The African Millennium Villages.
ABSTRACT We describe the concept, strategy, and initial results of the Millennium Villages Project and implications regarding sustainability and scalability. Our underlying hypothesis is that the interacting crises of agriculture, health, and infrastructure in rural Africa can be overcome through targeted public-sector investments to raise rural productivity and, thereby, to increased private-sector saving and investments. This is carried out by empowering impoverished communities with science-based interventions. Seventy-eight Millennium Villages have been initiated in 12 sites in 10 African countries, each representing a major agro-ecological zone. In early results, the research villages in Kenya, Ethiopia, and Malawi have reduced malaria prevalence, met caloric requirements, generated crop surpluses, enabled school feeding programs, and provided cash earnings for farm families.
- SourceAvailable from: cof.orst.edu[show abstract] [hide abstract]
ABSTRACT: To reach the Millennium Development Goal of halving hunger by 2015, the Millennium Project's Task Force on Hunger recommends seven mutually reinforcing actions: political action; national policy reforms; increased agricultural productivity for food-insecure farmers; improved nutrition for the chronically hungry; productive safety nets for the acutely hungry; improved rural incomes and markets; and restored natural resources essential for food security. The Task Force concludes that "it can be done"-the Hunger MDG can be attained. In this Policy Forum, the task force chairmen explain why achieving that goal will require unprecedented levels of effort, but it is well within the reach of our technical and financial capabilities.Science 02/2005; 307(5708):357-9. · 31.20 Impact Factor
- Science 04/2002; 295(5562):2019-20. · 31.20 Impact Factor
- The Lancet 01/2005; 365(9457):442-4. · 39.06 Impact Factor
The African Millennium Villages
Pedro Sanchez*†, Cheryl Palm*, Jeffrey Sachs*, Glenn Denning*, Rafael Flor*, Rebbie Harawa*, Bashir Jama‡,
Tsegazeab Kiflemariam*, Bronwen Konecky*, Raffaela Kozar*, Eliud Lelerai*, Alia Malik*, Vijay Modi*, Patrick Mutuo*,
Amadou Niang*, Herine Okoth*, Frank Place§, Sonia Ehrlich Sachs*, Amir Said*, David Siriri*,
Awash Teklehaimanot*, Karen Wang*, Justine Wangila*, and Colleen Zamba‡
*The Earth Institute, Columbia University, 405 Low Library, MC 4335, 535 West 116th Street, New York, NY 10027;‡United Nations Development
Programme, 1 UN Plaza, New York, NY 10017; and§World Agroforestry Centre, UN Avenue, PO Box 30677-00100 GPO, Nairobi, Kenya
Edited by Robert W. Kates, Brown University, Providence, RI, and approved August 23, 2007 (received for review January 17, 2007)
We describe the concept, strategy, and initial results of the Mil-
lennium Villages Project and implications regarding sustainability
and scalability. Our underlying hypothesis is that the interacting
productivity and, thereby, to increased private-sector saving and
investments. This is carried out by empowering impoverished
communities with science-based interventions. Seventy-eight Mil-
lennium Villages have been initiated in 12 sites in 10 African
countries, each representing a major agroecological zone. In early
results, the research villages in Kenya, Ethiopia, and Malawi have
reduced malaria prevalence, met caloric requirements, generated
crop surpluses, enabled school feeding programs, and provided
cash earnings for farm families.
agriculture ? environment ? hunger ? Millennium Development Goals ?
ways to achieve the MDGs (1). Although the MDGs have been
generally accepted as the world’s goals, there are questions and
controversies about how to attain them. sub-Saharan Africa is the
only region still severely off-track to reach the MDGs by 2015 (1).
Several biophysical and economic constraints impede sub-Saharan
Africa’s escape from extreme poverty, including extremely low
productivity of food production (2–4), heavy burden of infectious
disease (5, 6), and insufficient core infrastructure in water, roads,
power, and telecommunications (7, 8).
forth quantified and time-bound goals, the Millennium De-
The poverty trap in Africa results from the extreme shortage of
lives. A below-subsistence living standard means that survival
capital, a form of dissaving. Poverty prohibits African households
collateral and high transactions costs prohibit the finance of inputs
through market-based credits. Risk of drought adds to the barrier
of market-based financing of the poor. The result is that a high
(seeds, fertilizers), and soils become depleted of nutrients after
repeated crop cycles without sufficient nutrient replenishment (9).
There is little scope for net positive saving or environmental
Just as Africa’s food crisis is a reflection of the interaction of
burdens reflect an interaction of tropical climates, disease vectors
unique to Africa, and the lack of basic public health services (1).
Malaria and AIDS stand out as uniquely high burdens in Africa,
and both have run rampant in recent decades because of a lack of
adequate public health response, which in turn has been hampered
by the lack of adequate financing of public health measures.
Hunger and disease also interact with low food intake and
nutritional deficiencies, leading to reduced immune response. High
disease burdens (such as infection with soil-transmitted parasites)
result in reduced nutritional intake (2, 4). High child mortality
blocks the demographic transition to low fertility rates, and rapid
population growth and large families exacerbate poverty. Finally,
poverty also contributes to poor governance, which, in turn, exac-
erbates poverty. The result is a poverty trap in which poverty,
hunger, disease, rapid population growth, environmental degrada-
tion, and poor governance are all mutually reinforcing (10). The
‘‘Washington consensus’’ imposed by the Bretton Woods institu-
tions during the 1980s and 1990s did not address these underlying
factors and did not enable tropical sub-Saharan Africa to escape
from the poverty trap (1).
