|
At the global and continental levels, what
can be concluded from this analysis? Livestock production systems,
and the households that operate them, face major changes in the
next 5 decades. The spatial projections of human population growth, particularly
in SSA, are quite startling. Equally startling are the predicted changes in length
of growing period for SSA using the Hadley Global Circulation
Model (GCM). (It should be remembered that temperature increases may
be double what is simulated in the Hadley scenario; changes in LGP
simulations using the most recent GCM outputs would,
in all likelihood, be even more pronounced than those shown in Section 4.)
Add these projections to the increases in demand for
livestock products forecast globally, in SSA as well
as in Asia and South America (Delgado et al., 1999),
and the outlook is extremely dynamic.
In terms of
the numbers of poor and, so far as the analysis is capable of
distinguishing, the numbers of poor livestock keepers, the critical
regions are SA and SSA. Our analysis indicates that while the
rangeland systems contain relatively few poor (some 60 million),
most of these households are dependent on livestock for their
livelihoods. Almost half of the poor in rangeland systems are
located in SSA. The mixed systems contain large numbers of poor
(over 1 billion), and the numbers of poor who depend to some extent
on livestock are considerable; the mixed irrigated systems contain
approximately 103 million poor livestock keepers, and the mixed
rainfed systems some 366 million poor livestock keepers. In terms of
the magnitude of poverty and the importance of livestock to poor
households in the developing world, this analysis suggests that
there are at least 550 million poor livestock keepers globally.
Analysis
using the global data sets as outlined above can be of value, not
least as the first step in a two-tiered approach that involves
identification of hot spots of rapid change, a second step then
involving a zoom-in to these areas for more detail. At a global
level, and even with relatively coarse data sets, we can already
identify hot spots where system changes are likely to be substantial
over the next 3–5 decades, as a result of population
growth and climate change. The magnitude of these system changes,
particularly in SSA, may be so large as to be potentially
overwhelming.
At these
broader scales, our knowledge of poor livestock keepers is still
generally quite limited. At much higher resolutions, a great deal of
survey work and data collection has been carried out at the
community level. The data and maps for Kenya in Section 3
indicate that the conventional wisdom concerning livestock is not
necessarily correct. Poor households often have access to land; they
also often have cattle as well as smallstock—land and cattle are
not just the prerogative of the non-poor in these systems.
A major
issue remains: how can the links be made between such household
survey data and case studies and the broader picture, so that
reliable extrapolation and generalisation can be carried out,
thereby providing information at levels of aggregation appropriate
for making resource allocation decisions? While analyses based on
global data sets are useful, they can go only so far. As noted
above, methods based on small-area estimation (SAE) to produce
poverty maps at the level of census enumeration areas have been and
are being applied in various countries, but a concentrated effort is
needed. ILRI, in collaboration with WRI, the World Bank, the Rockefeller
Foundation, the International Food Policy Research Institute (IFPRI)
and national poverty teams, is currently undertaking SAE poverty
mapping in Kenya, Tanzania and Uganda. This may result in
preliminary high-resolution poverty maps by the end of 2002 (Appendix
4). IFPRI is engaged in producing similar maps for
Mozambique and Malawi. With the maps that were completed in 2000 in
South Africa, this is a reasonable coverage of countries for East
and Southern Africa, but more work is needed.
Potential
sites where livestock research could be focused
How might
poverty maps be used to prioritise and focus livestock research
appropriately? The case for poverty alleviation as a (if not the)
major research and development goal is given. But poverty is a
difficult concept, largely because of its multidimensionality. Is it
sufficient to locate poor livestock keepers, so that in DFID’s
target countries, for example, something can be said about the livestock systems where there are large
percentages of poor livestock keepers? But which criteria should
be used—absolute numbers of poor, or systems with high
poverty rates and where environmental issues are important? What about
the links between poverty and natural resources management, and
the nature of the interventions themselves—how likely are
these to affect the poor? And so on.
