...

PDF

by user

on
Category: Documents
3

views

Report

Comments

Description

Transcript

PDF
EPTD DISCUSSION PAPER NO. 45
PATHWAYS OF DEVELOPMENT IN THE HILLSIDES
OF HONDURAS: CAUSES AND IMPLICATIONS FOR
AGRICULTURAL PRODUCTION, POVERTY,
AND SUSTAINABLE RESOURCE USE
John Pender, Sara Scherr, and Guadalupe Durón
Environment and Production Technology Division
International Food Policy Research Institute
2033 K Street, N.W.
Washington, D.C. 20006 U.S.A.
May 1999
EPTD Discussion Papers contain preliminary material and research results, and are
circulated prior to a full peer review in order to stimulate discussion and critical comment. It
is expected that most Discussion Papers will eventually be published in some other form, and
that their content may also be revised.
ABSTRACT
Based on a survey of 48 communities in central Honduras, this paper identifies
the major pathways of development that have been occurring in central Honduras since
the mid-1970s, their causes and implications for agricultural productivity, natural
resource sustainability, and poverty. Six pathways of development were identified: 1)
basic grains expansion communities—where basic grains production is the dominant
activity and increased basic grains production has occurred; 2) basic grains stagnation
communities—where basic grains production is dominant but has stagnated or declined;
3) coffee expansion communities—where coffee production is important and has been
increasing in importance; 4) horticultural expansion communities—where substantial
adoption and expansion of horticultural crops has occurred; 5) forestry specialization
communities—where forestry activities are important and basic grains production is
stagnant or declining; and 6) nonfarm employment communities—where nonfarm
employment is a major and increasing source of income. The pathways were
distinguished by factors determining comparative advantage, including agricultural
potential, population density, and access to markets and technology. Changes in
agriculture and resource management differ significantly among these pathways, though
poverty was found to decline to a similar extent across all pathways. It appears that the
key causes of change in productivity and resource management are different and more
pathway-dependent than the key causes of change in poverty, which depends to a great
extent on provision of public services. Basic infrastructure and public services are badly
needed throughout most of central Honduras, while efforts to address sustainable
agricultural development may not be sufficient to solve poverty problems. There may not
be large tradeoffs between achieving more sustainable development and reducing
poverty, since the causes are different. The findings also imply that a “one-size-fits-all”
approach to technical assistance is unlikely to be successful, since different approaches
show promise in different pathways.
CONTENTS
1. Introduction....................................................................................................................1
2. Conceptual Framework And Methodology ...................................................................3
Conceptual Framework........................................................................................... 3
Research Hypotheses .............................................................................................. 4
1. Impacts of Population Growth ..................................................................... 5
2. Impacts of Improved Market or Technological Opportunities .................... 6
3. Pathways of Development ........................................................................... 7
Methodology ........................................................................................................... 8
3. Pathways of Development in Central Honduras..........................................................13
Characteristics of the Pathways ............................................................................ 15
Determinants of the Pathways .............................................................................. 20
4. Agriculture and Natural Resource Management..........................................................21
Crop Production .................................................................................................... 22
Use of Organic Inputs and Conservation Measures.............................................. 26
Collective Investment in Natural Resource Management .................................... 29
5. Development Outcomes...............................................................................................30
Productivity........................................................................................................... 30
Natural Resource Conditions ................................................................................ 33
Poverty .................................................................................................................. 34
6. Summary and Conclusions ..........................................................................................37
Findings Relative to the Research Hypotheses..................................................... 39
Population Pressure......................................................................................... 39
Market Access................................................................................................. 40
Access to Technology ..................................................................................... 41
Policy Implications ............................................................................................... 42
References..........................................................................................................................45
i
PATHWAYS OF DEVELOPMENT IN THE HILLSIDES OF HONDURAS:
CAUSES AND IMPLICATIONS FOR AGRICULTURAL PRODUCTION,
POVERTY, AND SUSTAINABLE RESOURCE USE
John Pender, Sara J. Scherr, and Guadalupe Durón*
1. INTRODUCTION
In Central America, rapid population growth and agricultural expansion and
intensification in steep hillside areas have led to growing concerns about deforestation,
soil erosion, watershed degradation, loss of soil fertility, sedimentation of reservoirs, and
other resource problems (Leonard 1987; Scherr and Neidecker-Gonzales 1997). Despite
the common perception that population growth and agricultural intensification are
primary causes of land degradation and other problems, evidence from many parts of the
world suggests that sustainable use of resources in fragile lands can be achieved at much
higher population densities than exist in most of Central America (Templeton and Scherr
1997; Tiffen et al. 1994; Turner et al. 1993). Whether such pressures lead to resource
and human welfare degradation or improvement depends upon a complex set of factors
that can vary substantially from one community to the next; including population density,
*
John Pender is a Research Fellow at the International Food Policy Research
Institute (IFPRI); Sara Scherr is a Visiting Fellow in the Agricultural and Resource
Economics Department, University of Maryland, and Guadalupe Durón is a research
assistant at the International Center for Research on Women. The authors gratefully
acknowledge the financial support of the Swiss Development Cooperation and the InterAmerican Development Bank for this research, and institutional supports from IFPRI, the
Inter-American Institute for Cooperation in Agriculture, and the University of Maryland.
We are grateful to the International Center for Tropical Agriculture for providing access
to census and other secondary data for Honduras. We are especially grateful to the rest of
the study team —Fernando Mendoza, Carlos Duarte, Juan Manuel Medina, Oscar
Neidecker-Gonzales, and Roduel Rodriguez — and to the many farmers and others in
Honduras who generously agreed to respond to our many questions.
2
access to markets and infrastructure, the development of local markets and institutions,
the nature and fragility of local resources, and the local incidence of government policies
and programs (Scherr et al. 1996; Lopez 1998).
Accounting for such a complex set of factors and their impacts on resource
management in the diverse circumstances of hillside areas is a major challenge for policy
research. To address this challenge, we employ the concept of “pathways of
development.” A development pathway represents a common pattern of change in
resource management, associated with a common set of causal and conditioning factors.1
For example, two of the pathways in central Honduras identified in this study are
expansion of basic grains production in areas far from roads and the urban market, and
intensification of horticultural production in areas close to roads and the urban market.
The causes and consequences of such pathways are likely different, and the
opportunities and constraints affecting natural resource management decisions likely also
differ across development pathways. For example, labor-intensive approaches may be
more effective in intensive basic grains production systems than in extensive grain
systems or in areas with substantial non-farm employment opportunities. Capitalintensive technologies are likely to be more effective in more commercialized pathways.
The policy and institutional requirements of sustainable development will depend upon
the types of technologies that are appropriate in different development pathways. The
appropriate technology and policy strategy to achieve sustainable development thus may
differ across development pathways.
1
This concept was inspired by the work on comparative patterns of economic
development by Morris and Adelman (1988).
3
The principal research questions addressed by this study are:
1. What are the major pathways of development in the central region of
Honduras, their causes, and implications for agricultural productivity, natural
resource sustainability, and poverty?
2. How can policies and technologies facilitate more productive, sustainable, and
poverty reducing pathways of development in this region?
2. CONCEPTUAL FRAMEWORK AND METHODOLOGY
CONCEPTUAL FRAMEWORK
The conceptual framework for this study is shown in Figure 1. In this framework,
pressures from population growth, markets, new technology or other external factors
induce change in local markets, prices and institutions within individual communities.
These shifts are conditioned by initial community characteristics, such as their human
and natural resource endowments, infrastructure, market linkages and local knowledge
base. Local-level conditions and changes may induce different patterns of change in
economic livelihood strategies--which we refer to as development pathways--resulting
largely from differences in comparative advantage.
Across and within development pathways there may be differences in agriculture
and natural resource management strategies at both household and collective levels, such
as land use, land investment, land management intensity, input mix, conservation
practices, and collective action. Within a given development pathway, these responses
may be affected by many of the same factors that determine which pathway is being
4
pursued (e.g., population pressure, access to market or technology), leading to both direct
and indirect impacts of such factors. For example, increased access to markets may
increase the likelihood that a community will produce high-value perishable vegetable
crops, which will tend to favor use of modern inputs. Beyond this, increased market
access may increase use of modern inputs directly by reducing farm level costs of inputs
or increasing farmers’ awareness of such inputs. It is important to consider both the
direct and indirect impacts of such factors on responses and outcomes.
Collective and household level responses affect outcomes, including natural
resource conditions, economic conditions and human welfare. These outcomes in turn
can have feedback effects on local conditions, institutions, development pathways, and
natural resource management (NRM) decisions.
Public policies can influence the process at various levels. For example,
agricultural research and price policies affect shift factors. Resource regulations and
infrastructure investment affect community conditions. Land titling and credit programs
affect local markets and institutions. Technical assistance influences response patterns.
Nutrition programs or state forest management directly affect outcomes. Policy research
must consider which types and sequences of policy action are likely to be most effective
in different circumstances.
RESEARCH HYPOTHESES
Several general hypotheses linking the driving forces of change to the nature and
outcomes of change can be suggested:
5
Figure 1 Conceptual framework
POLICY ENVIRONMENT AND INTERVENTIONS
Collective
Response
Pressure/Shift
Factors t
t
o... n
- Change in
Market/Prices
- Change in
Population
- Change in
Property Rights
- Change in
Technology
Community Baseline
Conditions t
o
- Natural Resource
Endowment and
Distribution
- Human Resource
Endowment and
Distribution
- Market Integration
- Local Institutions
t o... tn
Local Markets
and
Institutions
- Investment
- Management
- Self-Regulation
t o... t n
Household
Response
Productivity
t o... t n
Natural Resource
Conditions
t o... t n
Human Welfare
t o... t n
t o... tn
- Investment
- Management
- Land Use
Source: Scherr et al. 1996
1. Impacts of Population Growth
Population growth induces expansion of production into more marginal and
fragile lands, particularly where unsettled land is available and property rights are not
well enforced. Where land is limited, intensification of labor per unit of land will occur,
including shortening of fallow cycles and adoption of more labor intensive products and
practices (Boserup 1965). These changes may increase land productivity but, holding the
state of technology and market development constant, may reduce labor productivity
6
(Salehi-Isfahani 1988) and per capita income (Pender 1998).2 Effects on resource
management and resource conditions may be mixed. Expansion into marginal lands can
cause deforestation and land degradation, particularly where property rights are not well
defined; while population growth induces labor intensive land improvements where land
tenure is secure (Scherr and Hazell 1994; Tiffen et al. 1994). In addition, population
growth may promote institutional changes such as development of individual property
rights that contribute to improved resource management (Boserup 1965). Thus, there
may be a U-shaped relationship between population density and resource conditions, with
resource conditions first worsening and later improving as population rises (Scherr and
Hazell 1994; Pender 1998).
2. Impacts of Improved Market or Technological Opportunities
Increases in the profitability of agricultural products, whether resulting from
infrastructure investment, market development, changes in market prices, technological
innovation, or government policies affecting these factors, will promote expansion of
agriculture into marginal areas if the costs of productive factors are unaffected by the
change (Angelsen 1996). However, if the costs of factors rise, a reduction in agricultural
area is possible as productive factors are concentrated on the most profitable lands (ibid.).
If expansion of agricultural land is limited, increased profitability will cause
intensification of labor and/or capital per unit of land, though the effects on capital
relative to labor depend on the nature of factor markets and the nature of the change.
2
Krautkraemer (1994) offers a contrary view of the implications of populationinduced intensification for labor productivity, based on assuming that the production set
is non-convex.
7
Market or technology development may promote a shift to cash crops, and will tend to
increase farm incomes (unless offset by falling farm prices). The implications for
resource management and environmental conditions may be mixed. For example,
changes in commodity prices have a theoretically ambiguous effect on soil conservation
investments (LaFrance 1992; Pagiola 1996). Market or technology development may
increase externalities associated with demand for water and agricultural chemicals.
3. Pathways of Development
Although many pathways of change are theoretically possible, we hypothesize
that a relatively small number of pathways represent much of the variation in the
processes of change occurring in hillsides of Central America. These pathways are
expected to be determined primarily by differences in comparative advantage, which is
largely determined by agricultural potential, market access and population density
(Pender, Place and Ehui 1998). More commercially oriented pathways (such as intensive
production of vegetables) are expected to be found closer to urban markets, or where a
comparative advantage in high value non-perishable commodities exists (such as coffee
and some forest products). In areas of high market access, nonfarm development is also
likely to be important. In areas more remote from markets and lacking comparative
advantage in high value commodities, subsistence production of cereals and livestock is
likely to continue to be important. Where population density is low, there may be greater
comparative advantage in more extensive cereal and livestock production and/or in
forestry activities. The problems, constraints, opportunities, and responses to change
likely differ substantially across pathways; thus different technology, policy and
8
institutional strategies are likely to be required for more productive, sustainable, and
welfare-improving development to occur in different pathways.