The underlying hypothesis of the United Nations Millennium
be overcome through targeted public-sector investments to raise
rural productivity, leading to increased private-sector savings and
holds and the community in several dimensions, the poverty trap
can be escaped. Several kinds of capital need to be increased:
natural (soil nutrients), infrastructure (roads, power, telecommu-
nications), human (skills and health), and financial (household
above a threshold level, above which the village can move toward
self-sustaining economic growth. Ideally, this would be carried out
on a large scale, involving an entire region or nation in capital
accumulation along these lines. The United Nations Millennium
Project recommended that such capital investments be made at an
shared by communities, governments, and donors (1). This ap-
proach contrasts with projects that emphasize macroeconomic
stability or incremental steps in a single sector.
Such public investments at scale initiate a positive dynamic of
saving and growth that supports private-sector investments in two
ways. First, household incomes rise above subsistence, so that
household-based capital accumulation and microfinance become
feasible. Second, the existence of good roads, power, and telecom-
V.M., A.N., F.P., S.E.S., and A.T. designed research; P.S., C.P., J.S., G.D., R.H., T.K., A.M., P.M.,
D.S., K.W., and J.W. analyzed data; and P.S., C.P., and J.S. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Abbreviations: ha, hectare; IRD, integrated rural development; MDG, Millennium Devel-
opment Goal; ODA, official development assistance; t, ton.
†To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2007 by The National Academy of Sciences of the USA
October 23, 2007 ?
vol. 104 ?
no. 43 ?
munications encourages the inflow of capital from outside inves-
the transformation of below-subsistence communities into com-
mercial farming communities that are able to save, invest, and
diversify into productive nonfarm work The public-sector invest-
ments are designed to stimulate, rather than replace, private-sector
investments. In this sense, the project is consistent with a market-
based and mixed-economy strategy of economic development.
These investments are underpinned by three major science-based
initiatives: Africa’s green revolution (11, 12), the worldwide ad-
vances in treating malaria, HIV-AIDS, and neglected tropical
diseases (5, 6), and improvements in information technology.
the poverty trap can be overcome and the MDGs achieved by 2015
at the village-scale in rural Africa by applying the United Nations
Millennium Project’s recommended interventions in multiple sec-
tors at the investment level of $110 per capita per year sustained
over a period of 5–10 years (Table 1). The main principles of the
Y Science- and evidence-based, implementing technologies and
practices that have already been proven.
Y Community-based, with a participatory approach to planning,
implementation, and monitoring that contextualizes the specific
set of interventions for each village.
Y Enhanced by local capacity development in technical, manage-
rial, and participatory skills.
Y Based on multisectoral and integrated interventions.
Y Linked to district, national, and global strategies.
Y Supported by partnerships with other development groups.
Y Cost-shared by the community, government, and donors.
Y Supported by increased national-scale financing of public goods
in line with increased official development assistance (ODA)
made available to African governments.
productivity, public health, education, and infrastructure. The
interventions are undertaken as a single integrated project; the
synergies and tradeoffs are assessed and highlighted before deci-
sions are made. For example, higher food production has positive
impacts on health and education but might also result in children
missing school by working on farms. Impacts of interventions on
gender and the environment are sometime less obvious but critical
for long-term sustainability. Where adverse cross-sector tradeoffs
Not all interventions can start at the same time and are phased
12–18 months) involves the basics: food, health, water, and com-
munity empowerment. A generic list of interventions, which are
localized for each site, follows:
1. Increased food production. Subsidized provision of improved
seeds of high-yielding crop varieties or hybrids, the necessary
amounts of mineral and organic fertilizers, and training on
best agronomic practices to eliminate hunger months and
generate crop surpluses.
2. Malaria control. Free distribution of long-lasting insecticide-
impregnated bednets (LLINs) for all sleeping sites, preceded by
training and followed by monitoring of use, combined with
access to antimalaria medicines, to drastically reduce the disease
burden of malaria.
community health workers, to provide basic clinical services for
infectious diseases, nutritional deficiencies, antenatal care, and
attended normal delivery.
4. Safe drinking water points constructed with the eventual aim of
having access within 1 km of each household.
5. Community capacity-building, to empower villagers to manage
their own development more effectively; and to enhance the
sustainability of interventions.
Another set of interventions follows, building on the first set.
1. More robust and diversified agriculture using nitrogen-fixing
trees and cover crops, organic manures, crop rotations, soil
conservation practices, livestock, aquaculture, small-scale wa-
ter management, improved crop storage, and crop insurance.
2. Expanded health systems, including further malaria control
through indoor residual spraying, particularly in epidemic areas;
family planning; micronutrient supplementation for vulnerable
groups; treatment and prevention of HIV/AIDS and TB; and
improvements in the nearby referral hospital, including emer-
gency obstetrical care.
3. Functioning primary schools: Universal enrolment in primary
school, with adequate buildings, teachers, materials, separate
latrines for girls and boys, drinking water, and a nutritious
midday meal from locally produced food.
4. Improved clean water, sanitation and personal hygiene. Access
to sufficient clean water for domestic consumption, pit latrines
at home, and sanitary napkins for adolescent schoolgirls.
5. Infrastructure: Upgrading local roads and improving access
roads, connecting to the electrical grid and the internet; trans-
portation to markets.
6. Expanded links with government and other development part-
ners: Steering groups that coordinate local and district-level
activities, planning, and cost-sharing.
