Fractal poverty traps:

1
Fractal poverty traps implications for poverty
reduction strategies in Kenya
Brent Swallow Theme Leader for Environmental
Services ICRAF, Nairobi
2

• Outline
• Background (PRSP, Kenya)
• Research questions
• Conceptual approach
• Fractal poverty traps
• Data and empirical approach
• Results
• Conclusions

3

• Background
• 1. Poverty Reduction Strategy Papers (PRSP)
• IMF / WB requirement for debt relief under Highly
  Indebted Poor Country Initiative (HIPC)
• Detailed WB guidelines eg must be country-owned
  and involve broad consultation (www.worldbank.org/
  poverty/strategies/overview.htm)
• Bilateral donors see PRSPs as setting national
  priorities for public investment and services for
  poverty reduction (EU)
• Hey what about me?

4

• Kenya and the PRSP process
• National PRSP process Min. of Finance developed
  I-PRSP in 2000 using micro data from WMS (94, 97)
  addressing macro reform agenda
• District PRSP consultations, sector and theme
  consultations (each with 150 1500 persons) in
  2001
• Finalized PRSP, implementation matrix, matrix of
  issues and priorities (www.treasury.go.ke)
• V. impt. in Kenya with new government in place
• Some data indicating village priorities

5

• Research questions
• connections or disconnections between poverty and
  resource conservation implied by priorities at
  village, district and national levels
• correspondence between priorities articulated by
  village groups, district consultations national
  PRSP documents
• clarity of the fractal and public dimension of
  investment and service priorities at different
  scales.

6
Conceptual approach poverty traps
Low returns
?
Low and/or declining asset stock
Sub- threshold levels of investment
Policy, technology, markets, institutions, fina
nce
?
?
?
chronic poverty
7
Fractal Poverty Traps

• poverty traps, like poverty outcomes, are
  multi-scale individual, household, village,
  regions within countries, nations,

• poverty traps have similar origins, effects and
  dynamics at all scales

• some things cause poverty traps at multiple
  scales, while others cause traps at specific
  scales

8
Implications of Fractal Poverty Traps

• Multi-scalar poverty reduction plans, with some
  differences in priority across scale
• Need to identify target factors that constrain
  asset accumulation by / for the poor at each
  level
• Aggregate traps will reduce effectiveness at
  lower scales eg Kenya government trap of
  servicing the debt payable to domestic creditors

Source Kiringai and Manda, 2002
9
Source Kiringai and Manda, 2002
10
Data and analytical approach

• National PRSP for Kenya
• Problem ranking
• Judgement call on nature of gt 500 solutions
• Budgetary allocation in implementation matrix

• District PRSP Consultations
• 8 districts in Western Kenya (high poverty
  areas, most of which are known to me and others
  here)
• Listing and ranking (if any) of problem areas
• Judgement call on nature of solutions

11
(No Transcript)
12
NALEP Focal Area Priorities National
Agriculture and Livestock Extension Programme
(NALEP) has a shifting Focal Area approach
(FA) FA priorities established in a community
meeting at end of 5-day PRA We compiled FA
priorities and action plans for 10 Focal Areas
from each of the 8 districts 113 village
priorities aggregated into 16 priority areas.
Priority ranks used to weight priorities from
with weights from 12 (highest priority) to 1
(12th priority if listed)

13
Result 1 Poverty resource conservation
Village level aggregated for 80 villages
Priority area Weight Low crop
production 751 Human diseases health
care 576 Livestock diseases
production 551 Lack of clean water 547 Soil
degradation 489 Inadequate agric.
inputs 335 Community infrastructure
services 332 Poverty, low income,
unemployment 290 Agricultural pests diseases
221 Lack of fuelwood 199 Poor product
marketing 218 Poor roads 194 Environmental
problems 158 Culture and community
norms 134 Insecurity 85 Community
organization 65
14
Result 1 Poverty resource conservation
Village level (Table 4, across all 80
villages) Lack of clean water -- one of top 6
priorities Soil degradation one of top 6
priorities Lack of fuelwood 10th to 15th
priority
District level Lack of clean water listed in
most districts Soil degradation listed in a few
districts Lack of fuelwood not listed
Deforestation listed in some districts
National level Lack of clean water small
mention, paltry budget Soil degradation barely
mentioned, no budget Lack of fuelwood barely
mentioned, no budget Deforestation environ.
protection mentioned
15
Result 2 Correlation of priorities within and
between scales

• 2.1 Correlation of priority weights among
  villages
• correlation coefficients for priority weights
  between villages within districts (10 villages /
  district)
• Siaya -0.200 to 0.763, 17 of 45 sign. gt 0
• Vihiga -0.409 to 0.846, 6 of 45 sign. gt 0
• Nyando -0.142 to 0.897, 14 of 45 sign. gt 0
• correlation coefficients for priority weights
  (summed across villages) between districts (8
  districts)
• 0.307 to 0.911, 25 of 28 sign. gt 0

16
2.2 Match between village priorities and
priorities from district consultations (Table 9)
Overall fairly good matches in some districts
(eg Bondo) and poor matches in other districts
(eg Nyando)
High priority in villages districts low
agricultural production, lack of clean water
High priority only in villages human health,
availability of agr. inputs, fuelwood, soil
erosion, agr. extension, social organization
norms
High priority only in district consultations
education, roads
17
2.3 Match between village and national PRSP
priorities
Match between village priorities and national
priorities quite strong
Match between national priorities and national
budget allocation very weak
of budget
Priority rank
13 7 61 1 2 6 11
1 2 3 4 5 6 7
Ag. rural development Human resource
dev. Physical infrastructure Tourism Trade
industry Public safety, law order Public
administration
18
Result 3. Clarity of fractal and public dimension
of investment and services
Village level Activities to be jointly
undertaken by village groups with support from
locally-based extension providers Most
activities produce private benefits, with
possible thresholds in transaction costs Some
activities produce community investments
(clinics, roads, springs) or address community
social problems
19
District level Activities to be financed and
undertaken by the government (no concern about
financing, sustainability, prioritization)
Focus on construction of infrastructure,
subsidization of agric. inputs and outputs, free
services
National level Construction of roads, market
reform pluralistic extension all by central
government 84 of resources spent in Nairobi
little connection apparent to needs of the poor
20
Conclusions (1/2)

• Kenyas budget imbalance and interest payments
  to domestic creditors cause a major macro poverty
  trap
• local residents do put priority on conservation
  of resources most important to their livelihoods
• need for caution in extrapolating priorities
  from village to nearby village, sample of 80 OK
  for clarifying regional priorities
• Poverty reduction priorities should be based on
  review and discussion of best evidence about
  poverty from multiple sources

21
Conclusions (2/2)

• Central government road construction and market
  reforms likely to be little direct benefit to the
  poorest rural Kenyans
• Need more consideration of multi-level dimension
  of problems and solutions
• Consultation, per se, even among 1500 people,
  may be of limited value unless people are somehow
  accountable for resources, implementation or
  sustainability
• Donors should re-consider tying aid so closely
  to PRSP documents