METHODOLOGY
This study is based primarily on a survey of communities in the central region of
Honduras. The central region, defined to include 31 municipios (counties) in the
departments of Francisco Morazan and El Paraíso, was selected as the study region due to
the predominance of hillsides, the presence of several important watersheds, concerns
about poverty and resource degradation in the region, and the presence of the capital city
of Tegucigalpa and major road networks within the region (Map 1). A stratified random
sample of 48 of the 325 aldeas (villages) in the region was selected for the survey. The
stratification was based upon 1974 rural population density of the municipio in which
each aldea was located and the distance by road of the municipio county seat to
Tegucigalpa.3
To identify the pathways of development, we adopted a simple classification
system based on the primary and secondary occupations in the communities in 1996 and
changes since the mid-1970s. Six pathways were identified (these are described below).
In all but one of the sample communities, basic grains production was the primary or
secondary occupation, so this could not distinguish the communities. The other major
occupation (whether primary or secondary) was used to classify communities into five
types (basic grains and livestock, horticulture, coffee, forestry, and non-farm employment
communities). The basic grains and livestock communities were further disaggregated
3
(1997).
Details on the study sample and the survey are available in Pender and Scherr
9
10
between those where basic grains production had been increasing in importance (“basic
grains expansion”) and those where basic grains production had been constant or
declining (“basic grains stagnation”).
Although the classification for the other development pathways was based only
on occupational data in 1996, we found in most cases that the dominant non-grain
occupations in each pathway had been increasing in importance since the mid-1970’s.
For example, coffee production had increased in importance in all coffee communities
(and coffee area in almost all), and had declined in none of them. Thus, we chose to label
these communities as representing “coffee expansion”. We found similar results for the
other pathways, with the dominant non-grain occupation increasing or remaining constant
in importance in all cases.4
The empirical analysis was based upon comparisons of descriptive statistics
across the pathways, econometric analysis, and qualitative information from the survey.
The general form of the econometric analyses, based on the conceptual framework in
Figure 1, is as follows:
1.
4
Pathway = f (driving factors, conditioning factors, policies/programs)
We considered alternative methods of classification, such as using principal
components or cluster analysis. In a previous paper, Pender and Duron (1996) identified
pathways of development based on principal components analysis of secondary data on
causal and conditioning factors, responses and outcomes. We opted not to use that
approach in this paper because 1) many of these factors were measured as discrete or
ordinal variables and thus not appropriate for principal components analysis; and 2) we
wanted to test hypotheses about the causes and implications of the pathways, so did not
want to include such factors in the classification. We were also concerned about the
sensitivity of cluster and principal components analysis to arbitrary decisions such as the
number of variables included and their scaling, and felt that a simpler, more transparent
classification would be more useful.
11
2.
Household level responses = f (pathway, driving factors, conditioning
factors, policies/programs)
3.
Collective responses = f (pathway, driving factors, conditioning factors,
policies/programs)
4.
Outcomes = f (pathway, driving factors, conditioning factors,
policies/programs)
Many of the dependent variables in this system are measured as discrete variables,
including pathways (categorical), household responses (ordinal)5, collective responses
(binary), and some of the outcome variables (ordinal). Several different types of
regression models are used, depending on the nature of the dependent variable.6 In some
regressions the dependent variable is measured at a single point in time (e.g., use of
conservation practices in 1996), and in some the dependent variable is measured as a
change (e.g., perceived change in resource conditions since the mid-1970s).
The explanatory variables used to explain determinants of pathways include
factors affecting agricultural potential (altitude and number of rainfall days), population
density, access to markets (distance to the urban market and to the nearest road), and
access to technology (the presence of a technical assistance program). The crosssectional regressions for household responses and outcomes include as explanatory
variables the pathways, population density and density squared (the squared term
5
Most of the household response variables (such as adoption of various
conservation practices) were measured by an ordinal index from 0 to 6: 0 = no
households used the practice, 1=fewer than 10%, 2=minority, 3=about half, 4=majority,
5=more than 90%, 6=all.
6
For example, we use multinomial logit to estimate determinants of the pathways,
binary probit to estimate determinants of collective action, and ordered probit to estimate
determinants of household responses and some outcome variables.
12
included to check for possible U-shaped relationships), the distance to the urban market
and to the nearest road, adult literacy, and the presence of programs of technical
assistance, credit, agrarian reform, and land titling.7 In the regressions for changes, we
replaced population density and density squared with the change in these measures
between 1974 and 1988 (two census years), and replace the distance measures of market
access with indicators of whether a road has existed in the community since the 1970s, or
one was constructed since then.
Since the survey was conducted at a community level, we lacked data on intravillage variation in factors that may have affected measures of household level responses
and outcomes, such as land distribution or plot quality. This is an inherent limitation of
using community survey data to analyze household behavior and may introduce problems
of omitted variable bias, if the unobserved intra-village factors are correlated with the
included village-level explanatory variables. The only solution to this problem is to
collect household level data, which was not within the scope of the present study.
Endogeneity of some explanatory variables—particularly population growth and
the pathway variables—may be a problem. In all regressions including these explanatory
variables, we ran the regressions twice, using actual and predicted values of these
variables, to test the robustness of the findings.8 We discuss which results are significant
and robust below.
7
In regressions explaining changes in resource conditions, we also included
efforts to regulate forest use by the national or municipal governments.
8
In regressions using predicted values, the standard errors should be adjusted to
account for the additional error induced by this. Analytical formula for the appropriate
standard errors in a two stage model where the first stage is multinomial logit and the
second stage (in many of the regressions) is ordered probit have not been derived in
previous literature, as far as we have been able to determine. An alternative is to use
13
Given that the pathways of development are influenced by some of the same
factors that influence responses and outcomes directly (e.g., population pressure and
market access), these factors may have indirect as well as direct impacts on responses and
outcomes, via their impact on the pathways themselves. To address this, we estimate the
direct and indirect impact of marginal changes in such factors on various outcome
measures. We focus this analysis only on outcomes that are measured as continuous
outcome variables, since interpreting impacts on ordinal response variables is difficult.
3. PATHWAYS OF DEVELOPMENT IN CENTRAL HONDURAS
In five of the sample communities, horticultural production is the first or second
most important occupation, and horticultural area has been increasing since the mid1970s in all of these (Table 1). Horticultural area has not increased significantly in any
of the other sample communities. More than half of fruit and vegetable production is
commercialized only in this pathway. These five communities are classified as
representing “expansion of horticultural production.”
In 10 communities, coffee production is at least the second most important
occupation, and in all of these, coffee production has been increasing in importance.
More than half of coffee production is commercialized only in this pathway. These
communities are classified as representing “expansion of coffee production.”
In three communities, forestry activities rank first or second in importance as an
bootstrapping to compute the appropriate standard errors, but the small sample size (12
observations per stratum) made this option untenable. Thus, we do not report the results
of the two-stage regressions, but only use them to check the robustness of the findings.
14
occupation. Basic grains production has been decreasing in importance as an occupation
in all three communities, while forestry activities have increased or remained constant in
importance. There is significant commercialization of timber or lumber only in this
pathway, although commercialization of pine resin occurs in other pathways. These
communities are classified as representing “specialization in forestry.”
In 10 communities, off-farm employment is the first- or second-ranked
occupation. Off-farm employment has been increasing in importance in almost all of
these communities, and in no case has it declined. Commercialization of most
agricultural products is low in this pathway. These communities are classified as
representing the “nonfarm employment” pathway.
In the remaining 20 communities, basic grains production is the most important
occupation, and in nearly all cases, livestock production is the second most important. In
five of these communities, the importance of basic grains production has been increasing.
Basic grains and livestock production are most commercialized in this pathway. These
five communities are classified as representing “basic grains expansion.”
In the remaining 15 communities, the importance of basic grains production has
been constant or declining. Commercialization of basic grain and livestock production is
lower in these communities than in the basic grains expansion communities, though still
significant sources of income. These are classified as “basic grains stagnation”
communities.
15
Table 1 Development pathways in Central Honduras
Variable
Number of communities in
sample
Sample represents:
- % of population of region
- % of area of region
Dominant economic activities
- Basic grains
- Livestock
- Horticultural crops
- Coffee
- Forestry
- Nonfarm employment
Basic
grains
expansion
Basic
grains
stagnation
Horticultural
expansion
Coffee
expansion
Forestry
specialization
Nonfarm
employment
5
15
5
10
3
10
5
6
20
15
5
4
31
34
11
22
28
20
x
x
x
x
x
x
x
x
x
x
x
x
Change in economic activities since 1975
- Basic grains
↑
- Livestock
0↓
- Horticultural crops
0
- Coffee
↑↓
- Forestry
0↓
- Nonfarm employment
0
Index of proportion sold outside communitya
- Basic grains
3.4
(0.3)
0↓
0
0↑
0
0↓
0↑
0↓
0
0↑
0
0↓
0↑
0↓
0↓
0↑
↑
0↓
0↑
↓
0↑
0↑
↑
0↑
0
-- Mean (robust standard error in parenthesis) -1.9
0.0
0.6
2.0
(0.2)
(0.0)
(0.3)
(0.0)
↑↓
0↓
↑↓
0↓
↑↓
0↑
1.1
(0.3)
- Cattle/cattle products
2.9
(0.8)
2.6
(0.5)
2.7
(0.9)
1.4
(0.5)
2.0
(1.4)
2.2
(0.8)
- Vegetables
0.0
(0.0)
2.0
(0.4)
4.0
(0.0)
1.9
(0.7)
0.0
(0.0)
1.1
(1.0)
- Coffee
1.1
(0.5)
0.0
(0.0)
0.0
(0.0)
4.0
(0.0)
2.0
(1.4)
1.4
(1.1)
- Pine resin
6.0
(0.0)
0.0
(0.0)
0.0
(0.0)
6.0
(0.0)
6.0
(0.0)
6.0
(0.0)
a
Values of index: 0 = none, 1 = less than 10%, 2 = minority, 3 = half, 4 = majority, 5 = more than 90%, 6 =
all
CHARACTERISTICS OF THE PATHWAYS
Various indicators of determining factors, responses, and outcomes are shown in
Appendix Table 1 for the different pathways and the region as a whole. The basic grains
16
pathways are at lower altitude and have lower rainfall than the other pathways.
Population density was lowest in the forestry, horticultural and basic grains expansion
pathways in the mid-1970s, but rapid population growth in the horticultural pathway
caused this pathway to be among the most densely populated by the late-1980s. Access
to the urban market and roads is lowest in the basic grains pathways and highest in the
nonfarm employment and horticultural pathways. In the case of the horticultural
pathway, this has been a result of road construction since the mid-1970s. Access to
technical assistance and other government programs has been lowest in the basic grains
expansion pathway and generally highest in the coffee and forestry pathways. Literacy is
lowest in the basic grains and horticultural pathways.
Land use change has been moderate overall since the 1970s, but substantial
changes are apparent in a few pathways. The area in annual crops grew in all pathways,
but most rapidly in the horticultural and basic grains expansion pathways (Table 2). The
expansion in annual crops occurred at the expense of tree cover in all pathways, but
especially in the basic grains expansion and horticultural pathways. The area of devegetated land increased in the basic grains expansion pathway but declined in the others.
Changes in area of pasture and fallow were generally relatively small, with the exception
of fallow in the nonfarm employment pathway, which has increased significantly since
the 1970s. Survey respondents attributed land use changes to population growth and
burning practices (especially in the basic grains communities), increased knowledge
about land management options and improved market access (especially in the
horticultural pathway), increased profit potential of commercial crops (especially in the
coffee pathway), over-exploitation of forests by loggers and community members (in the
17
forestry pathway), poor soils (forestry pathway), decline in farming (nonfarm
employment pathway), and other factors.
