Swidden Agriculture in a Forest Society: Livelihood Strategies in the Maya Biosphere Reserve
Community of Uaxacti~n, Pet~n, Guatemala

Paul Litow
Former Graduate Student
Farming Systems Research and Extension
University of Florida
Email: jeroky@micron.net

Matt Baker
Professor & Chair, Agricultural Education and Communications
Texas Tech University
Box 42131
Lubbock, Texas 79409-2131
Phone: (806) 742-2816
Fax: (806) 742-2880
E-Mail: Matt.Baker@ttu.edu

Peter Hildebrand
Professor, Food and Resource Economics
University of Florida
P.O. Box 110240
Gainesville, Florida 32611-0240
Phone: (352) 392-1845 ext. 436
E-Mail: peh@gny.ifas.ufl.edu

Abstract

Milpa, or slash and burn agricutlture, is one of many livelihood strategies utilized by most households in
Uaxactdn, Guatemala. Participatory Rural Appraisal and Rapid Rural Appraisal revealed that although
household livelihood systems are primarily based on the extraction ofnon-timber forest products
(NTFPs), milpa is o~f fudamental importance. These approaches also suggest that households shif i in
and out ofsubsistence as they respond to changes in their sources ofcash earnings. Ethnographic linear
programming supports this premise, showing that householdrF also modify their livelihood strategies in
response to changes in the environment. How households choose to strategize and the intensity with
which they participate in each livelihood activity, however, is driven by their composition and the
household ratio ofconsumers to producers. These two factors not only determine their cash and
nutritional needs, but also the size of their labor force, enhancing or inhibiting their ability to support
themselves.

Introduction

Uaxactdin is a community located in Guatemala's Pet~n, its northern-most department (state) that shares
borders with Mexico and Belize. Although the Peten comprises one-third of the country's area, it was
largely ignored by the Colonial Spanish and early Guatemalan governments because of its remoteness and
inhospitable climate. It was not until the mid-1960s that colonization began in earnest (Schwartz, 1990).
Explosive growth continues, driven by severe land shortages in the country's highlands and the return of
refugees following Guatemala's civil war.

Uaxactdn is located within the borders of the Maya Biosphere Reserve (MBR), designated in 1990
through a cooperative effort between the Guatemalan government, Conservation Intemnational, and The










Nature Conservancy. The MBR covers more than 1.6 million hectares (ha) and is the largest contiguous
tract of tropical forest left within Central America. The reserve consists of five national parks and three
biotypes (biological reserves) surrounded by a large Multiple Use Zone (MUZ). While permanent
settlement is not allowed within parks or biotypes, it is permitted in the MUZ. Small-scale agriculture
and exploitation of above and below ground resources, subject to certain regulations, are also allowed
within the MUZ.

Uaxactdn is located in the MUZ some 83 kilometers north of Flores and is a small, relatively isolated
community of 136 families (Organizati6n Manejo y Conservaci6n, 1998). The village lies along an
unpaved road running from Tikal National Park, 24 kilometers to the south, to the Mexican border some
100 kilometers to the north. This community was originally the area center of a Maya city-state from
about 278 to 889 A.D. (Smith, 1950, cited in McNab, 1999). Contemporary civilization followed in the
late 1800s and early 1900s when Unxactin was used as a chicle (Manilkara zapota) camp. Today, most
families base their livelihoods upon a combination of hunting; wild allspice (Pimenta officinalis)
harvesting; xate cutting, an understory palm of the Chamaedora species; chicle extraction, used as a base
for chewing gum; and milpat, or slash and burn agriculture. It should be noted here that milpa is also used
as a descriptor of where maize and other crops are grown, much like the words 'field' or 'farm'.

Uaxactdin was awarded an 83,558 ha community forest concession in early 1998 by the Guatemalan
government, formally giving the community management and usufruct rights to all above-ground
resources. Under the terms of the concession, the community must meet governmental requirements
(outlined in McNab, 1999) for managing the area and pay a lease of $142,049 over a ten-year period.
Taxation of selected non-timber forest product (NTFP) sales, such as xate, is currently generating some
income for payment of the lease fee and management of the concession area. The community is also
pursuing other income-generating possibilities such as ecotourism and small local industry.

The Wildlife Conservation Society (WCS) has consistently worked with Uaxactdin for several years
throughout the concession solicitation process as well as in wildlife conservation. Presently WCS
partners with two other non-governmental organizations (NGOs) to assist the community with technology
transfer, training, and all aspects of concession management. WCS wished to learn the extent to which
village households depended upon milpa or swidden agriculture. While milpa was generally perceived to
be a subsistence activity, its importance to household livelihoods was not well understood.

