Get inspired by SYMBA Project: promoting Industrial Symbiosis
PEN: How global-comparative data challenges conventional wisdom
1. The PEN Team
Knowledge frontiers on poverty-forest linkages
PROFOR, World Bank, 3rd October 2014
PEN: How global-comparative data
challenges conventional wisdom
2. THINKING beyond the canopy
PEN is…
Large, tropics-wide collection of detailed & high-quality &
comparable data by PhD students on the poverty-forest
(environment) nexus, coordinated by CIFOR, with
numerous partners
It is the most comprehensive analysis of poverty-forest
linkages undertaken to date
3. Features of PEN
Approach: a network
• PhD students: Long fieldwork
& student engagement
• Supported by senior resource
persons
• Mutual benefits
Capacity building
• Majority of partners from
developing countries
State-of-the-art methods
• Quality data – short recall
• Comparable methods
• Methods summarised in a
2011 book
4. THINKING beyond the canopy
PEN: the numbers..
24 countries
38 PEN studies
239 households in the average study
364 villages or communities surveyed
>8,000 households surveyed
40,950 household visits by PEN enumerators
2,313 data fields (variables) in the average study
294,150 questionnaire pages filled out and entered
456,546 data cells (numbers) in the average study
17,348,734 data cells in the PEN global data base!
6. PEN sample: a delicate balance
Criteria for site selection:
• Within a tropical or sub-tropical developing region,
• Some access to forests (0 < forest cover < 100%)
Site selection was opportunistic (PhD students) – with
some posterior gap-filling (e.g. West Africa, Vietnam)
Within sites: stratified village selection (along pre-
defined gradients), random HH selection in villages
Broadly representative of smallholder-dominated
tropical and sub-tropical landscapes with moderate-to-
good access to forest resources.
Probably a slight bias toward areas with “good forests”
(vis-á-vis “rural developing world” baseline)
7. PEN methods and approaches
Research tools are online:
http://www.cifor.org/pen
Summarized in Angelsen et al.
(eds.) 2011 book (available for
free download from CIFOR
website)
PEN prototype questionnaires
available in eight languages
ESRC (UK) positive evaluation
2012: “methods and capacity
building may be more
important than actual PEN
results”
8. 0.0 0.1 0.2 0.3
income shares
other
environment
business
livestock
wages
forest
crops
Source: CIFOR-PEN dataset
Income sources in the PEN dataset
~22%
~6.4%
T = 27.5%
=> Clearly supports “high env. income” hypothesis
– …much more than some of us had thought!
9. Income share from forests and environment
(not country representative!)
sample mean=0.28
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Cameroon
Zambia
Nigeria
Bolivia
Brazil
Congo, Dem. Rep.
Mozambique
Peru
Cambodia
China
Ethiopia
Uganda
Ghana
Bangladesh
Burkina Faso
Malawi
India
Belize
Ecuador
Indonesia
Senegal
Guatemala
Vietnam
Nepal
forest and environmental income other sources
10. Forest and environmental income
shares by product type
Forest income (%) Other environmental
income (%)
Food 20.9 39.6
Fuel 37.2 21.8
Structural and
fiber
33.3 15.5
Medicines,
resins and dyes
5.1 5.8
Other 3.5 17.3 (fodder)
Total 100 100
11. 15.3 7.5
13.0 8.3
11.5 8.9
10.0 9.5
9.1 9.9
0 5 10 15 20 25
Forest income share (%)
Bottom 20%
20-40%
40-60%
60-80%
Top 20%
Forest reliance by income quintile, global
Subsistence Cash
Forest and environmental income
shares by wealth status
12. • Gender: Men generated at least as
much forest income as women do
(with product variations)
• Shocks: Forests less important as
“safety nets” and “gap fillers” than
portrayed in case-study literature
(other priority responses key)
• Tenure: more income from state
than private or community forests
(absolutely and per-ha)
• Deforestation: The poorest farmers
(“they cut because they must”) clear
much less forest than smallholder
middle class (investment motives)
Other myth-busting findings
15. 1. Analysis of fuelwood & charcoal in rural livelihoods (Univ. N. Carolina)
2. Migration, land-cover & climate change (Wageningen).
3. Local tenure institutions and forest benefit sharing (Univ. of Colorado
at Boulder).