Table 2 Land use change from aerial photographs: 1970s - early 1980sa
Mean change in area as percentage of total area
Land use
Basic grains
expansion
Tree coverb
-12.5
(--)
-4.2
(1.4)
-1.3
(3.8)
-0.8
(--)
-1.6
(5.6)
Cultivated
4.3
(--)
7.8
(4.3)
2.3
(2.1)
2.1
(--)
2.9
(2.1)
Fallow
-0.3
(--)
-0.02
(0.4)
1.5
(0.9)
-0.9
(--)
3.3
(2.0)
Pasture
0.7
(--)
-1.4
(0.7)
-1.4
(0.6)
-0.7
(--)
0.03
(0.7)
Devegetatedc
7.7
(--)
-2.4
(3.0)
-1.1
(2.6)
0.2
(--)
-4.8
(5.0)
1
4
7
1
10
No. of cases
Horticultural
Coffee
Forestry
expansion
expansion
specialization
(Standard error in parentheses)
Non-farm
employment
a
Based on comparison of aerial photographs at a scale of 1:40,000 or less from the 1975-1995 period and
from the 1990-1995 period. Recent photos were available for only 23 of the communities. Different dates
apply to different communities.
b
Tree cover includes land covered by forest as well as coffee, fruits, or other trees. No attempt was made
to distinguish different tree types or forest density.
c
Area with very little vegetation; sometimes recently deforested (with evidence of tree stumps) or with
small shrubs.
Crop production practices vary substantially across the pathways. Continuous (no
fallow) crop production is most common in the forestry pathway, despite low population
density, probably due to limited availability of arable land. Continuous cultivation has
increased in all pathways, but especially in the forestry and coffee pathways. Burning to
18
prepare fields is most common in the basic grains pathways, but has declined
significantly in all pathways. Crop production is most technified in the horticultural
pathway, in which use of chemical fertilizer, insecticides, herbicides, improved seeds and
irrigation is most common. Use of modern inputs has increased in most communities
since the mid-1970s, but especially in the horticultural pathway.
The most common livestock are poultry, pigs and cattle. While chickens are
common in all pathways, pigs are most common in basic grains communities and cattle in
forestry and basic grains expansion communities. Use of all types of livestock has
generally declined, though declines were more common outside of the basic grains
pathways.
Use of organic inputs and annual conservation practices are generally uncommon
in all pathways. The most common annual conservation practice is contour planting,
which is practiced by significant fractions of farmers (but less than half) mainly outside
the basic grains pathways. The most common land improving investments are live
barriers, stone walls, tree planting and terraces; all of which have been adopted by less
than half of households. Adoption varies by type of investment and pathway: live
barriers and terraces are most common in the forestry and coffee pathways, stone walls
most common in the forestry and basic grains stagnation pathways, and tree planting
most common in basic grains expansion and coffee pathways.
Local collective action related to natural resources includes formal or informal
regulation of use of common property resources, such as forests and water, and collective
investments, such as planting trees near water sources or building stone walls to control
runoff. Local regulation of common resources is strongest in the forestry and coffee
19
communities, where external institutional presence has been the greatest.9 Collective
investments in tree planting and runoff control have been greatest in the basic grains
expansion and nonfarm employment pathways, and no such investments were reported in
any of the horticultural or forestry communities.
In terms of outcomes, maize yields are lowest in the basic grains pathways, but
improvement in maize yields is more common in the basic grains expansion pathway
than other pathways. Agricultural wages and wage improvement have been lowest in the
basic grains pathways, and real wages have fallen in the basic grains stagnation pathway
since the mid-1970s. Measures of poverty are high but declining in all pathways, with
the greatest improvement occurring in the nonfarm employment pathway. Deforestation
on steep slopes is greatest in the basic grains stagnation pathway, while there is also
significant cultivation on steep slopes in the horticultural pathway. Perceptions of
changes in resource conditions (cropland quality, forest area, forest quality, water
availability and water quality) indicate a decline in most measures in all pathways, but
resource degradation is most common in the basic grains and forestry pathways.
These indicators suggest that the pathways have distinct causes or conditioning
factors, and different implications for resource management decisions, agricultural
productivity, poverty and resource sustainability. Below we investigate these
relationships using econometric analysis, controlling for the explanatory factors
mentioned previously.
9
In Pender et al. (1998b), we report a positive association between the presence
of external organizations and local organizational development in the study communities.
20
DETERMINANTS OF THE PATHWAYS
A multinomial logit model was used to investigate the determinants of the
pathways (Table 3). The results (together with the descriptive statistics discussed above)
imply that different determining and conditioning factors are critical for different
pathways. In general, the factors determining comparative advantage are important
distinguishing factors. Agroecological characteristics, especially altitude and rainfall,
distinguish the basic grains pathways from most other pathways, while horticultural
communities are at higher elevation than others. The basic grains expansion
communities have lower population densities, are more remote from roads and markets
and are less well served by external programs than all other pathways. Horticultural
communities have relatively good access to roads and markets, but this access has come
relatively recently, in contrast to most other communities outside the basic grains
pathways. Technical assistance does not appear to have been a major driving factor
behind the basic grains expansion or horticultural pathways; if anything, there are
opportunities for more external presence in this pathway. By contrast, external presence
appears to have been important in the coffee and forestry communities. Forestry
communities are also distinguished from the other pathways by their low population
density, relatively good access to markets, and high presence of forestry department
officials. Nonfarm employment communities have the best access to the urban market,
roads and public services.
21
Table 3 Determinants of the pathways
Multinomial logit regressionsa
Coefficient
Basic grains
expansion
Horticultural
Coffee
Forestry
expansion
expansion
specialization
(Standard errors in parentheses)b
Nonfarm
employment
Altitude at mid-point (m.a.s.l.)
-0.0187
(0.0127)
0.0187***
(0.0052)
0.00807
(0.00500)
0.00307
(0.0105)
0.00191
(0.00303)
Average number of rainfall days per year
0.1300**
(0.0627)
0.0499
(0.0308)
0.1556**
(0.0673)
0.1360
(0.0963)
0.0657***
(0.0210)
1974 population density (persons/km2)
-0.2263*
(0.1142)
-0.1429**
(0.0626)
0.0049
(0.0495)
-0.1269
(0.1870)
-0.0236
(0.0190)
Distance to Tegucigalpa (km)
-0.0969*
(0.0511)
-0.0605
(0.0557)
0.1810
(0.1111)
0.1425
(0.1461)
-0.0469
(0.0327)
Distance to nearest road (km)
1.133***
(0.347)
-59.03***
(4.76)
-0.975**
(0.450)
-0.7625
(0.4728)
-69.86***
(3.15)
Presence of technical assistance program
Mean predicted probability of correct
pathwayc
-35.35***
(10.67)
0.83
-50.85***
(9.02)
0.71
15.89
(20.68)
10.08
(35.71)
0.68
0.14
-53.22***
(9.20)
0.79
a
Reference category is basic grains stagnation pathway. Only 47 observations were used due to missing data on population
density for one community.
b
Coefficients and standard errors are adjusted to account for sampling weights, stratification, and finite population size.
Intercept not reported.
c
For example, the predicted probability that a community is in the basic grains expansion pathway is 0.83 for basic grains
expansion communities. The mean predicted probability of the correct pathway for basic gains stagnation communities is
0.71.
Note: *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level.
4. AGRICULTURE AND NATURAL RESOURCE MANAGEMENT
In this section we consider the factors determining the patterns and changes in
agriculture and natural resource management, including household level crop production
practices, conservation measures and collective action to manage common resources.
22
CROP PRODUCTION
Econometric analysis reveals significant variation among the pathways in crop
production practices in 1996, even after controlling for other factors such as population
density, access to markets, technical assistance and other programs (Table 4). Consistent
with the descriptive statistics, use of continuous cropping is most common in the forestry
pathway, while burning is most common in the basic grains stagnation pathway. Use of
fertilizer, herbicides, improved seeds, and irrigation are all more common in the
horticultural pathway than most other pathways.
Population density has a significant impact only on pesticide use. The impact is
positive, with higher population density associated with increased use, but at a
diminishing rate.10 Somewhat surprisingly, continuous cropping, insecticides and
herbicides are more common further from Tegucigalpa. Burning to prepare fields is more
common further from a road (probably because it is illegal), while use of fertilizers,
insecticides and improved seeds are more common close to a road, as one would expect.
Technical assistance programs are associated with less burning and greater (i.e., more
common) use of insecticides and herbicides. Agrarian reform programs are associated
with greater use of burning (probably because smallholder beneficiaries are oriented
towards basic grain production), fertilizer and irrigation, while land titling programs are
associated with greater use of fertilizer, herbicides and improved seeds. Literacy is
associated with greater use of burning, possibly because the opportunity cost of labor is
greater where literacy is higher.
10
The relationship reaches a maximum at a population density of 100 persons per
km , which is greater than the maximum population density of communities in the
sample. The same is true in all subsequent analyses in which the coefficients of
population density and population density squared are statistically significant.
2
23
Table 4 Determinants of cropping practicesa
Continuous
Cropping
Burning
Fertilizer
Insecticide
Herbicide
Improved
seeds
Irrigation
-.4258
(.7096)
-4.256 (R)
(.782)
4.175 (R)
(1.209)
2.847 (R)
(.690)
-.894
(.725)
-.470
(2.984)
-.992
(.836)
2.511
(1.123)
-4.435 (R)
(.889)
1.342 (R)
(.639)
2.186
(.906)
1.770 (R)
(.691)
5.678 (R)
(1.948)
1.951 (R)
(.785)
Coffee expansion
.292
(.723)
-2.129 (R)
(.580)
1.112 (R)
(.583)
.004
(.645)
-1.203 (R)
(.692)
2.179
(1.164)
-.215
(.617)
Forestry specialization
9.645 (R)
(.547)
-3.959 (R)
(.812)
.470
(.580)
-2.716
(.605)
2.341 (R)
(1.052)
-.242
(.620)
Nonfarm employment
1.603
(.735)
-2.686
(.606)
-.544
(.583)
1.653 (R)
(.653)
-.148
(.690)
Population density, 1988 (per km2)
-.0219
(.0218)
-.015
(.015)
-.0119
(.0155)
.0436 (R)
(.0162)
-.0547
(.0173)
Population density squared
.00015
(.00011)
.000033
(.000077)
.000027
(.000072)
Distance to Tegucigalpa (km)
.0274 (R)
(.0093)
.00949
(.00936)
Distance to nearest road (km)
-.066
(.091)
.224 (R)
(.077)
Technical assistance program
.740
(.729)
Credit program
.425
(.655)
Agrarian reform program
.030
(.482)
Land titling program
Variable
Basic grains expansion
Horticultural expansion
Proportion of adults literate, 1988
Number of observations
-1.694
(.588)
.670
(1.039)
-1.174
(.631)
.0206
(.0216)
.0156
(.0153)
-.000245 (R)) .000241
(.000087)
(.000081)
-.000110
(.000107)
-.000100
(.000068)
-.00502
(.00828)
.0251 (R)
(.0089)
.0321 (R)
(.0113)
-.00319
(.00761)
.00721
(.00927)
-.296 (R)
(.058)
-.186 (R)
(.065)
.131
(.082)
-3.793 (R)
(1.555)
-.0094
(.0731)
-2.404 (R)
(.602)
.048
(.631)
2.179 (R)
(.629)
1.508 (R)
(.560)
.484
(1.099)
-1.068
(.645)
-1.839
(.544)
.698
(.460)
.586
(.534)
-.315
(.450)
-.754
(.881)
-.573
(.459)
1.836 (R)
(.482)
.732 (R)
(.390)
.687
(.419)
.434
(.387)
-1.468
(.890)
.372
(.561)
.620
(.391)
1.021 (R)
(.383)
.389
(.488)
1.407 (R)
(.538)
.467
(2.107)
2.57 (R)
(1.22)
-2.057 (R)
(1.163)
1.794
(1.311)
45
45
47
46
-.817
(1.468)
46
2.419 (R)
(.923)
1.535
(2.633)
1.031 (R)
(.500)
.160
(.537)
1.169
(1.203)
47
46
a
Based on ordered probit regressions with coefficients and standard errors adjusted for sampling weights, stratification, and
finite population. Dependent variables take integer values from 0 (no one uses practice) to 6 (everyone uses).
Note: (R) indicates that the coefficient is statistically significant at the 5% level, and the same sign and significant at the 10%
level when pathways are replaced by predicted pathways.
The preceding results are based on cross-sectional regressions, and may be
affected by many possible sources of omitted variable bias.11 Fortunately, we are also
11
For example, the surprising positive association between distance to the urban
market and adoption of some purchased inputs may be due to omitted land quality or
climate characteristics that are correlated with distance to the market.