7. Commercial farming and business development: Diversifying
farm enterprise toward high-value products and linking pro-
ducer groups to markets. Enterprise development through ca-
pacity building, access to microfinance and microenterprise
8. Environmental rehabilitation: Increasing tree cover, soil con-
servation structures, agrobiodiversity and carbon sequestration.
The first Millennium Villages were established in Sauri,
Kenya, in December 2004, and in Koraro, Ethiopia, in February
2005, with additional villages in 2006, for a total of 12 located in
Ethiopia, Ghana, Kenya, Malawi, Mali, Nigeria, Rwanda, Sene-
gal, Tanzania, and Uganda [see supporting information (SI)
Appendix]. Each site is in a major agroecological zone that,
together, represent the farming systems used by 90% of the
agricultural population and 93% of the agricultural land area of
sub-Saharan Africa (calculated from ref. 13). The sites range
from slash-and-burn in rainforest margins to pastoralism in
Table 1. Recommended level of investment for rural African
villages by the United Nations Millennium Project (1)
U.S. $ per person
U.S. $ per year
per village of
Distribution by sector
Agriculture and nutrition (15%)
Infrastructure (energy, transport,
Water, sanitation, environment (15%)
The MVP contributes $50 of the $70 donor share.
www.pnas.org?cgi?doi?10.1073?pnas.0700423104Sanchez et al.
deserts and represent different situations of population density,
soils, climate, water access, disease complexes and burdens,
environmental degradation, market access, education levels,
cultures, religions, and gender issues. Together, the 12 research
villages provide a robust framework for proof of concept.
This section includes results from the Sauri baseline survey (14),
annual reports from Sauri (15) and Koraro (16) and crop-harvest
data from Malawi (17). In 2004, almost 80% of the population of
Sauri lived below the one-dollar-a-day poverty line. Land holdings
were ?0.6 hectare (ha) for 75% of the households, with 0.22–0.27
highest income quartile (Table 2). More than half of the blood
smears from the sample population in Sauri tested positive for
malaria parasites, with the highest prevalence found in the lowest
income quartile. Only 13% of the households reported having
lowest income quartiles, are another indication of the high levels of
malaria, undernutrition, and intestinal parasites.
Health. Before the project, the nearest health services were pro-
vided at the Yala Subdistrict Hospital (no electricity, water, or
doctors) at a distance of ?5 km. The Sauri Community Dispensary
was opened in July 2005, with the Ministry of Health providing a
clinical officer, the MVP funding two nurses and a laboratory
technician, and the community providing community health work-
ers. In the first year, 35,476 patient cases were seen, of whom
72% were Sauri residents, the others came from surrounding,
nonproject villages. The most common diseases treated were ma-
laria, respiratory tract infections, skin conditions, intestinal worms,
There was no marked difference in malaria prevalence among
residents from Sauri that had received bednets and those from
outside Sauri until October 2005 (Fig. 1). After that date, malaria
prevalence among Sauri residents tested at the clinic fell to one-
third of that of non-Sauri residents who did not receive bednets. In
Koraro, Ethiopia, a clinic did exist but had essentially no staff or
medicines. In 2005, the Ministry of Health assigned a medical
were treated as compared with 260 patients the previous year.
Before bednet distribution in September and October (the two
months of peak malaria transmission) of 2005, there were 281
same period of 2006, there were only 143, a 51% reduction.
Agriculture. The results of the initial harvests in three research
preproject year. In Sauri, 2005 maize yields increased 2.6-fold from
2004, averaging 5.0 t?ha?1at the village-scale (325 ha). Given the
that had been abandoned because of low soil fertility and related
weed infestation. The combined effect was a village-wide 3.9-fold
increase in maize production and a shift from 43% to 166% of the
basic caloric requirements (Table 3).
All income groups of farmers increased their yields by a similar
amount, and yields were similar among income quartiles. On
average, families in the lowest income quartile produced 1.8 t of
maize, surpassing the 1.1-t requirement for basic food security for
a family of 5.7 people. A minimum land area of 0.21 ha is needed
to produce that amount of maize for food security, and one-third
of the households had insufficient area.
The fact that preintervention yields were similar across income
groups and that yields increased similarly across income groups
suggests two important points. First, even the relatively wealthy
households ($2.60 per day) were not using fertilizer and high-yield
seeds before the project; and second, even the poorest households
In 2006, yields were even higher, with an average yield of 6.2
t?ha?1(Table 3) The area planted increased by ?10% above the
2005 area. In 2006, the village surplus was almost 1,300 t, and the
minimum area needed to produce 1.1 t was reduced to 0.18 ha,
meaning that only 27% of the households had insufficient area to
reach basic food security.
The cost of this up-front investment in fertilizer and seeds
supplied by the project was $50 per household planting an average
of 0.25 ha to maize in 2005. Approximately 11% was paid back
through contributions of surplus maize to the school meals pro-
gram, representing a net subsidy of 89%. For the 2006 maize crop,
did buy fertilizer and improved seed from the market. In the third
year, seed and fertilizer subsidies were eliminated for the house-
holds in the top three income quartiles while still fully subsidizing
the poorest and most vulnerable households. Farmers either pur-
chased inputs or obtained loans from a microfinance provider.
highly variable rainfall. The 2004 year was a severe drought year;
yields of the four main cereal grain crops averaged 0.13 t?ha?1,
producing only 418 t of grain which met only 13% of the village’s
caloric food requirement. In 2005, with inputs and good rainfall,
average yields quadrupled but still at a low level of 0.58 t?ha?1
food requirements for the village were met. A partition of the yield
Table 2. Baseline conditions (2004) by income quartiles for the
Sauri Millennium Research Village population
Yearly per capita income (U.S. $, PPP)
Income from agriculture, %
Average land holding, ha
Area planted to maize, ha
Maize yields, kg ha?1
Anemia (?11 mg Hb), %
Testing positive for malaria, %
Results from household surveys, maize harvest estimates, and blood sam-
pling (21). Q1 denotes the lowest income quartile. PPP, purchasing power
Research Village residents and those from outside. Prevalence is the percent-
of patients seen at the dispensary. No data were obtained in November and
December 2005 because of generator breakdown.