What this indicates to
us is the need for a consistent framework that can be used to set
priorities. Any realistic and convincing attempt to answer the
questions posed above will involve trade-offs between the criteria
used. One of the major outcomes of the recent ILRI priority-setting
exercise was that there are no ‘wonder’ livestock research solutions—that generate high returns to research investment, that can have a
huge impact on poor people and that score highly on environmental-impact criteria. While in one way
this is disappointing, in another it seems realistic, and
a powerful validation of the entire process. Obtaining consensus on what
such a framework should consist of, and the criteria used,
is a difficult though necessary task, even if different
agencies place different weights on the various criteria that are
used to arrive at appropriate portfolios of activities.
However such a
framework might be built, and whatever it might consist of,
there are key ingredients that would include basic information
on spatial and temporal distribution of crops and livestock; on the
numbers, location and characteristics of the poor; and on the
numbers, location and characteristics of highly vulnerable poor livestock
keepers. Despite the crucial importance of such information, our databases
are, by and large, very patchy and incomplete.
Assembling and
maintaining such databases is not given a very high priority by most
donors, although there are one or two exceptions. Co-ordinated and
coherent joint efforts to develop and maintain these types of
database would seem to be an obvious response to these data gaps, so that wheels are not continually
reinvented and the existence of key baseline data is
assured. Currently, there are no such co-ordinated efforts, but donor inputs
to promote truly inclusive database development and maintenance could be
a highly effective motor to drive the improvement of
global and regional data sets, not only for priority setting
but for a wide variety of other purposes.
Areas where the analysis
could be improved and further work
Several weaknesses can be identified
in the map of global livestock production systems, and these are
outlined in more detail in Appendix
3. In summary, the
classification depends on data of land cover/land use that could
certainly be improved. It should also be remembered that the
category classified as ‘other’ contains ecosystems that range from
arctic tundra to tropical rain forest to desert. There is also a
great deal of variation within all of the production systems
categories, particularly with respect to agricultural production
potential. There are also likely to be differences in the level of
poverty of livestock keepers within the same production system
associated with differences in livestock production potential, but
there is much that is unknown.
Despite
various weaknesses, the poverty maps in this report represent a
considerable step forward at the global level. We can identify
several areas of work to improve these in the future, including the
following:
-
A simple model of migration
patterns over the next 50 years should be developed, and this
could then be used to refine the human population projections to
2050. At the continental and global levels, human population
density is a very powerful proxy for a wide range of variables,
and the benefits of refined, spatial projections of human
population density could be expected to be considerable in many
fields.
-
Further refinement of production
systems categories could be made by accounting for different
levels of land-use intensity and different levels of productive
potential caused by soil fertility.
-
Further
sub-country and
country-level studies are needed to quantify rates of poverty
between and within different production systems. These measures
should attempt to associate different poverty rates with such
geographic variables as market access, natural resource endowment
and climate. These studies could usefully be at two levels: rapid
broad-scale community assessments and studies based on SAE. Both
types of study would greatly improve our understanding of the
proportion of income people in different production systems derive
from livestock, and thus the importance of livestock to their
livelihoods.
-
The conventional wisdom is that
the poor are concentrated where natural resources are either poor
or degraded, but whether this is correct or not is unknown. Better
understanding of the relationships between poverty and
natural resources would be of great value to development agencies
in the design of livestock interventions that are more sustainable
and that minimise adverse impacts. This work could usefully start
with an analysis of poverty and soil degradation, biodiversity and
carbon sequestration potential.
-
It seems to us that a fruitful
way to explicitly include the time dimension into studies of
poverty is through a marriage of notions of poverty with
vulnerability, perhaps partly in the context of marginal lands and
marginal areas (CGIAR, 1997). Some of the poor are bound to be
more vulnerable than others to such climatic shocks as drought or such political shocks
as revolution. For example,
pastoral people who live in areas with 300 mm of reliable annual
rainfall may be less vulnerable to other risks than those who live
in areas characterised by similar amounts of rainfall that is
highly erratic and unreliable. Global analyses of vulnerability
combined with poverty maps could contribute greatly to refining
the types of analyses attempted in this study.
|