Purpose and Objectives

This objective of this study was to determine the relative importance of milpa to households within the
context of household livelihood systems (Chambers & Conway, 1992; Hoon, Singh & Wanmali, 1997;
Scoones, 1998). This study also sought to determine if milpa's relative importance as a livelihood
strategy (Devereux, 1999) changed as intra- and extra-household conditions changed. The following
objectives were developed to guide the study: (1) ascertain the relative importance of milpa to other
household livelihood activities, namely harvesting allspice, chicle, xate, and cash employment to
Uaxactdin's families; and (2) examine the response of modeled households to three different scenarios (a
change in the natural environment, a shift in markets for NTFPs, and a new option for local employment).

Methods and Data Sources

Research for this investigation was carried out fr-om July to December of 1998 by the lead author of this
study (Litow, 2000). Data gathered were not used to build only a qualitative understanding of how
livelihood systems in Uaxactd~n work, but to descriptively and quantitatively model how households
function.










Direct observation, informal individual interviews, Rapid Rural Appraisal (RRA), Participatory Rural
Appraisal (PRA), and formal questionnaires were used to gather information about household livelihood
systems and strategies, the relative importance of activities comprising those systems, and resources used
and provided by household activities. Chambers (1997) contrasts RRA and PRA by suggesting that RRA
is utilized to 'find out' or elicit information, where PRA empowers people to use their knowledge, means,
and resources to solve their problems without relying on outsiders. The outsider's role in RRA is more
investigative in nature, where in PRA it is more facilitative in nature.

For this study, the RRA methods which were used included secondary data, direct observation, expert
opinion, livelihood systems diagrams, semi-structured interviews, household time lines, and transect
walks. PRA methods included ethnogmaphic drawings, daily and seasonal calendars, trend lines, and
matrix scoring and ranking.

Ethnographic linear programming (Bastidas, 2001; Kaya, Hildebrand, & Nair, 2000) was used to simulate
the response of modeled households to various scenarios. Traditional linear programming is a method of
maximizing or minimizing the outcome of a primary objective relative to selected constraints. For
example, if a household wants to earn as much cash as possible, this household will use available
resources to carry out the necessary activities to do so, resources that are also being used to do other
things such as produce needed food. The linear program calculates how much of those resources are
available after meeting household needs and then calculates how much cash can be earned using the
remammig resources.

This method diverges from traditional linear programming in the way the data are collected, and in the
sense that socio-cultural parameters, changing nutritional requirements, evolving household compositions,
and other factors, are added to enhance the models' dynamism. This, in turn, enables the programmer to
build models that better reflect reality. Six actual households and one fictitious "average household" were
modeled. This fictitious household was created to help determine whether the use of averages, maskmng
the diversity found in the real situation, can result in misleading conclusions. When the scenarios (a
change in the natural environment, a shift in markets for NTFPs, and a new option for employment) were
presented, the effect upon each household was analyzed based upon each household's composition, or its
unique characteristics in terms of its overall size, its members' ages, and their sex. These characteristics
determine household behavior because they affect, among other things, a household's nutritional
requirements and its ability to produce food and cash to sustain itself.

Findings

The first objective of this study was to determine the importance of milpa to households. Thirty-three
residents participated in this activity. Seventy percent of respondents considered milpa as the most
important household activity, and another 27% answered that xate harvesting was most important. Two
PRA sessions were conducted where milpa's importance was examined over time. The first group of
eight participants rated the importance of milpa much lower over time than other activities. This group
was comprised of individuals who worked as laborers on Maya ruin restoration projects during the 1980s.
The second group of nine participants, who did not have a history of working as employees, rated milpa
as their most important household activity from 1980 1998.

Regarding the second objective, the three linear programming scenarios collectively showed several
things. First, all six households and the seventh "average household" relied upon diverse sets of
livelihood strategies to survive. These strategies were used in various combinations to reduce risk of
extreme economic stress and hunger. The diversity of these strategies is paramount to household well
being. When one or more elements of the livelihood system fail due to environmental disturbances or
economic downturns, households can turn to other options for survival.











Second, the linear programming outcomes also showed that household composition and consumer-to-producer (C/P)
ratios (Chayanov, 1986) strongly affect household livelihood strategies (Table 1). Household 1, for example, was
able to narrow the scope of its livelihood activities (specialize) and increase its well-being in terms of cash earnings
and maize production because it had three able-bodied adult males (favorable household composition) and relatively
low household stress levels (a low C/P ratio). Households with less favorable household compositions and higher
C/P ratios (Households 3, 4, 5, 6) were driven to participate in a greater number of activities. This is owed to high
levels of household food and labor stress, embodied in numerous young and growing children who cannot
significantly add to the household's productive capacity but who consume its food. Paradoxically, Household 2's
comparatively favorable C/P ratio hides the fact that it is under high levels of stress. Because all but one of its
productive members are female, and because culturally women in Uaxacthsn are unable to participate in most major
food-producing and cash-earning activities, this household's composition prohibits its members from functioning at
their highest productive potential. Consumer-to-producer ratios must be examined in conjunction with household
composition.