Malaria prevalence from July 2005 to June 2006 for Sauri Millennium
Sanchez et al.
October 23, 2007 ?
vol. 104 ?
no. 43 ?
increases in teff, the main cereal crop, indicates that the effect of a
good rainy season without inputs resulted in a yield increase from
0.08 to 0.41 t?ha?1. The use of fertilizer without improved seeds
increased teff yields further to 0.59 t?ha?1, but the full intervention
resulted in an average yield of 1.20 t?ha, a 15-fold increase.
Mwandama. Mwandama, Malawi, suffered from a drought in the
year preceding the start of the project. Even in good rainy seasons,
however, the perennial nitrogen ‘‘drought’’ (i.e., shortage of nitro-
gen in the soils) results in low maize yields. In the first planting
season, maize yields increased from 0.8 to 6.5 t?ha?1, and the area
planted almost doubled, the total maize production increased
and fertilizers averaged 2.2 t?ha?1, showing that good rains alone
accounted for less that half of the yield increase. No one expects
these village average yields to be replicated at the country-scale.
maize production doubled, and average maize yields doubled from
0.73 to 1.61 t?ha?1(17).
Crop surpluses minimize risks of food shortages in subsequent
years but also serve as the entry point for entering the cash
prices, leaving farmers with their surpluses unsold (2, 3). In Sauri,
2005 by local middlemen, less than the official price of about $20.
In need of cash to buy essentials, farmers normally sell at these
prices, later to run out of food and buy back maize for as high as
$25 a bag. To buffer such price fluctuations, a cereal bank was
established by renting storage space and using project funds to pay
at $21 per bag in April 2006, farmers were paid the difference,
minus storage and management fees, effectively doubling the price
they would have received selling to middlemen. In Ethiopia a
five-ton truck provided by the project allowed farmers to transport
their surplus crop of 84 tons of teff and finger millet to the nearest
market, which was sold at $39,528.
After increased food production, diversification to higher-value
crops is being promoted. Sauri farmers have organized producer
groups for onion, tomato, banana, dairy, and mushrooms. Many
high-value products. Microfinance mechanisms, including but not
limited to microcredit, are being established, including training in
record-keeping and accounting. Crop insurance schemes are being
developed to reduce risks of future crop failures caused by drought.
The MVP aims to demonstrate the feasibility of practical economic
transformation in rural tropical Africa through targeted multisec-
toral investments. This approach has similarities and differences
and 1980s. The main similarities are simultaneous, complementary
interventions, which create synergies and a major initial focus on
agricultural productivity, including agroprocessing (18–22).
IRD projects were developed by a range of donors and govern-
ments with different priorities, resource allocations, designs, exe-
cution, and duration. In contrast, the MVP model is focused on
achieving time-bound and quantitative goals and promotes a more
comprehensive set of sectoral interventions. Many IRD projects
focused on more prosperous areas; World Bank programs, for
example, were often based in high-growth areas (18). In contrast,
the Millennium Villages are situated in hunger hotspots where at
least 20% of children under 5 are underweight (2, 3).
IRD projects were often based on insufficient experience with
local agricultural systems, and the new farming interventions had
seldom been tested with small holders (22). MVP interventions are
similar ecological and socioeconomic conditions. Whereas IRD
projects were typically top-down, government and community
participation is a major feature of the MVP. Ownership can be
generated by communal efforts and in-cash, in-kind contributions,
a feature of MVP.
year duration of IRD projects. Although the scale of investment of
IRD projects is unclear, they were probably funded at much lower
benefit from three key developments since the IRD era: (i) decen-
tralization and devolution of authority to local government; (ii)
internationally agreed financial commitments to double ODA in
Africa by 2010 and then further increases to 2015; and (iii) major
advances in environmentally-sound agriculture, health, and infor-
There have been many attempts to bring a green revolution to
Africa (23). The Asian green revolution benefited from high-
yielding crop varieties, fertile alluvial soils, and irrigation, with crop
improvement accounting for 66–88% of the yield increases (24).
The green revolution in Africa has not benefited similarly from
improved crop varieties but must also redress nutrient-depleted
soils and cope with little to no irrigation. There is abundant
evidence throughout Africa on crop yield response to fertilizers,
nutrient limitations by sufficient rates of both mineral and organic
sources of nutrients (33). Mineral fertilizers are distributed the first
nutrients to get farmers used to fertilizers but at rates insufficient
for high yields (34). Others promoted full rates of plant nutrients
and improved seed that have usually produced bumper crops (35),
but, in some cases, this resulted in collapsing prices, so farmers lost
money and were unable to pay back their loans (2). To address this
challenge, the MVP interventions include a cereal-storage and
marketing strategy to buffer price fluctuations in anticipation of
Table 3. Basic food production increases in the first three villages’ harvests
(times from *)
Data are for maize (Zea mays) in Kenya and Malawi and an average of teff (Eragrostis tef), sorghum (Sorghum vulgare), finger millet
people or 1,045 t of grains in Ethiopia).
www.pnas.org?cgi?doi?10.1073?pnas.0700423104 Sanchez et al.
