THE CHALLENGE OF

POVERTY REDUCTION

IN AFGHANISTAN

ARTF Strategy Group Meeting

Kabul, Nov 5th

Silvia Redaelli and Pedro Olinto

Monitoring Socio-Economic Outcomes

in Afghanistan – The NRVA

Significant progress over time in institutionalizing a monitoring system for socio-economic outcomes (including poverty!) in Afghanistan.

SOURCE OF DATA: • NRVA (National Risk and Vulnerability Assessment);

• 2 rounds of data currently available: 2007-08 and 2011-12

METHODOLOGY: • Ensure comparability over time (data, instruments and definition of indicators)

• Security and data quality issues – two provinces excluded from analysis of poverty and food security (Helmand and Khost) due to data quality problems in 2011-12

OUTPUT: • Compilation of indicators at the National and subnational level.

Development Trends:

Key Facts (1)

Significant progress in human development outcomes

(health and education), access to services and

infrastructure

27.6

50.3

67.2

32.6

55.8

68.3

Literacy rate - 14+(%)

Net attendancerate, primary school

(%)

Girls to boys netattendance rate,

primary (%)

Education and Gender

2007/08 2011/12

30.4 26.6

41.7

51.3 45.5

69.1

Access to skilledantenatal care (%)

Safe drinking water(%)

Electricity (%)

Heath and Access to Services

2007/08 2011/12

Development Trends:

Key Facts (2)

Lack of significant progress in poverty reduction, increase

in inequality and deterioration in food security.

29.3

18.5 20.9

34.7

25.3

19.9

Calorie deficiency (%) Protein deficiency (%) Inadequate dietarydiversity (%)

Food Security

2007/08 2011/12

36.3

7.9

29.7

35.8

8.4

31.6

Poverty rate (%) Poverty gap (%) Gini index (%)

Poverty and Inequality

2007/08 2011/12

Poverty Trends in Afghanistan

Survey year 95% CI

2011-12 2007-08 2011-12 2007-08

Poverty headcount 35.77 36.27 [34.14, 37.40] [34.94, 37.60]

Poverty Gap 8.44 7.86 [7.95, 8.94] [7.45, 8.27]

Squared Poverty Gap

2.91 2.55 [2.69, 3.13] [2.37, 2.73]

NO evidence of poverty reduction in Afghanistan “despite”

- High economic growth ~ 7% annual

- Substantial inflow of aid to the economy ~ 10 billion annual (civilian and military)

i. Poverty Headcount & Poverty Gap unchanged

ii. Significant increase in Poverty Severity

Scope of the poverty challenge

Survey year

2011-12 2007-08

Number of poor 9.1 million 8.5 million

Poverty Deficit(*) 5.5% GDP** 5.8% GDP**

The “cost” of eliminating poverty (if perfect targeting was possible) is about 6 % of GDP;

- Same order of magnitude than other Low income countries (Africa ~5%)

- …but

- High relative to the country’s revenue capacity

Number of Afghans living below the poverty line has increased at an annual rate of 1.6 percent Poverty Deficit, i.e. the additional annual income that would, in principle, bring every poor person exactly up to the poverty line

(*) Annual,. expressed in current prices, Afs. (**) Non Opium GDP

Poverty in Afghanistan:

Key Facts (1) Large and persistent urban-rural disparities:

• Rural areas: higher poverty rates, lower level of wellbeing (pc

consumption), lower level of inequality (gini)

• Trends: no change in urban/rural poverty rates, but increase in the number

of poor living in urban areas (urbanization; 18% increase in urban pop)

• Profile 2011: 4 out of 5 poor people live in rural areas (rural areas

account for 82% of the poor and 76% of the population)

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

45.0

Urban Rural Total

Poverty Headcount, by year – area of residence

2007-08 2011-12

Poverty in Afghanistan:

Key Facts (2) • Large disparities at the subnational level:

• Between Regions: North East, West Central and East – lagging regions: high poverty rates and low levels of wellbeing

• Between Provinces: Poverty concentration the highest in more urbanized/densely populated provinces (Kabul, Nangarhar, Herat) and in the most populated amongst poorest (Takhar, Badakhshan).