24
able to investigate determinants of change in some cropping practices, and these
regressions will be less affected by omitted variables (at least those which do not change
over time).12 These results are presented in Table 5. Some of the results explaining
changes in practices confirm results in the cross-sectional analysis. We find that burning
has declined more commonly in the horticultural and nonfarm employment pathways
than in the basic grains stagnation pathway. Population growth has a positive but
diminishing impact on insecticide use. Road access and land titling programs contribute
to increasing use of insecticides. Technical assistance is associated with reduced burning,
while agrarian reform programs are associated with increased burning. However, some
of the factors that have a significant impact in the cross-sectional regressions have an
insignificant impact in the analysis of changes, reducing our confidence in those crosssectional results. In some cases, we find significant impacts in the analysis of changes
that we did not find in the cross-sectional analysis. For example, we find that increase in
continuous cropping is more common in the basic grains expansion than the basic grains
stagnation pathway (possibly because continuous cropping was initally already more
common in the latter), and more common where technical assistance has operated. Such
impacts could easily be missed in cross-section regressions.
12
Unfortunately, our measures of change are relatively rough, in most cases only
indicating whether a practice had increased, remained constant, or decreased. For some
types of practices, such as use of fertilizers and herbicides, it was so common for use to
increase that regression analysis was not possible. For other practices, such as use of
improved seeds and irrigation, changes were so uncommon that regression analysis was
not possible.
25
Table 5 Determinants of change in cropping practicesa
Variable
Continuous
cropping
Burning
Insecticide
Basic grains expansion
2.247 (R)
(1.062)
-1.521
(.976)
2.575
(1.246)
Horticultural expansion
.729
(1.066)
-4.229 (R)
(1.307)
.713
(3.972)
Coffee expansion
1.314
(.852)
-.939
(.748)
.326
(1.053)
Forestry specialization
.533
(.774)
-2.473
(.922)
.961
(.974)
Nonfarm employment
.261
(.505)
-2.520 (R)
(.714)
-.725
(.492)
Change in population density, 1974-88
(per km2)
.0401
(.0266)
.0128
(.0296)
.0409
(.0319)
Change in (population density squared)
-.000131
(.000173)
-.000222
(.000164)
-.000337 (R)
(.000161)
-.970
(1.114)
2.003
(1.276)
Already had road access in 1975
-.435
(.717)
.960
(.688)
Technical assistance program
2.295 (R)
(.775)
-1.482 (R)
(.622)
1.921
(1.343)
Credit program
.188
(.574)
-2.150 (R)
(.520)
-.842
(.612)
Agrarian reform program
1.016
(.602)
2.470 (R)
(.505)
1.510
(.558)
Land titling program
.115
(.659)
.254
(.395)
12.13 (R)
(1.39)
.343
(1.738)
-.224
(1.542)
-16.08 (R)
(8.77)
New road access since 1975
Change in proportion of adults literate,
1974-88
No. of observations
a
38
43
16.62 (R)
(3.17)
3.511 (R)
(1.037)
43
Based on ordered probit regressions with coefficients and standard errors adjusted for sampling weights,
stratification, and finite population. Dependent variables take values of +1 (increase), 0 (no change), and -1
(decrease). There was insufficient variation in the change in other cropping practices for regressions to be
feasible.
Note: (R) indicates that the coefficient is significant at the 5% level, and the same sign and significant at
the 10% level when pathways are replaced by predicted pathways and population growth replaced by
predicted population growth.
26
These results imply that changes in market access and technical assistance have
been important factors influencing intensification of crop production, both directly (i.e.,
controlling for the pathways) and via their influence on development of the pathways
themselves. Market access has had most influence on adoption of purchased inputs,
while technical assistance and other programs have affected the use of labor intensive
practices as well. Population growth appears to be of lesser importance in directly
causing these changes. However, by influencing the choice of development pathway,
differences in population density (and other initial conditions) still have an important
influence.
USE OF ORGANIC INPUTS AND CONSERVATION MEASURES
Different pathways are associated with different types of conservation measures.
Use of contour planting is more common in the horticultural and forestry pathways than
the other pathways, controlling for other factors (Table 6). Minimum tillage is most
common in the coffee pathway. Mulching is least common in the horticultural and
nonfarm employment pathways. Terraces are most common in the coffee and nonfarm
employment pathways, while live barriers are also most common in the coffee pathway
(Table 7). These differences suggest that the potential for adoption of particular
conservation measures varies across the pathways, and suggests the importance of a
targeted approach to promotion of conservation measures in different pathways.
Population density has an insignificant direct impact on most conservation or
organic measures, except use of cattle manure and tree planting. Both of these are more
common at higher population density, but at a diminishing rate. This may be because the
27
relative returns to such investments are higher at higher population densities, and/or the
labor costs of the measures are lower (particularly for manuring).
Table 6 Determinants of annual conservation or organic practicesa
Variable
Contour
planting
Minimum till
Mulching
Incorporation
of crop
Cattle manure
residues
.659
(.852)
-.469
(1.134)
.462
(.793)
18.69
(3.83)
.360
(.919)
-.154
(.835)
-2.302
(1.224)
Basic grains expansion
.158
(.844)
1.310
(1.528)
Horticultural expansion
3.60 (R)
(1.34)
-2.697
(1.449)
-3.094 (R)
(.908)
Chicken
manure
Coffee expansion
1.944
(.781)
1.721 (R)
(.791)
.200
(.557)
.654
(.602)
.815
(.565)
1.029
(.766)
Forestry specialization
1.634 (R)
(.774)
3.575
(.718)
-.359
(.712)
.106
(.905)
.254
(.713)
-8.037
(.571)
Nonfarm employment
.357
(.706)
.559
(.551)
-1.181 (R)
(.532)
.092
(.699)
-.258
(.707)
.635
(.663)
Population density, 1988
(per km2)
-.0214
(.0144)
-.0074
(.0187)
.0124
(.0176)
.0200
(.0165)
.0412 (R)
(.0170)
.0225
(.0182)
Population density squared
.000067
(.000071)
.000016
(.000090)
-.000045
(.000081)
-.000089
(.000068)
-.000181 (R)
(.000075)
-.000118
(.000082)
Distance to Tegucigalpa
(km)
-.0408 (R)
(.0096)
-.0562 (R)
(.0149)
-.0254 (R)
(.0102)
-.0307 (R)
(.0103)
-.0136
(.0136)
-.0323
(.0196)
Distance to nearest road
(km)
.1501
(.0820)
.397 (R)
(.119)
.0234 (R)
(.0874)
.183
(.104)
.0979
(.0813)
-15.96 (R)
(1.64)
Technical assistance
program
2.284
(1.024)
1.434
(.677)
-.951
(.489)
1.610 (R)
(.635)
.949
(.608)
-.975
(.799)
Credit program
-1.063
(.629)
1.007
(.655)
.504
(.429)
-.769
(.584)
-.851
(.524)
.832
(.807)
Agrarian reform program
-.505
(.527)
.841
(.492)
1.025
(.402)
.388
(.470)
.045
(.414)
.727
(.630)
Land titling program
1.098 (R)
(.452)
.116
(.655)
.464
(.503)
-.272
(.438)
.103
(.547)
-.513
(.529)
Proportion of adults
literate, 1988
5.379 (R)
(1.999)
-2.317
(1.909)
-2.929
(1.156)
3.467
(1.575)
3.09 (R)
(1.49)
-2.725
(2.086)
47
47
45
No. of observations
46
47
44
a
Based on ordered probit regressions with coefficients and standard errors adjusted for sampling weights, stratification,
and finite population. Dependent variables take integer values from 0 (no one uses practice) to 6 (everyone uses).
Note: (R) indicates that the coefficient is statistically significant at the 5% level, and the same sign and significant at
the 10% level when pathways are replaced by predicted pathways.
28
Table 7 Determinants of land improving investmentsa
Variable
Terraces
Live barriers
Stone walls
Tree planting
Basic grains expansion
1.452
(.716)
1.827
(.834)
-.182
(.683)
1.258
(.751)
Horticultural expansion
1.635
(.985)
.854
(.595)
1.491
(.852)
-.278
(.782)
Coffee expansion
2.057 (R)
(.884)
1.972 (R)
(.676)
.408
(.515)
.898
(.643)
Forestry specialization
2.589
(.723)
2.886
(.941)
1.270
(1.226)
-.350
(.605)
Nonfarm employment
2.104 (R)
(.726)
.322
(.485)
-.426
(.656)
-.312
(.618)
Population density, 1974b (per km2)
-.0263
(.0293)
.0465
(.0182)
.0255
(.0227)
.0531 (R)
(.0175)
Population density squared
.000302
(.000244)
-.000163
(.000128)
.000075
(.000131)
-.000307 (R)
(.000117)
Distance to Tegucigalpa (km)
-.0237
(.0167)
-.0118
(.0091)
.0002
(.0125)
.0163
(.0089)
Distance to nearest road (km)
.142
(.097)
-.1292 (R)
(.0623)
.0109
(.0667)
-.1001
(.0486)
Technical assistance program
2.227 (R)
(.563)
1.604 (R)
(.456)
1.531 (R)
(.494)
-.720
(.438)
Credit program
1.766 (R)
(.601)
-.580
(.437)
-1.372 (R)
(.595)
.280
(.427)
Agrarian reform program
1.140 (R)
(.415)
-.226
(.430)
Land titling program
1.078 (R)
(.541)
.817
(.456)
3.081 (R)
(1.289)
.028
(1.244)
Proportion of adults literate, 1974b
No. of observations
a
46
46
-.372
(.510)
-1.274 (R)
(.430)
2.274
(1.608)
46
.819
(.385)
-.687
(.462)
1.215
(1.284)
46
Based on ordered probit regressions with coefficients and standard errors adjusted for sampling weights,
stratification, and finite population. Dependent variables take integer values from 0 (no one uses practice)
to 6 (everyone uses).
b
1974 population density and proportion of adults literate used because some investments may have
occurred before 1988.
Note: (R) indicates that the coefficient is statistically significant at the 5% level, and the same sign and
significant at the 10% level when pathways are replaced by predicted pathways.
29
Several conservation measures are less common further from the urban market,
including contour planting, minimum tillage, mulching, and incorporation of crop
residues. This could be because the returns to such efforts are lower where returns to
agriculture may be lower (due to lower market access), but it also may be due to lower
intensity of involvement of programs promoting conservation in more remote areas.
Minimum tillage is more common further from a road, perhaps because of difficulties
plowing in more remote locations. Use of chicken manure and live barriers is more
common close to roads. The effect of road access on chicken manure (which is marketed
by commercial poultry operations) is similar to the effect of road access on inorganic
fertilizer use, as one might expect.
As one would expect, several conservation measures are positively associated
with technical assistance programs, including incorporation of crop residues, terraces,
live barriers, and stone walls. Credit, agrarian reform, land titling programs and literacy
are also positively associated with adoption of terraces. Credit and titling programs are
negatively associated with use of stone walls. Perhaps farmers feel less need to build
stone walls around their fields for tenure security if they have a title to land and access to
credit based on land ownership.
COLLECTIVE INVESTMENT IN NATURAL RESOURCE MANAGEMENT
Collective investment in tree planting and to control runoff is higher at moderate
rates of population growth than at low or very high rates.13 This may be because low and
high rates of population growth are associated with high rates of emigration and
13
These results are reported in Pender et al. (1998b). The probit model was not
estimable with the pathway variables included.
30
immigration, respectively; which may reduce the ability to achieve collective action by
reducing the stability and homogeneity of the community (Baland and Platteau 1996).
Higher initial population density reduces collective investments, possibly because the
number of required participants and hence organizational costs of collective investments
are higher where population density is greater. The presence of local organizations
involved in NRM appears to stimulate collective investments in NRM, while external
organizations are negatively associated with local collective investment. This latter
finding suggests that external organizations may undermine local collective efforts, and
implies that such organizations should take a cautious approach when intervening in local
communities, to avoid such displacement.
5. DEVELOPMENT OUTCOMES
The above patterns of resource management are the proximate causes of change in
the “critical triangle” of development outcomes: economic productivity or growth,
sustainability of the natural resource base, and poverty alleviation (Vosti and Reardon
1997). These outcomes thus also vary by development pathway and other factors.