4. Local tenure institutions, forest product types, and income share
(e.g. poor, ethnic minorities, women etc.) (University of Alberta).
6. Theoretical model on safety-net uses of NTFPs (Columbia Univ.).
7. What share of hh nutrition comes from forest foods at PEN sites?
(CIFOR)
8. Relative importance of forest and non-forest food supply (Univ.
Copenhagen).
9. Safety net response strategies to different types of shocks (ZEF-
Bonn)
10. WB 2015 report: Climate Change & Poverty – CIFOR analysis PEN &
climate data (CIFOR)
11. How useful are environmental service models for decision making?
(University of Southampton).
Further analyses: CIFOR & partners
16. Not without a big startup
grant!
Sentinel sites discussion:
larger scale, more
extensive interest field
(mismatch).
“Would you do PEN again?”
Angelica Almeyda, PEN
PhD student: “Yes, but
only in the past!”
Gathering more PEN (panel) data?
17. Outcomes & impact pathways
• Help World Bank and
statistical bureaus engaged
in Living Standard
Measurement Surveys
(LSMS) to do a better job in
(environmental) ‘bean
counting’…
• Joint project working with
FAO, World Bank, PROFOR,
IFRI, Univ. Copenhagen:
develop forestry module for
LSMS.
• CIFOR pilot testing new
forestry module in Indonesia
(+Tanzania), Oct 2014
18. Environmental income is
key for rural small-holders
in developing world is
central claim: what policy
impact?
But what is PEN non-
random sample really
representative of? So
much tropics-wide variation
Are PEN results arguments
for conservation?
Perhaps, but not if
forestland is abundant
Final perspectives
19. The role of extractive incomes
“More than 10,000 years
after the agricultural
revolution, millions of
rural smallholders across
the developing world may
still derive as much
income from foraging
forests and wildlands as
from cultivating crops”
(Wunder, et al.World Development 2014)
Editor's Notes
Approach: a network
Collaborative effort between CIFOR & 40+ universities & other institutions
PhD students: Long fieldwork & student enthusiasm
Capacity building
Most partners from developing countries
The standardised definitions, questionnaires and methods used by PEN Partners across multiple sites represent a new approach to data collection, and enabled direct comparison between sites and the subsequent integration of case studies into a compatible global data set. Our methodological advances have already influenced a number of high profile research projects (e.g. CIFOR’s REDD multimillion GBP global-comparative project) and some significant donors/actors (e.g. the Finnish Ministry of Foreign Affairs Forestry program).
This is the global average results.
Forest (natural) income = 21.1%
Forest (plantation) income = 1%
Non-forest environmental income = 6.4
Environmental income (natural forest & non-forest environmental) = 27.5, about the same as crop income 28.7%
The poor are more reliant on forest income, env income, and subsistence income
These findings (all published in the WD Special issue) confirmed some suspicions & challenged some conventional wisdom about environmental income.
The world’s rural poor are more dependent on forest & environmental resources than is commonly realized.
Income from natural forests & other natural areas ~28% of total household income, nearly as much as crops.
Men generated at least as much income from forests as women do, contradicting long-held assumptions.
Forests were less important as “safety nets” in response to shocks and as gap fillers between seasonal harvests than previously believed.
State forests generated more income than private or community forests.
While the most destitute of poor farmers are often blamed for deforestation, they played only a modest role in forest clearing.
The complete PEN dataset has been shared with
7 PEN articles: 1 summary/overview article (Wunder et al. 2014), 5 thematic global articles, 1 regional case study (Duchelle et al. 2014),
6
One of the key knowledge products generated from the PEN project is the PEN global dataset, which consists of data from ~8,000 households from 333 villages in 58 sites, in 24 countries. As part of CIFOR’s open-access policy and mandate to share global public goods, this dataset is now being made available to share with others from outside of the PEN project. The data has thus far been shared with a total of 10 ‘external’ parties, who are currently analyzing the data with the aim of producing a range of publications that will acknowledge the use of PEN data. Promotion of the dataset’s availability has currently been limited to ‘word-of-mouth’, however it is expected that in the near future its availability will be promoted more broadly, and even made available for public download from our website.
The complete PEN dataset has been shared with
Largely funded by Difid, although it involved very many partners.