ZabulLaghman

Ghor

Takhar

Urozgan

Kunarha

Badakhshan

Sar-I-Pul

BamyanBadghisNangarhar

JawzjanSamangan

DaykundiKunduz

PaktikaBaghlan

Wardak

Paktya

Ghazni

Nooristan

Kandahar

PanjsherHerat

Nimroz

Faryab

KapisaLogar

Farah

Parwan

BalkhKabul

02

04

06

08

0

Hea

dco

un

t P

ove

rty R

ate

1500 2000 2500 3000 3500 4000Average Per Capita Consumption

Note: Area of symbol proportional to the size of poor population

Source: NRVA 2011-12

Central

South

East

Northeast

North

West

Southwest

WestCentral

25

30

35

40

45

50

% p

op

ula

tio

n th

at is

poo

r

1500 2000 2500 3000Median per capita expenditure

Note: Area of symbol proportional to the size of poor population

Source: NRVA 2011-12

# of poor % of poor

Kabul 1,033,730 11.37

Nangarhar 663,455 7.3

Herat 660,397 7.26

Takhar 610,028 6.71

Badakhshan 588,060 6.47

Why Growth did not contribute to poverty

reduction?

Because growth had an “un-equalizing effect” – had

growth been equally distributed, poverty would have

declined by 4.4 percentage points.

Lack of pro-poor growth No “Shared Prosperity”

-20

24

6

Gro

wth

rate

of p

er

cap

ita c

onsum

ption

0 20 40 60 80 100Cumulative population share

Upper 95% confidence bound/Lower 95% confidence bound

Growth rate in mean

Growth rate of the

poor was negative

(-0.17).

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

POOREST 2ND POOREST

MIDDLE 2ND RICHEST

RICHEST TOTAL

Bottom 40%

Annualized growth rate pc consumption, by quintiles

Why hasn't growth reduced poverty?

3 hypotheses:

i. Was growth skill biased? Poor people do not have human capital endowments – an increase in returns to education over time could explain lack of poverty reduction and widening inequalities. NO

ii. Did growth increase urban/rural gap? Poverty rates are higher in rural areas where most of population lives – an “urban” bias in growth could explain lack of poverty reduction and widening inequalities. NO

iii. Did growth increase regional disparities? Poverty has a strong spatial connotation – poorest regions lagging behind/not growing could explain lack of poverty reduction and widening inequalities. YES

Lagging regions, Poverty and Growth

Growth in Afghanistan is highly volatile and vulnerable to high

incidence of shocks (agriculture, security). As poor people are more

vulnerable to shocks, lagging regions (NorthEast, WestCentral,

East)– having the highest % of poor – are likely to be “left behind”

1. Annual growth rate lagging regions -1.11% (1.97% rest of country)

2. Incidence of Shocks: 2011/12 was a drought year, security incidents

increased over time

3. Lagging regions seem most affected by the increase in shock incidence

between the two survey years

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

2007 2008 2009 2010 2011 2012

# Afghans killed or injured (civilian, military & AGE)

0

0.1

0.2

0.3

0.4

0.5

0.6

2007 2011 2007 2011

Agriculture Security

Leading regions Lagging regions

% population affected by shock

Lagging regions: the North East

• Amongst the lagging regions, North East seems the region in which households are less able to cope with shocks: main evidence - • Average pc consumption dropped (no consumption smoothing)

• Poverty increased from 36.3 to 49.7 while no change in poverty in East and West Central.

Hypotheses to be further explored: role of geography and aid in explaining widening regional inequalities.

• NE less connected than other lagging regions to the rest of the country

• NE households are less resilient to shocks

• NE receiving less aid than other regions

Towards a poverty reduction strategy

What do we know:

1. Lack of poverty reduction despite growth is due to widening inequalities across regions.

2. Lagging regions are not growing and poverty increased in the NE

3. High shocks incidence and volatility of growth

What needs further analysis:

1. Role of aid in explaining poverty trends and regional disparities (Data constraints) • Main question: what would have been the evolution of poverty over time without

aid?

2. Role of shocks vs geography in explaining existence of “lagging regions” • Main question: are some regions lagging behind because of transitory shocks or

structural differences.