PRODUCTIVITY
In the cross-sectional analysis, we did not find significant and robust differences
in land productivity (as measured by maize yields) among the pathways (Table 8). The
only factor found to be significantly associated with maize yields was population density,
with higher density associated with lower yields, though at a diminishing rate. This
suggests that population pressure is causing a decline in soil productivity, perhaps
31
Table 8 Determinants of outcomesa
Resource conditions, late 1970sc
Productivity, 1996
Variable
High maize yield
Basic grains expansion
Proportion of steep
land in forest
Proportion of steep
land de-vegetated
Proportion of steep
land cultivated
Proportion of
houses with
dirt floor
Proportion of
households whose
last child died
.117
(.130)
.292
(.169)
-.3698 (R)
(.1402)
.0684
(.0423)
.0562
(.0609)
.0209
(.0168)
Horticultural expansion
7.147
(9.205)
.515 (R)
(.160)
.320 (R)
(.149)
-.433 (R)
(.159)
.1670
(.0639)
.1062
(.0874)
.0113
(.0184)
Coffee expansion
5.503
(8.374)
.761 (R)
(.136)
.3354 (R)
(.0834)
-.3688 (R)
(.0816)
.0551
(.0357)
.0879
(.0807)
.0212
(.0146)
Forestry specialization
-3.641
(8.440)
.377 (R)
(.218)
.716 (R)
(.217)
-.524 (R)
(.216)
-.1193
(.0387)
.1528 (R)
(.0823)
.0133
(.0121)
Nonfarm employment
11.65
(8.00)
.630 (R)
(.103)
.255
(.156)
-.233
(.161)
-.0007
(.0290)
.0369
(.0702)
.0070
(.0192)
.00355
(.00156)
-.00282
(.00196)
.000292
(.000363)
-.0000161
(8.31e-06)
9.54e-06
(8.59e-06)
-1.50e-06
(1.62e-06)
Population density, (per km2)
-.659 (R)
(.243)
.00482
(.00412)
-.00916
(.00374)
.00630
(.00379)
Population density squared
.00285 (R)
(.00116)
-.0000175 (R)
(.0000178)
.0000591 (R)
(.0000218)
-.0000498 (R)
(.0000226)
Distance to Tegucigalpa (km)
.194
(.153)
.00730
(.00313)
-.00121
(.00524)
-.00011
(.00559)
.00090
(.00104)
.002153
(.001007)
-.000170
(.000210)
Distance to nearest road (km)
-1.838
(1.048)
-.0615 (R)
(.0149)
-.0091
(.0181)
.0264
(.0174)
-.01172
(.00572)
-.00506
(.00702)
.00237
(.00183)
Technical assistance program
-12.253
(7.512)
.1623
(.0803)
NE
NE
NE
.1091
(.0644)
.0006
(.0176)
Credit program
-3.079
(6.143)
-.0944
(.0880)
NE
NE
NE
-.0831
(.0585)
-.0203
(.0137)
Agrarian reform program
4.597
(8.110)
-.0482
(.0868)
NE
NE
NE
.0269
(.0470)
.0229 (R)
(.0081)
Land titling program
-3.814
(6.973)
.189
(.110)
NE
NE
NE
.0578
(.0519)
-.00535
(.00960)
-22.39
(15.87)
.473
(.371)
.410
(.361)
-.088
(.384)
-.1887 (R)
(.0915)
-.390 (R)
(.136)
-.0091
(.0424)
Proportion of land which is steep
(30% slope)
NE
NE
1.289
(1.199)
-.674
(1.218)
-.972 (R)
(.546)
NE
NE
No. of observations
45
46
37
37
37
47
47
Proportion of adults literate, 1988
a
-14.93
(9.18)
High male wageb
Poverty, 1988
Based on least squares regressions with coefficients and standard errors adjusted for sampling weights, stratification, and finite population. Intercept not reported.
The dependent variable is actually the natural logarithm of the daily wage in lempiras.
c
Population density and proportion of adults literate are for 1974 in these regressions, for 1988 in the other regressions. Programs are not included as explanatory variables because they were usually
present after the land uses were observed in aerial photographs.
Note: (R) indicates that the coefficient is statistically significant at the 5% level, and the same sign and significant at the 10% level when pathways are replaced by predicted pathways. NE means not
estimated.
b
32
Table 9 Determinants of change in outcomesa
Productivity, 1975-96
Variable
Maize yield
High male
wageb
Perceived natural resource conditions, 1975-96
Cropland
quality
Poverty, 1974-88
Water
availability
Water quality
Proportion of
houses with
dirt floor
-7.832
(.992)
-7.388 (R)
(.670)
NE
.0503
(.0680)
-.0009
(.0406)
2.385 (R)
(.949)
.806
(1.196)
-2.571 (R)
(1.004)
-.679
(1.628)
.1275
(.0867)
.0113
(.0508)
-.705
(.362)
-.203
(.307)
.0204
(.0482)
-.0029
(.0294)
-1.349
(.423)
-.679
(.284)
.0595
(.0471)
.0082
(.0362)
-1.039
(1.081)
.0142
(.0396)
-.0349
(.0196)
Forest area
Proportion of
households where
last child died
Basic grains expansion
.995
(.939)
-.688
(.485)
1.570
(1.133)
Horticultural expansion
1.792 (R)
(.984)
1.885 (R)
(.358)
2.670 (R)
(1.169)
Coffee expansion
.560
(.919)
.354
(.309)
1.404
(.958)
3.637
(.879)
1.960 (R)
(1.124)
Forestry specialization
-7.743
(.851)
.240
(.327)
-7.409
(.802)
.851
(.586)
-7.337
(.894)
Nonfarm employment
.130
(.719)
1.069 (R)
(.250)
.455
(.822)
1.674 (R)
(.674)
.468
(.798)
-.959 (R)
(.431)
Change in pop. Dens., 1974-88
-.0502
(.0363)
.0160
(.0074)
-.0501
(.0354)
-.0805 (R)
(.0247)
-.0272
(.0405)
.0414
(.0186)
.0962
(.0435)
.00352
(.00250)
-.00031
(.00225)
Change in (pop. Dens. squared)
.000035
(.000196)
-.000071
(.000026)
.000076
(.000204)
.00040 (R)
(.00013)
-1.51e-06
(.00026)
-.00022
(.00015)
-.00054
(.00035)
-7.79e-07
(.00002)
1.68e-06
(.00001)
New road access since 1975
-.103
(.844)
-2.707 (R)
(.631)
-.168
(1.015)
.226
(1.034)
Already had road access in 1975
-.589
(.628)
-1.443 (R)
(.375)
-1.077
(.874)
-1.569 (R)
(.538)
Technical assistance program
.092
(.626)
.146
(.172)
1.080
(.906)
Credit program
.146
(.649)
.275
(.214)
Agrarian reform program
-.699
(.788)
Land titling program
Change in proportion of adults literate,
1974-88
-9.289
(.873)
Forest quality
1.001
(.700)
1.128
(.779)
NE
-.151
(.087)
.0156
(.0732)
NE
1.343
(.663)
1.210
(.883)
-.0213
(.0415)
.0036
(.0390)
-1.029
(.587)
-.048
(.807)
-.377
(.581)
-.393
(.912)
.1452 (R)
(.0399)
-.0074
(.0199)
-.311
(.491)
-.718
(.556)
-.480
(.640)
.550
(.447)
-.826
(.991)
-.0435
(.0397)
-.0170
(.0229)
-.454
(.133)
-.248
(.766)
.989
(.703)
.222
(.833)
-.959
(.620)
-.0010
(.0348)
.0252
(.0212)
-.191
(.427)
.053
(.106)
-.312
(.573)
NE
.701
(.582)
1.357
(.620)
.0006
(.0455)
.0135
(.0267)
1.180
(1.427)
.867
(.576)
1.908
(1.496)
2.344
(2.301)
-.225
(2.287)
-2.812 (R)
(1.614)
-1.529
(2.695)
.214
(.162)
.1166
(.1235)
.906 (R)
(.258)
NE
NE
46
46
-1.247 (R)
(.567)
Presence of national or municipal forest
regulation
NE
NE
NE
-.205
(.668)
-.163
(.723)
No. of observations
45
36
44
45
44
a
142
-2.278 (R)
(.717)
1.008 (R)
(.398)
79
Based on ordered probit regressions for change in maize yield and change in resource conditions, and least squares for other regressions. Coefficients and standard errors adjusted for sampling weights,
stratification, and finite population. Dependent variables for ordered probit regressions take the values +1 (increase), 0 (no change), and -1 (decrease). Intercept not reported.
b
The dependent variable is actually the change in the natural logarithm of the daily wage in lempiras.
Note: (R) indicates that the coefficient is statistically significant at the 5% level, and the same sign and significant at the 10% level when pathways are replaced by predicted pathways and population
growth replaced by predicted population growth.
33
because of shortening fallow periods without sufficient use of organic or inorganic
sources of soil fertility. This hypothesis could not be confirmed in the regression
explaining changes in maize yields, since the population growth variables were found to
have an insignificant effect (Table 9). The only significant and robust finding in that
analysis was that growth in maize yields was more likely in the horticultural pathway
than in any others. This is consistent with the finding above of increasing use of modern
inputs in the horticultural pathway. Such inputs are not restricted only to horticultural
crops in this pathway, but are also used in maize production.
Labor productivity, as measured by wage rates, is highest outside the basic grains
pathways (Table 8). Real wages have also been growing more rapidly in horticultural
and nonfarm employment communities (Table 9). Wages are lower further from a road,
as one would expect. Surprisingly, however, we find that wage growth has been slower
in communities that have access to a road than in those that do not.
NATURAL RESOURCE CONDITIONS
Natural resource conditions vary substantially across the development pathways.
In the late 1970s, forests were most likely to be found on steep lands (with slopes over 30
percent) in the horticultural, coffee and forestry pathways (Table 8).14 De-vegetation of
steep lands is more common in the basic grains stagnation pathway than most other
pathways, controlling for other factors, as was found in the descriptive statistics. Higher
population density is associated with greater cultivation and less forest on steep lands, but
14
Cross sectional regressions explaining land use on steep slopes were not
possible for recent land use because of the limited number of communities for which
aerial photos were recently available (23).
34
at a diminishing rate. Literacy is associated with less cultivation on steep lands.
Perceived changes in natural resource conditions also varied across the development
pathways. Cropland quality was most likely to improve (or least likely to decrease) in the
horticultural pathway (Table 9). Forest area was most likely to increase in the
horticultural and nonfarm employment pathways. However, water availability was more
likely to decline in both of these pathways, as well as in the basic grains expansion
pathway, than in the other pathways. Other factors also affect resource degradation.
Population growth is associated with greater likelihood of deforestation, though at a
diminishing rate. Road access and land titling are both associated with deforestation, as
one might expect. Water quality was more likely to decline where there had been an
agrarian reform program, while the presence of national or municipal efforts to protect
forests was associated with improvement (or less reduction in) water availability and
quality.
POVERTY
By most available measures, socioeconomic conditions improved in all pathways
between 1974 and 1988. The measures that we examine through econometric analysis
include the percentage of houses with a dirt floor and the percentage of households whose
last child died.15 We do not find statistically significant and robust differences among the
pathways in these measures of poverty or changes in these measures, controlling for other
factors (Tables 8 and 9). The only factor that we find significantly associated with the
15
These measures are plausibly more affected by agricultural and resource
management decisions and productivity than other available measures from the census
data (such as availability of sanitation, water, and electricity), which are more related to
public services.
35
proportion of houses with a dirt floor is literacy (negative effect). We find that the
presence of an agrarian reform program is positively associated with child mortality,
while technical assistance programs are associated with an increase (or less decrease) in
the proportion of households having a dirt floor. These last two findings may be a result
of such programs focusing in areas where poverty is severe and perhaps worsening.
Overall, these results do not show strong differences across the pathways in terms
of changes in poverty, which has declined fairly broadly in the region. Perhaps public
and NGO investments in social services have been more powerful determinants of
changes in poverty than differences in comparative advantage and changes in agriculture
occurring across the different pathways of development.