Third, linear programming and actual observations showed that households adjust their livelihood
strategies in response to adverse environmental change or economic difficulty. This is the case in
Scenarios I and 2 where normal household activity mixes were disrupted. Their responses ranged from
minimizing cash spending and investing more household labor in agriculture (a return to subsistence), to
maximizing discretionary cash earnings by shifting household labor away ~from milpa into cash-earning
activities (a move away from subsistence), or a mix of the two. Fourth, results of these analyses indicated
that some decrease in NTFP harvests occurred in Scenario 3 when household members were given the
opportunity to work for the community's forest concession (Figure 1). It appears that new employment
options may indirectly ameliorate the rate of NTFP exploitation by diverting available household labor
away from extractive activities.

Fifth, total milpa area changed very little throughout the three scenarios (Table 2). This indicates that
even when bio-physical and socio-economic conditions favor specialization in cash-earning activities,
purchasing needed maize, and decreasing labor investment in milpa, households continue raising milpa.

Table 1: Change in the number of household livelihood activities performed by households when
comparing the three modeled scenarios"

Household Number of Number of Number of
Consumer to activities activities activities
Producer ratio performed in performed in performed in
Household Scenario 16 Scenario 2c Scenario 3d
I 1.6:1 3 2 2
2 1.5:1 2 1 1
3 2.3:1 3 4 5
4 2.3:1 3 4 4
5 2.5:1 3 2" 4
6 3.5:1 3 3 3
7 2.0:1 2 2 4
a Household activity es not included in this table are those undertaken to maintain the household, such as
hauling water, cooking, etc., as these remain constant.
b This scenario modeled household effects based upon changes in the environment.
This scenario modeled household effects based upon shifts in markets for NTFPs.
d This scenario modeled household effects based upon a new employment option.
e This household would have to drastically reduce food consumption and expenses in order to survive in
this scenario











1000-
900
800 +- W Scenario I
700 -H E Scenario2-
600 1 8 Scenario3
S500-

300-
200
100-

Harvest xate Harvest Harvest chicle Paid worker Concession
allspice worker
Activity


Figure 1: Total male labor use in seven modeled households in Linear Progrrammingr Scenarios 1-3

The slight differences in total milpa area comparing Scenario I to Scenarios 2 and 3 is attributed to
differing household livelihood strategies. In Scenario 1, households were more subsistence-oriented
because there were fewer cash-generating options due to bio-physical and socio-economic factors. In the
latter two scenarios, households had a greater choice of cash-earning options, and thus, overall, chose to
raise less maize.

An important point regarding Table 2 is that the size of Household 7's milpa (the "average household") is
often quite different from that of the other six households. While Household 7 is not representative of all
households in Uaxactdn, it is evident that the size of its milpa in these three scenarios is not representative
of the other six households' milpas. In Scenario 1, only Household 3 has a milpa similar in size to that of
Household 7. The same pattern emerges in the other scenarios, with two households in Scenario 2 and
one in Scenario 3 having similar milpa sizes to Household 7.

Table 2: Changes in Household and Total Milpa Area in Scenarios 1-3.

Scenario 1: Changes in Scenario 2: Shift in Scenario 3: New
Household # Environment Markets for NTFPs EmlyetOton
1 11.40 12.56 12.56
2b 0.00 0.00 0.00
3 2.00 2.96 1.77
4 4.20 1.33 0.79
5 0.00 0.00 2.72
6 2.90 2.25 2.25
7b 1.98 2.91 1.84
Totals 22.48 22.01 21.93
a Milpa area measured in manzanas. One manzana approximately equa s 0.7 hectares.
bNote that Household 2 is a poor, exceptionally marginalized, female-headed household that barely
survives; Household 7 is the "average household."


Conclusions, Implications, and Recommendations











It is concluded from the first objective of this study that because individuals in the initial PRA session
were receiving a steady flow of cash and likely had little time to work in their milpas, they focused more
upon cash-earning activities outside of work easily undertaken for short lengths of time on weekends or
vacations (harvesting chicle, xate, and allspice). Although they still tended to focus on cash activities
after restoration projects ended, one notes that milpa began gaining importance relative to other activities
as they shifted their livelihood strategies towards subsistence and away fr~om earning cash. Participants of
th~e second PRA session, who did not have consistent cash earnings, tended to rank xate and milpa
equally.