Results are still early, although the initial agriculture revolution
has begun in most sites with high yields and crop surpluses. Crop
surpluses were used to initiate school meals programs. Positive
interaction between improved labor productivity (less malaria) and
land productivity allow agricultural activities to intensify. The
subsequent phases of crop diversification and linking farmers to
Targeted input subsidies are necessary to jump-start increased
food production and to sustain yields of both food crops and
high-value agricultural products (2, 3, 17). The strategy is to
subsidize fertilizers, improved seeds of maize, vegetables, grain
legumes, and nitrogen-fixing trees, treadle pumps, and other inputs
through local dealers and provide them with technical training,
making a network of agrodealers in rural Africa. One approach is
to use vouchers with information on the farmer’s eligibility for
subsidized inputs, limiting them to one or two bags of fertilizer and
10 kg of improved seed, magnitudes of inputs that are too small to
interest large-sale commercial farmers (17). The vouchers are
redeemed for inputs at agrodealers, who get paid by the banks
where the subsidies are deposited by governments or donors. Such
and empower the private sector. Partnerships are being developed
with companies working across the food chain to ensure quantity,
quality, and timely delivery of products.
Limitations of Project Design. The MVP is a highly complex project,
we mention scale and controls. First, the scale of each project site
(a cluster of a several villages totaling 5,000–55,000 people) is
determined by available project financing and implementation
to justify large-scale infrastructure (e.g., a new water-treatment
facility) or to justify inflows of foreign buyers for sourcing agricul-
tural-based primary commodities, or to promote foreign invest-
ments in processing facilities. They are vulnerable to inflows of
people from surrounding areas and might generate resentment in
surrounding communities. Finally, the project remains vulnerable
with national-scale interventions. For these reasons, we should
expect that a larger-scale effort than MVP would benefit from
important economies of scale that will not be evident in the MVP
alone. The scaling-up of the key interventions to district and
national level is discussed below.
For ethical and practical reasons, there are no formal ‘‘control’’
villages. Instead, project impact is assessed by rigorous before-and-
after comparisons and detailed studies by sector. Also, we collect
district-level indicators, impacts measured by other rural projects,
and household longitudinal surveys taking place in several of the
countries. The ethical reasons relate to the fact that many core
interventions (e.g., malaria control, access to safe water) are
village. The project also discovered that in some of the sites it was
village without also offering a substantial package of beneficial
Nor are individual interventions randomized across or within
villages to try to get intervention-specific impacts. In some cases
(e.g., the impact of fertilizers and seed varieties on crop yields), the
data will allow us to assess the impact of specific interventions. In
other cases (e.g., anemia), the outcome will result from several
interventions (deworming, food intake, nutritional supplements,
fact that key outcomes have multiple and synergistic causes is even
more true for more complex outcomes such as school attendance
and school performance.
Sustainability. One of the complex issues regarding the MVP is the
exit strategy for external donors. The project commits to 5 years of
funding and anticipates that the community will become econom-
ically self-sustaining in commercial farming and nonfarm activities
within that period. We hypothesize that farmers will be able to
procure inputs and sell outputs on a normal market basis at the
completion of the first 5 years of the project, based on higher
productivity and greater product diversity. Subsidies of farm inputs
may continue as part of national policy but should no longer be
necessary from the project itself.
The community, however, will not be able to bear the financing
for the interventions in health, education, and infrastructure.
Rather, the national governments should, by the 5th year of the
project, have national financing of such interventions out of an
expanded budget for such priority investments. All governments in
the world are committed to achieving the MDGs, and the rich
countries are committed to a significant increase of ODA to
support the MDGs in the poorest countries. Thus, by 2011 (the 6th
year of the project for the current Millennium Villages), the aid
received by Africa should be more than double the 2005 levels,
reaching at least $50 billion per year (a per-capita equivalent of
$75–100). This increased aid will be directed in part to expanding
government support for critical services in health, education, and
infrastructure, and the expanded national budget should substitute
for financing provided by the MVP in the years to 2011.
Scalability. Chances for success of the MVP depend, naturally, on
whether the increased government budgets or ODA for public
goods actually materializes and whether the types of interventions
pioneered by the MVP are thereby expanded to other rural areas
in the host countries. If the villages remain islands of prosperity in
by in-migration and undermined by neighborhood jealousy. If
instead, the Millennium Villages are part of an expanded national-
scale effort to achieve the MDGs, then they will benefit from an
expanding national market, increased social and political stability,
improved national-scale infrastructure, and success in attracting
private investments that connect the villages with regional and
As part of the overall Millennium Project, a senior country
advisor based in the capital of each of the 10 countries works with
government to assist in bringing the MDGs into the budget and
To encourage that expanded scale, two additional initiatives are
being launched. The Millennium Cities Initiative works with the
major regional city closest to the Millennium Villages to improve
the urban business environment, to attract foreign investments into
the city, and, in part, to create a market for the increased outputs
of the villages and surrounding areas. After the 1st year of the
project, several countries in which Millennium Villages are located
are working with the donor community and using their own
resources to establish additional villages and also to scale up from
the village-scale (5,000 people) and ‘‘cluster’’-scale (30,000–50,000
people) to a Millennium District-scale (300,000–500,000 people).