DIRECT AND INDIRECT EFFECTS
As mentioned previously, factors such as population pressure and market access
may have both direct and indirect effects, by affecting which pathway of development
occurs (indirect) as well as by affecting natural resource management given the pathway
of development (direct). In Table 10, we present the predicted direct and indirect impacts
on outcomes of an increase in population density and in distance from a road, based on
the regression results presented in Tables 3 and 8. The predicted impacts of greater (by 1
person per square km.) population pressure are negative for maize yield and forest cover
on steep lands. In most cases, the direct and indirect effects are in the same direction, and
the direct effects are larger in magnitude. Surprisingly, wages and housing conditions are
predicted to improve with increased population pressure. However, since these results
are based on coefficients with low statistical significance, we cannot place high
36
Table 10 Predicted effects of population pressure and market access on outcomes
Productivity, 1996
Factor
Higher population
density (by 1
person/km2)
Further from road
(by 1 km)
Effect
Resource conditions, late 1970s
Percentage of Percentage of
steep land
steep land
in forest
de-vegetated
Percentage of
steep land
cultivated
Poverty, 1988
Percentage of Percentage of
houses with households where
dirt floors
last child died
Maize
yield
ln (male
wage)
(kg/ha)
(Lps/day)
-31.3
0.0037
-0.55
0.32
0.25
-0.22
0.02
6.6
0.0028
-0.37
0.29
0.03
-0.09
0.00
Total
-24.7
0.0065
-0.92
0.61
0.28
-0.31
0.02
Direct
Indirect
Total
-119.4
-366.2
-485.6
-0.0615
-0.0930
-0.1545
-0.91
2.52
1.61
2.64
-4.14
-1.50
-1.17
1.22
0.05
-0.51
1.01
0.50
0.24
0.35
0.59
Direct
Indirect
37
confidence in them. These results are more useful for understanding the general
magnitude of direct and indirect impacts implied by the coefficients, than in making
precise predictions.
The effects of poorer market access (by 1 km from the nearest road) are highly
negative in terms of productivity and poverty, but positive in terms of preserving forest
on steep lands. In contrast to the case of higher population growth, the indirect impacts
of changed market access are stronger than the direct impacts. Market access appears to
affect outcomes more via its impact on the pathways of development.
6. SUMMARY AND CONCLUSIONS
The general pattern of agricultural change in the central region of Honduras is
toward increased specialization and commercialization of production based upon
comparative advantage. Differences across the pathways in comparative advantage and
other factors led to differences in land use, agriculture and resource management
decisions. Given such differences, it is not surprising that we find also significant
differences across the pathways in their outcomes for agricultural productivity and
natural resource conditions. In general, productivity outcomes were more favorable in
the horticultural, coffee and nonfarm employment pathways, while the implications for
resource conditions were more mixed, with some resource conditions improving in these
pathways and others becoming worse.
While many natural resource and environmental problems were getting worse in
the central region, many aspects of human welfare were improving. In general, the
changes in agricultural productivity and resource conditions were more mixed, and more
38
often negative, than changes in measures of poverty. A major reason for this seems to be
that the factors influencing these different outcomes are different. The factors
influencing agricultural productivity and natural resource conditions are very locationspecific. By contrast, welfare conditions are largely affected by provision of services by
public agencies and NGOs, and these interventions were fairly broad in impact. In
addition, migration has tended to reduce differentials in wages across communities
(although large disparities remain), while general growth in economic opportunities
resulting from market liberalization and changes in market conditions have led to
increasing real wages in almost all pathways.
Although changes in natural resource conditions and human welfare are moving
in opposite directions in much of the central region, this does not imply that there must be
large tradeoffs between these objectives. As mentioned above, the differences in these
outcomes appear to result from different causal factors, suggesting that these causes can
be addressed separately.16 In some cases, direct tradeoffs do exist—for example,
increased use of irrigation and agrochemicals in the horticultural pathway leads to rising
productivity and incomes, but also to problems of water scarcity and contamination. In
these cases, careful consideration of the impacts on farmers’ welfare is needed when
considering measures to address the resource and environmental issues, and vice versa.
16
Of course, to the extent that public expenditures are required to finance both
efforts to improve public services and improvements in natural resource management,
there will be tradeoffs in the use of such expenditures.
39
FINDINGS RELATIVE TO THE RESEARCH HYPOTHESES
Population Pressure
The results of the regression analysis and the explanations provided by survey
respondents support the hypothesis that population growth induces deforestation and
expansion of agricultural area at relatively low levels of population density. Survey
responses also support the expected impact of population growth on reducing fallow
periods, though surprisingly this was not supported by the regression analysis. In general
we found few significant and robust impacts of population density or population growth
on agricultural and resource management practices. Agricultural change in central
Honduras appears to have been more market and technology induced than population
induced.
Population pressure was found to be associated with deforestation on steep lands,
as expected. This relationship had the hypothesized U-shape, although the predicted
turning point is at levels of population density above those found in central Honduras.
Population pressure was also associated with lower maize yields, contrary to the
prediction that population pressure induces more intensive land use and thus greater land
productivity (if not labor productivity). This result may be due to population-induced
land degradation.
Population pressure was not significantly associated with wages or measures of
poverty, perhaps because populations respond to differences in wages and poverty via
migration. Large differences in the rate of population growth across the pathways are
indicative of this. Thus the potentially negative welfare effects of population growth
appear to have been mitigated, or at least shared among communities to some extent, as a
result of migration.
40
Overall, these results support the concerns raised by many observers about the
negative implications of population growth for agricultural productivity and the
environment, although the impacts are not all negative or large. While we do find some
evidence that population-induced investments in land improvement are occurring, these
responses are not sufficient to compensate for all of the negative effects, particularly on
soil fertility and forest resources.
Market Access
Access to markets, as measured by road access, is associated with deforestation
and expansion of agricultural area. As expected, road access is also associated with
greater use of purchased inputs. The impacts of market access on changes in resource
conditions are mixed, as hypothesized. As mentioned above, deforestation is greater
where there is good road access. On the other hand, the use of burning is lower in such
areas, which will tend to lead to better resource conditions. Communities closer to the
urban market were more prone to adopt several conservation measures; however, this
may be a reflection more of access to information and technical assistance than access to
markets or economic opportunities. The impacts of road access on measures of welfare
are not as favorable as we expected to find. However, wage growth has been higher in
the horticultural and nonfarm employment pathways, indicating that the indirect effects
of market access (as a determinant of these pathways) must also be considered. Road and
urban market access did not have significant direct effects on measures of poverty.
Overall, these results confirm our expectations that market access should lead to
intensified use of agricultural inputs, while having mixed impacts on natural resource
conditions. The findings do not show the impacts of market access on changes in wages
41
and poverty that we expected, but this may be because these effects are mitigated by
migration.
Access to Technology
The presence of technical assistance programs in agriculture has contributed
significantly to several aspects of agricultural intensification and natural resource
management, including reduced use of fallow and burning and adoption of many
conservation measures. We found no significant impacts of technical assistance on most
outcome measures. Overall, the results suggest that technical assistance has been most
effective in promoting more labor intensive practices, but that these changes have not
produced large measurable impacts on agricultural productivity, resource conditions, or
poverty.
Pathways of Development
As hypothesized, we found that a relatively small number of development
pathways exist in central Honduras and that these are largely determined by factors
affecting comparative advantage, including agricultural potential, market access and
population pressure. Basic grains and livestock production dominate in lower rainfall
and more remote areas, while commercially oriented farming and non-farm employment
activities are more common in areas closer to roads and the urban market. Higher
population pressure distinguishes the basic grains expansion communities, where basic
grains productivity and production has been increasing, from the basic grains stagnation
communities, where productivity has been declining.
42
We found substantial differences in agriculture, resource management practices
and outcomes across the pathways. Agriculture is most labor intensive in the forestry
pathway, where arable land is scarce, and most capital intensive in the commercially
oriented horticultural pathway. Although adoption of conservation measures is generally
low throughout the central region, different measures showed potential in different
pathways. For example, labor intensive investments such as terraces and live barriers
appear to have more potential in the forestry specialization pathway than most other
pathways, while fruit tree planting appears to have more potential in more extensive
production systems, such as in the basic grains expansion pathway. The resource
problems also vary significantly across pathways. For example, there are greater
problems associated with agricultural burning practices and deforestation in the basic
grains pathways than most other pathways, while water access and quality is a serious
concern in the horticultural pathway. Agricultural productivity and wages also differ
substantially across the pathways.
POLICY IMPLICATIONS
These research findings suggest several important implications for policymakers
seeking to increase agricultural productivity, ensure resource sustainability, and reduce
poverty in hillside areas of Honduras. The findings with regard to poverty suggest that
basic infrastructure and public services are critical and badly needed throughout most of
the central region, and that interventions to promote sustainable agricultural production
are likely to be insufficient to address the problems of poverty in the region. Efforts to
provide these services should be continued and expanded to poorly served areas,
43
particularly more remote areas, regardless of what is done to address resource
management issues.
Our findings with regard to technical assistance programs indicate the importance
of increasing productivity as a primary objective if such programs are to have a
substantial and long-term impact. The low adoption of most conservation measures
despite substantial efforts to promote them is often due to their high labor costs and
limited near-term economic benefits, according to survey respondents. This has often
been the case with similar measures promoted elsewhere (for example, see Lutz et al.
1994).
Part of the problem with technical assistance appears to be a lack of targeting of
different types of measures to different situations. A “one-size-fits-all” approach is
unlikely to be successful, given the pathway-specific nature of opportunities and
constraints. Production and conservation practices that are appropriate for labor intensive
agriculture, such as in forestry communities, may not be appropriate in more extensive
production systems such as in the basic grains expansion pathway, or in more external
input-intensive production such as in the horticultural pathway. Less labor-intensive
practices with income earning potential, such as fruit production, need to be developed
and promoted in such pathways.
Credit, agrarian reform and land titling programs have had limited impact in
central Honduras, probably because of their limited presence. The importance of
education as a conditioning factor in adoption of some soil conservation measures
suggests that educational improvement may have important “spin-off” benefits for
resource conservation, in addition to its direct impacts on reducing poverty.
44
The widespread decline in use of burning is an example where policies and
technical assistance programs seem to have had a large beneficial impact for
environmental objectives. The increase in continuous cropping, access to alternative
techniques (particularly use of herbicides), and shifts into cash crop production
undoubtedly were also important contributing factors to this change.
In conclusion, there are many ways to promote more productive, sustainable, and
poverty-reducing development in the central region of Honduras. While some of these
can be applied across the board, most will need to be tailored more specifically to the
particular problems and opportunities available in different development pathways.
45
REFERENCES
Angelsen, A. 1996. Deforestation: Population or market driven? Different approaches
in modelling agricultural expansion. Fantoft, Norway: Chr. Michelsen Institute.
Baland, J-M., and J-P. Platteau. 1996. Halting degradation of natural resources. Is there
a role for rural communities. Oxford: Clarendon Press.
Boserup, E. 1965. The conditions of agricultural growth. New York: Aldine Publishing
Co.
Krautkraemer, J. A. 1994. Population growth, soil fertility, and agricultural
intensification. Journal of Development Economics 44: 403-428.
LaFrance, J. T. 1992. Do increased commodity prices lead to more or less soil
degradation? Australian Journal of Agricultural Economics 36 (1): 57-82.
Leonard, H. J. 1987. Natural resource and economic development in Central America:
An environmental perspective. New Brunswick, N.J.: Transaction Books.
Lopez, R. 1998. Where development can or cannot go: The role of poverty-environment
linkages. Annual World Bank Conference on Development Economics 1997.
Washington, D.C.: World Bank.
Lutz, E., S. Pagiola, and C. Reiche, eds. 1994. Economic and institutional analysis of
soil conservation projects in Central America and the Caribbean. World Bank
Environment Paper No. 8. Washington, D.C.: World Bank.
Morris, C. T., and I. Adelman. 1988. Comparative patterns of economic development.
Baltimore, Md.: Johns Hopkins University Press.
Neidecker-Gonzales, O., and S. J. Scherr, eds. 1997. Desarrollo agrícola, sostenibilidad
y alivio de la pobreza en América Latina: El papel de las regiones de laderas.
Memoria de la conferencia celebrada del 4 al 8 de diciembre, 1995 en
Tegucigalpa, Honduras. Washington, D.C: IFPRI, IICA, SRN, DSE.
Pagiola, S. 1996. Price policy and returns to soil conservation in semi-arid Kenya.
Environmental and Resource Economics 8: 251-271.
Panayotou, T. 1993. Green markets: The economics of sustainable development. San
Francisco, Calif.: ICS Press.
Pender, J. 1998. Population growth, agricultural intensification, induced innovation and
natural resource sustainability: An application of neoclassical growth theory.
Agricultural Economics 19: 99-112.
