• “What are the economic returns of social protection in the mid-and

long term?”

• Supported by UNICEF and Royal Government of Cambodia


Data and methodology

• Rates of Return: relation between net benefits and costs

• Data: CSES (2004 and 2009)• Microsimulation (steps):

• Static (cost-effectiveness): changes on poverty and inequality (direct distributional effect)

• Returns of human capital (education) at the household level• Behavioural (income) effects

• School attendance (education)• Nutrition (health)• Labour (participation and supply)

• Dynamic: 20 periods


RoR study Cambodia – Model

Household consumption

Poverty and inequality

Education(school attendance)

Health(underweight)

Labour participation

Human capital

Labour productivity

Social protection

Economic performance

Return

Direct (distributional) effects

Behavioural (income) effects


Returns to human capital

• An additional year of education is related with a 4.1% higher wage.• Low in international comparison mainly low-

skilled employment with limited productivity

• An additional year of education is related with a 1.8% higher household consumption for a poor rural household.


Behavioral (income) effects

• A 10% increase in the level of consumption is related with a 5.6 percentage points higher probability to attend lower secondary school for a poor rural person

• 10% increase in consumption is related with 0.4 percentage points lower probability of being underweighted.

• A 10% increase in household consumption is related with a 7.8 percentage points higher probability of formal work for poor persons between 18 and 64 years old in rural areas.


Behavioral (income) effects

Model: 2SLS for all households

Dependent variable (independent variable) Urban(non-poor)

Urban (poor)

Rural(non-poor)

Rural(poor)

log of household consumption per capita (maximum level of education within the household)

0.042*** 0.016** 0.026*** 0.018***(0.005) (0.007) (0.003) (0.002)

Model: Probit model for individuals 6-25 years old

Dependent variable (independent variable)Education level (rural-poor)

Primary Lower secondary

Upper secondary

School attendance (log of household consumption per capita)

0.226** 0.560** 0.373(0.089) (0.262) (0.516)

Model: Probit model for children under 5 years old

Dependent variable (independent variable) National National (poor)

Rural(poor)

Underweight (log of household consumption per capita)-0.043*** -0.048 -0.038

(0.015) (0.038) (0.041)

Underweight (no toilet facility in the house = 1)0.026 0.062** 0.081***

(0.016) (0.029) (0.029)


RoR study Cambodia – Policy

Social Protection Instrument

Target Population Benefit

Transfer Total Cost

KHR billion

% of GDP

KHR billion

% of GDP

Cash transferPoor children 0-6 years old in rural areas, up to two per household

USD 12 per month (60% rural food poverty line) 391 0.9 430 1.0

Social pension

Poor persons 65+ in rural areas

USD 20 per month (100% rural food poverty line) 139 0.3 153 0.4

Scholarship Poor children at lower secondary in rural areas

USD 50 per year (20% rural food poverty line) 25 0.1 28 0.1

Public works

Poor persons 18-64 years old in rural areas, up to 1 per household (80 days per year)

USD 2.3 per day 50 0.1 75 0.2

• Total costs is around 1.6% of GDP (USD 166 million, 2009).


RoR study Cambodia – Results

Benefit Scenario Period 1 Period 5 Period 10 Period 15 Period 20

Average years of education (18-64 years old)

With social protection 6.52 7.67 9.00 10.40 11.62

Without social protection 6.52 7.65 8.89 10.22 11.41

Benefit (difference) 0.00 0.02 0.11 0.19 0.21

Total household consumption average annual

growth rate (%)



Benefit (difference) 1.55 0.26 0.12 0.07 0.04

Poverty headcount (%)



Benefit (difference) -6.0 -6.0 -4.2 -3.8 -2.6

Inequality (Gini of consumption)



Benefit (difference) -0.016 -0.014 -0.013 -0.012 -0.010

Cost Policy Period 1 Period 5 Period 10 Period 15 Period 20

Cost (% of GDP) Social protection package 1.6 1.4 1.2 0.9 0.8

RoR Discount rate Period 1 Period 5 Period 10 Period 15 Period 20

Rate of Return (Absolute benefit on total

household consumption / absolute cost) (%)

2% -11.6 -10.0 -4.1 5.8 14.7

3% -11.6 -10.1 -4.3 5.0 13.3

4% -11.6 -10.1 -4.6 4.3 11.9


Results: benefit (human capital)• Labour force’s median education level

increases faster due to social transfers.Labour force median education level (schooling)

8.5

6.0

8.6

0.08

-0.0200.0000.0200.0400.0600.0800.1000.1200.1400.1600.180

6.0

6.4

6.8

7.2

7.6

8.0

8.4

8.8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Without SPI With SPI Bs (schooling - years)

Years of education Difference (Bs)


Results: Benefit (hh consumption)• Total household consumption grows faster if

SPI are implemented.Total household consumption average growth rate

2.67

1.55

2.71

0.04

-0.02

0.18

0.38

0.58

0.78

0.98

1.18

1.38

1.58

1.78

0.00

0.50

1.00

1.50

2.00

2.50

3.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Without SPI With SPI Bc (hh consumption growth rate)

Household Consumption average growth rate Difference (Bc)


Results: Benefit (poverty reduction)

• Poverty headcount decreases faster because of SPI Head count

29.7

25.5

19.8

14.7

10.4

23.720.7

15.6

10.97.8

-6.0

-5.5

-5.0

-4.5

-4.0

-3.5

-3.0

-2.50.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Without SPI With SPI Bp (poverty head count)

Poverty (head count) Difference (Bp)


Results

• Total cost of SPI decreases over time.

684

327

1.6

0.8

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

300

350

400

450

500

550

600

650

700

750

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Total cost (KHR billion) Total cost (% of GDP)

KHR (billion) % of GDP


Rates of Return – Results

14.711.9

-15.0

-10.0

-5.0

0.0

5.0

10.0

15.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

RoRC (d=2%) RoRC (d=4%) RoRC (d=6%)

Period

• A basic package of SPI for poor rural individuals in Cambodia has a RoR of between 12% and 15%, after 20 periods (years). It becomes positive after 12 periods (years).


RoR study Cambodia – Final remarks

• Dynamic microsimulation provides a novel approach to analyse economic returns of social protection.

• Modelling options depends on data constraints.• Costs and specific impacts have been estimated. The model can be used to

analyze potential economic returns in the mid- and long- term.

• Any model is always a simplification of real life.• Effects, benefits and returns may be higher if complementary

policies are also implemented.• Improving health and education coverage and quality.• Enhancing sanitation conditions.• Fostering economic productivity, formal labour market, industrialization,

innovation and technical change.


RoR study Cambodia – Final remarks

• Additional effects may increase benefits and RoR. • Behavioral (non-economic) effects due to SPI design.• Spillover effects and regional multiplier.• Institutional change and social cohesion.• Health status improvements (e.g. nutrition).

• Financing aspects (taxation), administrative issues (inefficiency) and targeting errors may reduce RoR.

• Specific SPI design (e.g. targeting, conditionality, payment mechanism) may affect RoR.

Rates of Return of Social Protection The case for non-contributory social transfers in Cambodia Franziska Gassmann Arusha, Tanzania – 17 December 2014.

Documents

social protection investments

social protection costs

social cohesion

human capital education

indirect slide

social inclusion sp

human development objective

rapid economic development