With interventions applied at a district-scale, not only will there be
a greater scope of the market and improved chances for a produc-
tive division of labor, but also improved chances for attracting
inflows of foreign investment and for undertaking public invest-
ments in roads, power, telecommunications, and other sectors
where the scale of investments exceeds what can be accomplished
at the village-scale.
The following methods are described for the first research village
in Sauri, Kenya, but apply as standard protocols and generic
intervention strategies for all villages. See SI Appendix for site
Sanchez et al.
October 23, 2007 ?
vol. 104 ?
no. 43 ?
Baseline Assessments. Before intervention activities begin, compre-
hensive baseline assessments are conducted through household
surveys, anthropometric and biophysical measurements (see SI
Appendix). These surveys are used to determine baseline MDG
initial conditions and targets to meet the MDGs and to monitoring
impact of interventions. Repeat surveys will be administered in
years 3 and 5 (and 10). The content of the survey modules is
samples to determine nutritional status, levels of anemia, malaria
parasite types and infection levels, weights and heights of children
under 5 years of age, and stool samples for determining parasite
An initial demographic survey provides information to stratify
the population according to geographic representation and wealth
categories. The wealth categories are determined by identifying
assets considered important by each of the communities. A com-
whereas the remaining surveys are administered to a stratified
subsample of 300 households, determined by geographic and
Malaria Control, First Phase. In Sauri, from May to July 2005, 3,000
Olyset long-lasting insecticide-impregnated bednets were distrib-
uted free to all households to cover all sleeping sites, roughly three
nets for every five people. Nets were distributed after training
sessions on malaria prevention and treatment and bednet use by
community health workers. The bednets were distributed free of
charge because they constitute a public good, like immunizations,
The impact of the long-lasting insecticide-impregnated bednets
came to the clinic in Sauri between those that had received bednets
(Sauri residents) and those that had not received bednets (nonres-
idents). Malaria was diagnosed by clinical symptoms and, for the
questionable cases, was confirmed by examination of blood smears
for presence of malaria parasites.
Agriculture, First Phase. Sauri farmers were provided with 96 kg of
urea for top dressing, the full recommended rate of mineral
fertilizers for these nitrogen-depleted and high-phosphorus-fixing
Oxisols and oxic Alfisols. In addition, they received high-yielding
hybrid maize seed (WS 502 and WS 505) from a local seed
company. Farmers also requested training on the best agronomic
practices, which was done by agricultural extension agents before
a document agreeing to use them in their fields and return 10% of
their crop surplus to the village schools for the feeding program.
Similar arrangements took place in the other villages.
farms that used fertilizer, improved seed, and training. This sample
was again stratified according to wealth and geographic criteria. In
addition, crop yields were estimated from a minimum of 30 farms
that did not use improved agricultural inputs. Crop estimates per
farm are obtained by taking two to three samples from 5 ? 5-m
quadrants randomly placed within the maize field. Subsamples of
the cobs and grain are taken and dried to determine grain yields at
14% moisture content.
Basic food security for the maize-based villages of Sauri and
maize produced and the caloric requirements for 5,000 people,
caloric requirement is met through consumption of maize, as
reported for Western Kenya (38), and that 1 kg of maize provides
3,500 calories (39).
The preparation of this paper was supported by a Special Initiative Grant
from the Bill and Melinda Gates Foundation. Funding for the 12
Millennium Research Villages was provided by the Lenfest, Blaustein,
Sara McCune, Stephen Lewis/MACAID, and Yara Foundations and the
Government of Japan through the United Nations Develoment Pro-
gramme’s Human Security Trust Fund.
1. UN Millennium Project (2005) Investing in Development. A Practical Plan to Achieve the
Millennium Development Goals (Earthscan, London).
2. UN Millennium Project (2005) Halving Hunger–It Can Be Done. UN Millennium Project Task
Force on Hunger (Earthscan, London).
3. Sanchez PA, Swaminathan MS (2005) Science 307:357–359.
4. Sanchez PA, Swaminathan MS (2005) Lancet 365:442–444.
5. UN Millennium Project (2005) Combating AIDS in the Developing World. UN Millennium
Project Task Force on HIV/AIDS, Malaria TB, Access to Essential Medicines (Earthscan,
6. UN Millennium Project (2005) Coming to Grips with Malaria in the New Millennium. UN
Millennium Project Task Force on HIV-AIDS, Malaria TB, Access to Essential Medicines
7. UN Millennium Project (2005) Health, Dignity and Development: What Will It Take? UN
Millennium Project Task Force on Water and Sanitation (Earthscan, London), 206 p.
Bank, Washington, DC).
9. Sanchez PA (2002) Science 295:2019–2020.
10. Sachs JD (2005) The End of Poverty (Penguin, New York).
11. UN Millennium Project (2004) Proceedings of the July 5, 2004, High-Level Seminar, Addis
Ababa, Ethiopia (Millennium Development Goals Technical Support Centre, Nairobi,
Kenya) pp 1–64.
12. United Nations (July 6, 2004) Secretary-General Calls for ‘Uniquely African Green Revolution’
in 21st Century, to End Continent’s Plague of Hunger, in Addis Abada Remarks, press release,
13. Dixon J, Gulliver A, Gibbon D (2001) Farming Systems and Poverty (Food and Agriculture
14. Mutuo P, Lelerai E, Okoth H, Oule J, Ouma BA, Wangila J, Palm CA, Wang K, Konecky
Columbia University, New York).