46
Pender, J., F. Place, and S. Ehui. 1998a. Strategies for sustainable agricultural
development in the East African highlands. Paper presented at the International
Conference on Strategies for Poverty Alleviation and Sustainable Resource
Management in the Fragile Lands of Sub-Saharan Africa, co-sponsored by IFPRI,
DSE, European Commission, and the National Agricultural Research
Organization of Uganda from May 25-29 in Entebbe, Uganda.
Pender, J., S. J. Scherr, O. Neidecker-Gonzales, and G. Durón. 1998b. Pathways of
development in hillside communities of central Honduras. International Food
Policy Research Institute, Washington, D.C. Mimeo.
Pender, J., and S. J. Scherr. 1997. Community survey of hillside development patterns in
the Central region of Honduras. Environment and Production Technology
Division (EPTD). International Food Policy Research Institute, Washington, D.C.
Mimeo.
Salehi-Isfahani, D. 1988. Technology and references in the Boserup model of
agricultural growth. Journal of Development Economics 28 (2): 175-191.
Scherr, S. J., G. Bergeron, J. Pender, and B. Barbier. 1996. Policies for sustainable
development in fragile lands: Methodology overview. Fragile Lands Program.
Environment and Production Technology Division. International Food Policy
Research Institute, Washington, D.C. Mimeo.
Scherr, S. J., and P. Hazell. 1994. Sustainable agricultural development strategies in
fragile lands. EPTD Discussion Paper No. 1. International Food Policy Research
Institute, Washington, D.C.
Templeton, S., and S. J. Scherr. 1997. Population pressure and the microeconomy of
land management in hills and mountains of developing countries. EPTD
Discussion Paper No. 26. International Food Policy Research Institute,
Washington, D.C.
Tiffen, M., M. Mortimore, and F. Gichuki. 1994. More people, less erosion:
Environmental recovery in Kenya. Chichester, England: John Wiley and Sons.
Turner, B. L., G. Hyden, and R. Kates. 1993. Population growth and agricultural
change in Africa. Gainesville, Fla.: University Press of Florida.
Vosti, S. A., and T. Reardon, eds. 1997. Sustainability, growth, and poverty alleviation:
A policy and agroecological perspective. Baltimore, Md.: Johns Hopkins
University Press.
47
Appendix Table: Means of variables used in the analysis
Variable
Mid point altitude (m.a.s.l.)
Percentage of area with >30%
slope
Number of rainfall days per year
Basic
grains
expansion
Basic
grains
stagnation
848
(13)
866
(64)
1427
(37)
1168
(79)
6.46
(1.82)
7.39
(0.89)
6.39
(3.19)
7.11
(1.54)
HorticulForestry
Coffee
tural
specializa
expansion
expansion
-tion
Robust standard errors in parenthesesa
Nonfarm
employment
Region
1050
(144)
1161
(56)
1104
(33)
4.51
(1.40)
3.12
(0.63)
5.72
(0.68)
122.4
(7.8)
89.7
(4.5)
154.6
(24.5)
152.7
(10.5)
153.0
(22.9)
129.5
(7.9)
132.0
(6.2)
Population density (per km2)
-
1974
20.0
(6.8)
44.9
(7.0)
19.9
(6.2)
23.8
(5.6)
11.3
(3.9)
42.9
(10.6)
31.0
(4.1)
-
1988
19.6
(5.9)
47.0
(7.3)
42.8
(8.9)
35.5
(9.3)
18.4
(4.5)
44.7
(6.1)
38.3
(3.7)
103.2
(10.6)
69.9
(6.2)
50.6
(7.0)
76.9
(7.2)
76.6
(12.0)
40.2
(7.4)
64.7
(3.3)
5.75
(2.01)
2.08
(0.79)
0.0
(0.0)
0.21
(0.17)
0.14
(0.12)
0.0
(0.0)
0.84
(0.16)
Distance to Tegucigalpa (km)
Distance to road (km)
Percentage of communities
-
where road constructed since
1975
0.0
(0.0)
4.3
(3.5)
92.2
(6.9)
0.0
(0.0)
0.0
(0.0)
17.4
(13.3)
15.0
(6.0)
-
that already had road access in
1975
0.0
(0.0)
63.6
(11.7)
7.8
(6.9)
87.2
(9.0)
71.6
(24.3)
82.6
(13.3)
66.2
(7.2)
-
that still do not have road
access
100.0
(0.0)
32.1
(10.8)
0.0
(0.0)
12.8
(9.0)
28.5
(24.2)
0.0
(0.0)
18.8
(4.0)
-
with agricultural technical
assistance
46.0
(21.3)
100.0
(0.0)
61.7
(28.0)
100.0
(0.0)
100.0
(0.0)
55.8
(17.4)
80.4
(6.9)
-
with credit program
0.0
(0.0)
25.3
(9.8)
100.0
(0.0)
100.0
(0.0)
71.5
(24.2)
38.4
(17.3)
60.2
(7.9)
-
with agrarian reform program
0.0
(0.0)
12.9
(5.9)
38.3
(28.0)
18.3
(11.3)
28.5
(24.2)
24.0
(14.9)
20.6
(6.6)
-
with land tilling program
0.0
(0.0)
5.9
(5.0)
46.1
(27.8)
40.2
(14.0)
0.0
(0.0)
24.0
(14.9)
23.9
(6.9)
-
with enforcement of forest
regulations by national or
municipal government
15.3
(13.0)
17.2
(6.7)
53.9
(27.8)
81.7
(11.3)
100.0
(0.0)
57.4
(17.8)
57.6
(8.0)
Percentage of adults literate
-
1974
50.6
(8.9)
52.8
(4.5)
42.3
(3.2)
52.9
(5.2)
51.5
(3.3)
57.3
(5.3)
52.7
(2.5)
-
1988
56.0
(8.9)
54.8
(5.9)
48.8
(6.7)
65.7
(5.4)
64.3
(0.1)
68.7
(4.9)
62.0
(2.3)
48
Variable
Basic
grains
expansion
Basic
grains
stagnation
Horticultural
expansion
Coffee
expansion
Forestry
specializa
-tion
Nonfarm
employment
Region
Index of proportion of households usingb (1996)
-
continuous cropping without
fallow
2.77
(0.46)
3.08
(0.33)
4.50
(0.37)
4.18
(0.43)
6.00
(0.00)
3.76
(0.75)
3.96
(0.29)
-
burning to prepare fields
3.23
(0.37)
4.03
(0.16)
1.31
(0.31)
2.29
(0.23)
1.71
(0.93)
2.85
(0.49)
2.66
(0.20)
-
chemical fertilizer
3.85
(0.94)
1.49
(0.17)
4.31
(0.21)
3.71
(0.28)
3.00
(0.39)
1.76
(0.53)
2.78
(0.23)
-
insecticides
3.23
(0.58)
2.48
(0.40)
4.31
(0.21)
3.45
(0.39)
1.14
(0.56)
3.16
(0.70)
3.05
(0.27)
-
herbicides
3.53
(0.61)
3.17
(0.33)
3.53
(1.12)
2.57
(0.51)
1.14
(0.56)
1.68
(0.62)
2.48
(0.29)
-
improved seeds
0.00
(0.00)
0.34
(0.24)
5.08
(0.56)
2.15
(0.61)
1.43
(0.48)
1.25
(0.59)
1.66
(0.32)
-
irrigation
0.54
(0.21)
0.72
(0.21)
2.46
(0.83)
0.50
(0.20)
0.57
(0.28)
0.50
(0.23)
0.76
(0.17)
-
oxen
1.23
(0.37)
1.88
(0.45)
2.31
(0.21)
1.74
(0.32)
1.57
(0.28)
1.85
(0.14)
1.81
(0.14)
-
cows
3.23
(0.37)
2.55
(0.34)
2.38
(0.82)
2.33
(0.31)
3.43
(0.48)
1.91
(0.15)
2.43
(0.18)
-
pigs
4.15
(0.40)
3.48
(0.31)
2.38
(0.82)
2.60
(0.31)
2.43
(0.28)
1.96
(0.34)
2.66
(0.19)
-
chickens
4.61
(0.36)
4.39
(0.16)
3.62
(1.36)
4.16
(0.10)
4.86
(0.56)
4.15
(0.15)
4.24
(0.17)
-
contour planting
0.31
(0.18)
1.19
(0.62)
2.69
(1.37)
2.52
(0.47)
2.15
(0.47)
1.88
(0.65)
1.94
(0.27)
-
minimum tillage
1.23
(0.72)
0.53
(0.27)
0.08
(0.07)
1.23
(0.49)
2.58
(0.73)
0.77
(0.28)
0.98
(0.22)
-
mulch
1.23
(0.72)
1.12
(0.33)
0.62
(0.42)
1.83
(0.62)
0.57
(0.48)
1.15
(0.53)
1.23
(0.26)
-
incorporation of crop residues
0.61
(0.52)
1.65
(0.54)
0.93
(0.54)
1.45
(0.57)
1.14
(0.97)
1.04
(0.30)
0.98
(0.22)
-
cattle manure
0.92
(0.54)
1.20
(0.61)
0.31
(0.21)
1.06
(0.28)
0.57
(0.48)
1.04
(0.50)
0.98
(0.22)
-
chicken manure
0.00
(0.00)
0.24
(0.09)
0.62
(0.39)
0.73
(0.30)
0.00
(0.00)
2.01
(0.67)
0.89
(0.26)
-
terraces
0.15
(0.13)
0.27
(0.15)
1.62
(0.28)
2.42
(0.49)
2.29
(0.85)
1.75
(0.57)
1.60
(0.26)
-
live barriers
0.85
(0.13)
1.20
(0.22)
1.39
(0.60)
2.71
(0.50)
3.15
(0.73)
1.61
(0.48)
1.91
(0.24)
-
stone walls
1.46
(0.39)
2.72
(0.40)
1.32
(0.55)
1.08
(0.24)
2.86
(0.97)
1.47
(0.52)
1.70
(0.25)
-
tree planting
2.93
(0.63)
1.29
(0.30)
0.77
(0.32)
2.55
(0.53)
0.85
(0.47)
1.18
(0.34)
1.63
(0.21)
49
Variable
Basic
grains
expansion
Basic
grains
stagnation
Horticultural
expansion
Coffee
expansion
Forestry
specializa
-tion
Nonfarm
employment
Region
Index of proportion of households usingb (1975)
-
continuous cropping
2.44
(0.39)
2.75
(0.43)
4.14
(0.28)
2.88
( 0.55)
5.00
(0.69)
4.31
(0.55)
3.48
(0.31)
-
burning to prepare fields
4.46
(0.29)
4.62
(0.21)
4.16
(0.14)
4.26
(0.18)
5.00
(0.69)
4.20
(0.21)
4.36
(0.09)
Index of change useb (1996-75)
-
chemical fertilizer
0.69
(0.18)
0.56
(0.13)
1.00
(0.00)
0.86
(0.12)
1.00
(0.00)
0.60
(0.20)
0.76
(0.06)
-
insecticides
0.07
(0.43)
0.75
(0.10)
1.00
(0.00)
0.58
(0.24)
0.57
(0.28)
0.78
(0.16)
0.67
(0.11)
-
herbicides
0.85
(0.13)
0.94
(0.05)
0.92
(0.07)
0.83
(0.12)
0.57
(0.28)
0.60
(0.19)
0.78
(0.07)
-
improved seeds
0.00
(0.00)
0.06
(0.05)
0.46
(0.28)
0.27
(0.13)
0.43
(0.28)
0.18
(0.16)
0.23
(0.07)
-
irrigation
0.23
(0.32)
0.20
(0.21)
-0.42
(0.45)
0.21
(0.13)
0.57
(0.28)
0.13
(0.29)
0.19
(0.11)
-
oxen
-0.07
(0.43)
-0.28
(0.10)
-0.38
(0.29)
-0.63
(0.16)
-1.00
(0.00)
-0.85
(0.14)
-0.60
(0.09)
-
cows
-0.69
(0.18)
0.11
(0.17)
-1.00
(0.00)
-0.63
(0.25)
-0.43
(0.48)
-0.85
(0.14)
-0.58
(0.11)
-
pigs
0.08
(0.38)
-0.29
(0.29)
-1.00
(0.00)
-0.73
(0.13)
-1.00
(0.00)
-0.79
(0.14)
-0.66
(0.10)
-
chickens
-0.39
(0.24)
-0.21
(0.09)
-0.84
(0.10)
-0.14
(0.12)
-0.43
(0.28)
-0.13
(0.09)
-0.27
(0.07)
24.0
(9.4)
0.0
(0.0)
Percentages of communities
taking collective action to
control runoff or improve
common land
54.0
(21.3)
32.1
(15.4)
0.0
(0.0)
41.3
(17.4)
29.5
(7.7)
Typical maize yield (kg/ha)
-
in a bad year
886
(190)
649
(57)
1028
(93)
709
(188)
1249
(251)
905
(269)
834
(101)
-
in a good year
1792
(376)
1667
(169)
2345
(182)
2259
(536)
2339
(209)
2504
(474)
2195
(210)
Index of change in typical
maize yieldc (1996-75)
0.54
(0.29)
-0.50
(0.25)
0.08
(0.56)
-0.36
(0.20)
-1.00
(0.00)
-0.41
(0.27)
-0.35
(0.13)
Adult male agricultural wage, 1996 (Lps /day)
-
low (slack season)
15.8
(2.0)
16.6
(2.0)
22.3
(1.4)
18.2
(0.7)
20.0
(0.0)
24.4
(1.0)
20.1
(0.8)
-
high (peak season)
19.6
(2.6)
22.1
(1.0)
34.2
(7.1)
58.3
(6.3)
39.9
(14.0)
38.0
(3.1)
39.6
(2.9)
50
Variable
Basic
grains
expansion
Basic
grains