A conclusion related to the second objective is that basing assumptions about how much milpa area
households will produce (and need) upon averages is misleading. Milpa size varies according to how
much labor households have available and how much maize they need annually, factors that continuously
change as household members mature--and hence labor supply---changes. By first analyzing this
variability based on household composition, diverse households can later be grouped into household
domains of similar characteristics for purposes of aggregation into community characteristics. Finally,
these research findings show that milpa is an invaluable and fundamental building block of household
livelihood strategies, one which is relied upon at varying levels according to household composition, the
socio-economic environment with which it interacts, or bio-physical disturbances. It is important to note
that except for the female-headed household, no household in any of the three scenarios stopped raising
milpa. Even in times when its relative importance wanes, as in Scenario 3, households continue to rely
upon milpa to reduce food expenses and as insurance against oft-occurring times of difficulty.

The diversity of household livelihood systems--and thus livelihood strategies--must be maintained. It is
this diversity that allows Uaxactd~n's families to draw upon numerous survival strategies to sustain
themselves over time. The circumstances in which households find themselves are as dynamic as the
households themselves. Not all of a household's possible strategies are available or useful in every
circumstance, such as when one or more elements o fthe livelihood system fail; or the circumstance itself
may inhibit or prevent use of livelihood strategies or their usefulness as survival tools. It is imperative
that households continue to have multiple options to ensure their food and economic security.

With diversity in mind, those engaged in extension programming should use participatory approaches to
increase the likelihood of successful planned change. Such approaches can help people and agencies
identify, highlight, and prioritize problems, and determine the best courses of action. Because they are
participatory, diverse and complex relationships between households, livelihood systems, and
communities are recognized and included throughout diagnostic, problem-solving, and monitoring and
evaluation processes. These methods aspect and use indigenous knowledge, uncover the role and
importance of gender, and encourage bi-directional information flows between assisting agencies and
stakeholders.

Ethnographic linear programming can be an integral part of the evaluation of such participatory
processes. When done correctly, it is a useful tool that allows researchers, extension professionals, and
clients to analyze the intended (and unintended) effects of planned change on individual households and
domains of similar households. Together, the results of these participatory approaches and this tool can
empower stakeholders in developing solutions more adaptable and applicable to local situations without
simplifying the complexities of livelihood systems.

References

Bastidas, E.P. (2001). Assessing potential response to changes in the livelihood system of diverse,
limited-resource farm households in Carchi, Ecuador: Modeling livelihood strategies using participatory
methods and linear programming. Doctoral Dissertation, University of Florida, Gainesville.











Chambers, R. & Conway, G. (1992). Sustainable rural livelihoods: Practical concepts for the 21st
century. IDS Discussion Paper 296. Institute of Development Studies, University of Sussex, England.

Chayanov, A. V. (1986). Peasant farm organization. (In D. Thorner, B. Kerblay & R. E. F. Smith (Eds.),
The theory of the peasant economy (pp. 29-269). Madison, WI: University of Wisconsin Press.

Devereux, S. (1999). Making less last longer: Informal safety nets in Malawi. IDS Discussion Paper
373. Institute of Development Studies, University of Sussex, England.

Hoon, P., N. Singh, S., & Wanmali, S. (1997, August). Sustainable livelihoods: Concepts, principles,
and approaches to indicator development. Unpublished discussion paper prepared for the Workshop on
Sustainable Livelihoods Indicators. New York, United Nations Development Programme.

Kaya, B., P.E. Hildebrand and P.K.R. Nair (2000). Modeling changes in farming systems with the
adoption of improved fallows in southern Mali. Agricultural Systems, 66. (51-68).

Litow, P. (2000). Food security and household livelihood strategies in the Maya Biosphere Reserve: The
importance of milpa in the community of Uaxactirn, Pet~n, Guatemala. Masters thesis, University of
Florida, Gainesville.

McNab, R. (1999). Comparative impacts of chicle and xate harvests on wildlife of the Maya Biosphere
Reserve, Guatemala. Masters thesis, University of Florida, Gainesville.

Organizacii~n Manejo y Conservacign. (1998). Justificaci6n T~cnica, Solicitud Concesii~n Forestal
Comunitaria de Uaxactdin (Technical Justification for the Application of a Community Forest Concession
for Uaxactinln. Unpublished report, CATIE/CONAP, Flores, Pet~n, Guatemala.

Schwartz, N. B. (1990). Forest society (Philadelphia, University of Pennsylvania Press).

Scoones, I. (1998). Sustainable rural livelihoods: A framework for analysis. Working Paper 72. Institute
of Development Studies, University of Sussex, England.