15. Mutuo P, Okoth H, Makomere C, Oule J, Oduong G, Ombai W, Wariero J, Akinyi B (2006)
Annual Report for Sauri, Kenya Millennium Research Village July 2005–June 2006 (Earth
Institute, Columbia University, New York).
16. Kiflemariam T, Gebretsadik A, Kebede G, Mehari G, Habtetsion T, Ye-ebiyo Y, Said A,
Desta H (2006) Annual Report for Koraro, Ethiopia Millennium Research Village July
2005–June 2006. (Earth Institute, Columbia University, New York).
17. Denning GL, Kabambe P, Harawa R, Nkhoma P, Zamba C, Banda C, Magombo C, Keating
M, Malik A, Flor R, et al. Malawi: Start of an African Green Revolution? Science, in press.
18. Mkandawire T (1980) The World Bank and Integrated Rural Development in Malawi, Working
Paper No. 1 (Council for the Development of Economics and Social Research in Africa,
19. Sallinger-McBride J, Picard LA (1989) Comparative Politics 22:1–22.
20. Binns JA, Funnell DC (1983) Geograf Ann 65:57–63.
21. Department for International Development (DFID (2004) Synthesis of Integrated Rural
Development Projects, Evaluation Summary 438 (DFID, London). Available at www.dfid.
gov.uk/aboutdfid/performance/files/ev438s.pdf. Accessed June 4, 2007.
22. DFID (2004) Rural Development in Africa: A Synthesis of Project Experience, Evaluation
Summary 400 (DFID, London). Available at http://www.dfid.gov.uk/aboutdfid/
performance/files/ev400s.pdf. Accessed June 4, 2007.
23. Djurfeldt G, Holmen H, Jistro M, Larsson R (2005) The African Food Crisis. Lessons from
the Asian Green Revolution (CABI, Wallingford, UK).
24. Evenson RE, Gollin D, eds (2001) Crop Variety Improvement and Its Effect on Productivity:
The Impact of International Agricultural Research (CABI, Wallingford, UK).
25. Buresh RJ, Sanchez PA, Calhoun FJ, eds (1997) Soil Fertility Replenishment in Africa, Spec
Pub 51 (Soil Sci Soc of America (Madison, WI).
26. Sanchez PA, Buresh RJ, Leakey RRB (1997) Philos Trans R Soc London Ser B 352:949–961.
27. Shepherd KD, Soule MJ, (1998) Agric Ecosyst Environ 71:131–145.
28. Nziguheba G, Palm CA, Buresh RJ, Smithson PC (1998) Plant Soil 198:159–168.
29. Buresh RJ, Cooper PJM, eds (1999) Agroforestry Syst 47:1–368.
30. Vanlauwe B, Diels J, Sanginga N, Merckx R, eds (2002) Integrated Nutrient Management in
Sub-Saharan Africa: From Concept to Practice (CABI, Wallingford, UK).
31. Gichuru MP, Bationo A, Bekunda MA, Goma HC, Mafongoya PL, Mugendi DN, Murwira
HM, Nandwa SM, Nyathi P, Swift MJ, eds (2003) Soil Fertility Management in Africa: A
Regional Perspective (Tropical Soil Biology and Fertility, Nairobi).
32. Bationo A, Waswa B, Kihara J, Kimetu J, eds (2006) Nutr Cycl Agroecosyst 76:1–381.
eds Monis AC, Furlani AMC, Schaffert RE, Fageria NK, Rosolem CA, Cantarella H
(Brazilian Soil Sci Soc, Piracicaba, Brazil), pp 19–28.
34. Seward PD, Okello D (1998) Methods to Develop an Infrastructure for the Supply of
Appropriate Fertilizer for Use by Small-Scale Farmers in Sub Saharan Africa (Sustainable
Community-Oriented Development Program, Ukwala, Kenya).
35. Quin ˜ones MA, Borlaug NE, Dowswell CR (1997) Soil Sci Soc Am Spec Pub 51:81–95.
36. Teklehaimanot A, Sachs JD, Curtis C, (2007) Lancet 369:2143–2146.
37. Food and Agriculture Organization (2004) Food and Nutrition Technical Report Series 1
38. Rommelse R (2001) Natural Resources Problems, Priorities and Policies Working Paper
2001-2 (International Centre for Research in Agroforestry, Nairobi).
39. Okoruwa AE, Kling JG (1996) IITA Research Guide 33 (International Institute of Tropical
Agriculture, Ibadan, Nigeria).
www.pnas.org?cgi?doi?10.1073?pnas.0700423104Sanchez et al.
CELL BIOLOGY. For the article ‘‘Activation of the PTEN/mTOR/
STAT3 pathway in breast cancer stem-like cells is required for
viability and maintenance,’’ by Jiangbing Zhou, Julia Wulfkuhle,
Hao Zhang, Peihua Gu, Yanqin Yang, Jianghong Deng, Joseph
B. Margolick, Lance A. Liotta, Emanuel Petricoin III, and Ying
Zhang, which appeared in issue 41, October 9, 2007, of Proc Natl
Acad Sci USA (104:16158–16163; first published October 2,
2007; 10.1073?pnas.0702596104), the authors note that the leg-
end for Fig. 1 appeared incorrectly in part. The figure and its
corrected legend appear below. This error does not affect the
conclusions of the article.