stagnation
Horticultural
expansion
Coffee
expansion
Forestry
specializa
-tion
Nonfarm
employment
Region
Ratio of 1996 to 1975 real adult male wage
-
low (slack season)
1.07
(0.25)
0.69
(0.21)
1.19
(0.13)
1.10
(0.19)
1.05
(0.16)
1.08
(0.11)
1.05
(0.08)
-
high (peak season)
1.27
(0.27)
0.76
(0.21)
1.71
(0.33)
1.90
(0.71)
1.41
(0.30)
1.47
(0.24)
1.55
(0.26)
Percentage of households with a dirt floor
-
1974
98.6
(0.9)
85.9
(2.4)
90.2
(3.3)
88.2
(5.2)
94.7
(3.4)
82.8
(3.9)
87.8
(2.0)
-
1988
88.6
(2.7)
77.8
(3.0)
80.2
(4.9)
82.6
(5.6)
93.1
(4.8)
66.6
(7.2)
78.5
(2.9)
Percentage of households in which last child died
-
1974
10.2
(1.4)
8.0
(1.0)
5.7
(2.7)
7.9
(1.9)
6.7
(1.9)
10.2
(1.4)
8.4
(0.8
-
1988
6.1
(1.0)
4.1
(0.6)
3.8
(1.7)
3.6
(0.6)
3.6
(1.6)
4.2
(1.5)
4.0
(0.5)
Percentage of steep land (over 30% slope ) in late 1970s
-
tree cover
66.1
(16.0)
27.3
(4.5)
76.1
(5.8)
77.1
(5.7)
100.0
(0.0)
64.6
(14.1)
66.1
(5.5)
-
devegetated
25.8
(12.5)
60.0
(4.7)
3.4
(1.7)
11.2
(4.4)
0.0
(0.0)
28.0
(14.5)
23.1
(5.6)
-
cultivated
2.5
(1.0)
6.2
(1.3)
17.7
(8.7)
6.4
(2.7)
0.0
(0.0)
5.0
(1.8)
6.9
(1.8)
-
pasture/fallow
5.6
(3.5)
6.5
(1.6)
2.7
(1.7)
5.3
(1.5)
0.0
(0.0)
2.3
(1.6)
3.9
(0.8)
Index of change inc (1996-75)
a
-
cropland quality
0.54
(0.29)
-0.72
(0.24)
0.08
(0.56)
-0.36
(0.24)
-1.00
(0.00)
-0.48
(0.31)
-0.41
(0.14)
-
forest area
-1.00
(0.00)
-0.82
(0.09)
-0.23
(0.14)
-0.14
(0.30)
-0.72
(0.24)
-0.32
(0.17)
-0.43
(0.12)
-
forest quality
-1.00
(0.00)
-0.72
(0.24)
-0.46
(0.29)
0.01
(0.34)
-1.00
(0.00)
-0.44
(0.30)
-0.47
(0.15)
-
water availability
-1.00
(0.00)
-0.73
(0.11)
-0.75
(0.14)
-0.62
(0.10)
-0.77
(0.06)
-0.83
(0.09)
-0.73
(0.05)
-
water quality
NA
-0.48
(0.13)
-0.50
(0.13)
-0.60
(0.09)
-0.60
(0.0)
-0.69
(0.20)
-0.61
(0.09)
Means and standard errors are adjusted to account for sampling weights, stratification, and finite population of communities
Values of index: 0 = none, 1 = less than 10%, 2 = minority, 3 = half, 4 = majority, 5 = more than 90%, 6 = all
c
Index values: -1 = decrease, 0 = no change, +1 = increase
b
List of EPTD Discussion Papers
01
Sustainable Agricultural Development Strategies in Fragile Lands, by Sara J.
Scherr and Peter B. R. Hazell, June 1994.
02
Confronting the Environmental Consequences of the Green Revolution in Asia, by
Prabhu L. Pingali and Mark W. Rosegrant, August 1994.
03
Infrastructure and Technology Constraints to Agricultural Development in the
Humid and Subhumid Tropics of Africa, by Dunstan S. C. Spencer, August 1994.
04
Water Markets in Pakistan: Participation and Productivity, by Ruth MeinzenDick and Martha Sullins, September 1994.
05
The Impact of Technical Change in Agriculture on Human Fertility: District-level
Evidence from India, by Stephen A. Vosti, Julie Witcover, and Michael Lipton,
October 1994.
06
Reforming Water Allocation Policy Through Markets in Tradable Water Rights:
Lessons from Chile, Mexico, and California, by Mark W. Rosegrant and Renato
Gazmuri S., October 1994.
07
Total Factor Productivity and Sources of Long-Term Growth in Indian
Agriculture, by Mark W. Rosegrant and Robert E. Evenson, April 1995.
08
Farm-Nonfarm Growth Linkages in Zambia, by Peter B. R. Hazell and Behjat
Hojjati, April 1995.
09
Livestock and Deforestation in Central America in the 1980s and 1990s: A Policy
Perspective, by David Kaimowitz (Interamerican Institute for Cooperation on
Agriculture), June 1995.
10
Effects of the Structural Adjustment Program on Agricultural Production and
Resource Use in Egypt, by Peter B. R. Hazell, Nicostrato Perez, Gamal Siam and
Ibrahim Soliman, August 1995.
11
Local Organizations for Natural Resource Management: Lessons from
Theoretical and Empirical Literature, by Lise Nordvig Rasmussen and Ruth
Meinzen-Dick, August 1995.
12
Quality-Equivalent and Cost-Adjusted Measurement of International
Competitiveness in Japanese Rice Markets, by Shoichi Ito, Mark W. Rosegrant,
and Mercedita C. Agcaoili-Sombilla, August, 1995.
13
Role of Inputs, Institutions, and Technical Innovations in Stimulating Growth in
Chinese Agriculture, by Shenggen Fan and Philip G. Pardey, September 1995.
14
Investments in African Agricultural Research, by Philip G. Pardey, Johannes
Roseboom, and Nienke Beintema, October 1995.
15
Role of Terms of Trade in Indian Agricultural Growth: A National and State
Level Analysis, by Peter B. R. Hazell, V. N. Misra and Behjat Hojjati, December
1995.
16
Policies and Markets for Non-Timber Tree Products, by Peter A. Dewees and
Sara J. Scherr, March 1996.
17
Determinants of Farmers’ Indigenous Soil and Water Conservation Investments
in India’s Semi-Arid Tropics, by John Pender and John Kerr, August 1996.
18
Summary of a Productive Partnership: The Benefits From U.S. Participation in
the CGIAR, by Philip G. Pardey, Julian M. Alston, Jason E. Christian and
Shenggen Fan, October 1996.
19
Crop Genetic Resource Policy: Towards a Research Agenda, by Brian D. Wright,
October 1996.
20
Sustainable Development of Rainfed Agriculture in India, by John M. Kerr,
November 1996.
21
Impact of Market and Population Pressure on Production, Incomes and Natural
Resources in the Dryland Savannas of West Africa: Bioeconomic Modeling at the
Village Level, by Bruno Barbier, November 1996.
22
Why Do Projections on China’s Future Food Supply and Demand Differ? by
Shenggen Fan and Mercedita Agcaoili-Sombilla, March 1997.
23
Agroecological Aspects of Evaluating Agricultural R&D, by Stanley Wood and
Philip G. Pardey, March 1997.
24
Population Pressure, Land Tenure, and Tree Resource Management in Uganda,
by Frank Place and Keijiro Otsuka, March 1997.
25
Should India Invest More in Less-favored Areas? by Shenggen Fan and Peter
Hazell, April 1997.
26
Population Pressure and the Microeconomy of Land Management in Hills and
Mountains of Developing Countries, by Scott R. Templeton and Sara J. Scherr,
April 1997.
27
Population Land Tenure, and Natural Resource Management: The Case of
Customary Land Area in Malawi, by Frank Place and Keijiro Otsuka, April 1997.
28
Water Resources Development in Africa: A Review and Synthesis of Issues,
Potentials, and Strategies for the Future, by Mark W. Rosegrant and Nicostrato
D. Perez, September 1997.
29
Financing Agricultural R&D in Rich Countries: What’s Happening and Why, by
Julian M. Alston, Philip G. Pardey, and Vincent H. Smith, September 1997.
30
How Fast Have China’s Agricultural Production and Productivity Really Been
Growing? by Shenggen Fan, September 1997.
31
Does Land Tenure Insecurity Discourage Tree Planting? Evolution of Customary
Land Tenure and Agroforestry Management in Sumatra, by Keijiro Otsuka, S.
Suyanto, and Thomas P. Tomich, December 1997.
32
Natural Resource Management in the Hillsides of Honduras: Bioeconomic
Modeling at the Micro-Watershed Level, by Bruno Barbier and Gilles Bergeron,
January 1998.
33
Government Spending, Growth and Poverty: An Analysis of Interlinkages in
Rural India, by Shenggen Fan, Peter Hazell, and Sukhadeo Thorat, March 1998.
34
Coalitions and the Organization of Multiple-Stakeholder Action: A Case Study of
Agricultural Research and Extension in Rajasthan, India, by Ruth Alsop, April
1998.
35
Dynamics in the Creation and Depreciation of Knowledge and the Returns to
Research, by Julian Alston, Barbara Craig, and Philip Pardey, July 1998.
36
Educating Agricultural Researchers: A Review of the Role of African
Universities, Nienke M. Beintema, Philip G. Pardey, and Johannes Roseboom,
August 1998.
37
The Changing Organizational Basis of African Agricultural Research, Johannes
Roseboom, Philip G. Pardey, and Nienke M. Beintema, November 1998.
38
Research Returns Redux: A Meta-Analysis of the Returns to Agricultural R&D,
Julian M. Alston, Michele C. Marra, Philip G. Pardey, and T. J. Wyatt, November
1998.
39
Technological Change, Technical and Allocative Efficiency in Chinese
Agriculture: The Case of Rice Production in Jiangsu, Shenggen Fan, January
1999.
40
The Substance of Interaction: Design and Policy Implications of NGOGovernment Projects in India, Ruth Alsop with Ved Arya, January 1999.
41
Strategies for Sustainable Agricultural Development in the East African
Highlands, John Pender, Frank Place, and Simeon Ehui, April 1999.
42
Cost Aspects of African Agricultural Research, Philip G. Pardey, Johannes
Roseboom, Nienke M. Beintema, and Connie Chan-Kang, April 1999.
43
Are Returns to Public Investment Lower in Less-favored Rural Areas? An
Empirical Analysis of India, Shenggen Fan and Peter Hazell, May 1999.
44
Spatial Aspects of the Design and Targeting of Agricultural Development
Strategies, Stanley Wood, Kate Sebastian, Freddy Nachtergaele, Daniel Nielsen,
and Aiguo Dai, May 1999.
45
Pathways of Development in the Hillsides of Honduras: Causes and Implications
for Agricultural Production, Poverty, and Sustainable Resource Use, John
Pender, Sara J. Scherr, and Guadalupe Durón, May 1999.
Fly UP