(B) Proliferation inhibition effects of LY294002 (2.5 ?M), rapamycin (5 ?M), and IS3 295 (50 ?M) on MCF7 SP and non-SP cells. (C) Colony formation inhibition
effects of LY294002 (2.5 ?M), rapamycin (5 ?M), and IS3 295 (25 ?M) on MCF7 SP and non-SP cells. (D) Western blot analysis of expression of STAT3 and mTOR
in mTOR knockdown cells. (E) Decrease of SP fraction in stable mTOR and STAT3 knockdown MCF7 cells. (F) Western blot analysis of expression of STAT3 and
mTOR in STAT3 knockdown cells. (G) Negative regulation of STAT3 and mTOR expression by PTEN, shown by Western blot. (H) Increase of SP fraction in MCF7
cells by PTEN knockdown. (I) Increase of SP fraction in MCF7 cells treated with PTEN-specific inhibitor bpV(pic).
Importance of PI3K/mTOR, STAT3, and PTEN signaling for the SP cells. (A) Decrease of the SP fraction within MCF7 cells by pathway-specific inhibitors.
December 4, 2007 ?
vol. 104 ?
no. 49 ?
MICROBIOLOGY. For the article ‘‘Isolates of Zaire ebolavirus from
wild apes reveal genetic lineage and recombinants,’’ by Tatiana
J. Wittmann, Roman Biek, Alexandre Hassanin, Pierre Rou-
quet, Patricia Reed, Philippe Yaba, Xavier Pourrut, Leslie A.
Real, Jean-Paul Gonzalez, and Eric M. Leroy, which appeared
in issue 43, October 23, 2007, of Proc Natl Acad Sci USA
(104:17123–17127; first published October 17, 2007; 10.1073?
pnas.0704076104), the affiliation for Jean-Paul Gonzalez should
have appeared as ‘‘Institut de Recherche pour le De ´veloppe-
ment, UR178, Center for Vectors and Vector-Borne Diseases,
Faculty of Science, Mahidol University at Salaya, Phutthamon-
thon 4, Nakhonpathom 73170, Thailand.’’ The corrected affili-
ation line appears below.
†Centre International de Recherches Me ´dicales de Franceville, BP 769
Franceville, Gabon;‡Department of Biology and Center for Disease Ecology,
Emory University, 1510 Clifton Road, Atlanta, GA 30322;§Muse ´um National
d’Histoire Naturelle/Centre National de la Recherche Scientifique, Unite
Mixte de Recherche 5202, 75005 Paris, France;¶Wildlife Conservation
Society, 2300 Southern Boulevard, Bronx, NY 10460;?Institut de Recherches
pour le De ´veloppement, UR178, Centre International de Recherches
Me ´dicales de Franceville, BP 769 Franceville, Gabon; and††Institut de
Recherche pour le De ´veloppement, UR178, Center for Vectors and
Vector-Borne Diseases, Faculty of Science, Mahidol University at Salaya,
Phutthamonthon 4, Nakhonpathom 73170, Thailand
NEUROSCIENCE. For the article ‘‘Dissociating the neural mecha-
nisms of memory-based guidance of visual selection,’’ by David
Soto, Glyn W. Humphreys, and Pia Rotshtein, which appeared
in issue 43, October 23, 2007, of Proc Natl Acad Sci USA
(104:17186–17191; first published October 16, 2007; 10.1073?
pnas.0703706104), the authors note that, due to a printer’s error,
the third line of the Abstract appeared incorrectly in part. ‘‘How
WM and implicit priming affects influence visual selection
remains poorly understood, however’’ should read: ‘‘How WM
and implicit priming influence visual selection remains poorly
understood.’’ Additionally, on page 17186, in the second line of
the second paragraph, left column, ‘‘This work has led to the
development of the influentially biased competition model of
visual selection (1), where it is hypothesized that memory acts to
bias the competition for selection between different objects in
the visual scene’’ should read: ‘‘This work has led to the
development of the influential biased competition model of
visual selection (1), where it is hypothesized that memory acts to
bias the competition for selection between different objects in
the visual scene.’’ These errors do not affect the conclusions of
MEDICAL SCIENCES. For the article ‘‘The EGF receptor is required
Wagner, and Maria Sibilia, which appeared in issue 43, October
23, 2007, of Proc Natl Acad Sci USA (104:17081–17086; first
published October 16, 2007; 10.1073?pnas.0704126104), the
authors note that, due to a printer’s error, line 3 of the
Acknowledgments appeared incorrectly. The sentence should
read as follows: ‘‘This project was supported by Austrian Na-
tional Bank O ¨NB-10556 and the European Community Grant
SUSTAINABILITY SCIENCE. For the article ‘‘Poverty and Hunger
Special Feature: The African Millennium Villages,’’ by Pedro
Sanchez, Cheryl Palm, Jeffrey Sachs, Glenn Denning, Rafael
Flor, Rebbie Harawa, Bashir Jama, Tsegazeab Kiflemariam,
Bronwen Konecky, Raffaela Kozar, Eliud Lelerai, Alia Malik,
Vijay Modi, Patrick Mutuo, Amadou Niang, Herine Okoth,
Frank Place, Sonia Ehrlich Sachs, Amir Said, David Siriri,
Awash Teklehaimanot, Karen Wang, Justine Wangila, and Col-
leen Zamba, which appeared in issue 43, October 23, 2007, of
Proc Natl Acad Sci USA (104:16775–16780; first published
October 17, 2007; 10.1073?pnas.0700423104), due to a printer’s
error, the author name Vijay Modi did not appear in the Table
of Contents. The online version has been corrected.