Land Use Policy 23 (2006) 147–160
Impact of forestland allocation on land use
in a mountainous province of Vietnam
Jean-Christophe Castella
a,b,c,Ã
, Stanislas Boissau
a,c
, Nguyen Hai Thanh
c
, Paul Novosad
c
a
Institut de Recherche pour le De
´
veloppement (IRD), 213 rue Lafayette, 75480 Paris, Cedex 10, France
b
International Rice Research Institute (IRRI), DAPO 7777, Metro Manila, Philippines
c
Mountain Agrarian Systems (SAM) Program, Vietnam Agricultural Science Institute (VASI), Thanh Tri, Hanoi, Vietnam
Received 19 January 2004; received in revised form 15 May 2004; accepted 28 July 2004
Abstract
In the early 1990s following the decollectivization of agriculture, the Vietnamese government distributed forest land-use rights to
individual households. The new forestland policy had three related objectives: (i) the introduction of a sedentary livelihood system
for those populations who had traditionally relied on shifting cultivation and regular migration; (ii) the development of the village
economy through tree plantations; and (iii) the protection of forest resources. In this paper, we discuss the changes in land use that
resulted from the new forestland policy and the effectiveness of the policy in achieving each of its three objectives.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Land policy; Forestland allocation; Natural resource management; Livelihood systems; Mountain agriculture; Bac Kan; Vietnam
Introduction
The allocation of forestland to individual households
was the final step in the changes in land-use rights that

accompanied the doi moi (renovation) reforms in
Vietnam. In a context of declining agricultural produc-
tivity and food shortages, the cooperative system was
undergoing a major crisis. A dual system of production
had developed, with collectively managed work in
paddyfields complemented by individually managed
work on the hillsides (Sadoulet et al., 2002). Although
the ‘‘people’’ (in other words, the State, Article 19 of the
1980 constitution) officially owned all land, sloping
lands had not been integrated into the cooperatives. For
this reason, sloping lands continued to be cultivated
according to rules of free-access,
1
providing a supple-
mentary income for many households. In times of
shortage, farmers focused on the private economy,
increasing the area cultivated on the slopes, which led
to major deforestation in northern regions (De Koninck,
1999; Castella et al., 2002).
On April 5, 1988, the Vietnamese Communist Party
politburo adopted Resolution 10 to address the agri-
cultural crisis in the country. Resolution 10 emphasized
the importance of private property rights, as well as the
need for each region to design a development model
suited to its own natural, economic and social environ-
ment. In recogni tion of the importance of the private
agricultural economy, the new policy dismantled the
failing cooperatives and beg an to restore ownership of
means of production to individuals (Bergeret, 1995;
Kerkvliet and Porter, 1995). In the mountainous areas

of northern Vietnam, the restoration of the rights of the
ARTICLE IN P RESS
www.elsevier.com/locate/landusepol
0264-8377/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.landusepol.2004.07.004
Ã
Corresponding author. IRD, B.P. 64501, 34394 Montpellier, Cedex
5, France. Tel.: +33 4 67 63 69 80; fax: +33 4 67 63 87 78.
E-mail address: (J C. Castella).
1
Traditional land access was free, with the absence of rules made
feasible by the low population pressure on mountainous areas at this
(footnote continued)
period ( Mellac, 2000). The farmer who cleared a given upland field
secured for him- or herself the right to use that field until it was
fallowed for regeneration.
individual to use paddyland in the early 1990s was
intended to stimulate paddyfield intensification while
reducing pressure on the surrounding upland areas
(Donovan et al., 1997). The policy was successful in that
it gave farmers an incentive to invest more in their own
ricefields, and resulted in substantial productivity gains
in the inter-mountain valleys (Sikor, 1999; Mellac, 2000;
Sadoulet et al., 2002).
However, the implementation of the national land
allocation policy varied from region to region. Rather
than distributing land as the policy suggested, the Black
Thai villages of Son La Province left the village in
control of the land, and periodically redistributed land
among households ( Sikor, 1999). In contrast, in 1990 the

Tay ethnic group in the Cao Bang Province began a
movement to reclaim the lands of their ancestors, a
movement that soon spread to neighbouring provinces.
As Tay farmers repossessed the land that their
forefathers had contributed to the agricultural coopera-
tives, other groups were left without paddyfields,
particularly households who had not owned paddyfields
before the institution of the cooperatives, but who had
nonetheless been contributing to the cooperatives in
recent years. These groups (mostly Kinh from the delta
region and Dao and H’mong who had emigrated from
the uplands) were forced to turn to the uplands to meet
their food needs.
Despite the paddyland allocations, many farmers
turned to extensive slash-and-burn systems in the
uplands, which had not yet been allocated (Rambo et
al., 1995; Jamieson et al., 1998). The State hoped to
regulate the runaway exploitation of the uplands by
applying the same solution that had worked in the
lowland areas (Pillot, 1995; Fforde and de Vylder, 1996;
De Koninck, 1999). The forestland, which encompassed
all land that was or should be covered in forest (as
defined by the State), was allocated to individual
households as stipulated in the 1993 land law. This final
stage of the land allocation policy had three major
objectives: (i) converting the populations of migratory
cultivators to a sedentary livelihood system, (ii) increas-
ing agricultural production in the uplands by giving
farmers incentives to grow perennial plantations, and
(iii) preserving the deteriorating forest resource base

(Government of the Socialist Republic of Vietnam,
1994; Nguyen et al., 1995).
In this paper, we examine the goals of the State in
forestland allocation. Based on a case study undertaken
in four villa ges in a mountainous region of northern
Vietnam, we then analyse to what extent this land
allocation achieved its stated goals and what impacts it
had on the livelihood systems of farmers. Finally, we
discuss possible trajectories of future production sys-
tems, and their implications for development interven-
tions.
Methods
The study draws on diverse sources of data including
(i) a review of published literature, official statistics,
legal and policy documents, (ii) monographic studies
and analysis of land cover maps derived from aerial
photographs, and (iii) qualitative data collected in
interviews, oral histories and participatory observation.
In four remote mountain villages within Na Ri
District of Bac Kan Province, we investigated the
process and effects of forestland allocation at the
grassroots level. None of the four villages is accessible
by road or motorized transportation. Unlike many other
ethnic minority villages in Bac Kan province, these
villages have not been the targets of the major
government projects (i.e., infrastructure, education,
production innovations) associated with the sedentar-
ization policy . Fig. 1 shows the location of the four
villages within Lang San and Luong Thuong communes
of Na Ri District, along the Khuoi Sung and Ngan Son

rivers. We selected these villages to cover a large range
ARTICLE IN P RESS
Khuoi Sap
Nam Ca
Na Hiu
Khuoi Noc
5km
District boundary
Commune boundary
Hamlet
Village boundary
River, stream
Road
Case study site in Na Ri District
Bac Kan Province
Vietnam, with
province boundaries
Fig. 1. Study site in Bac Kan Province.
J C. Castella et al. / Land Use Policy 23 (2006) 147–160148
of diversity in ethnic composition, natural resource base,
and livelihood systems. The objective of the case study
was to analyse the way that the same land policies
resulted in notably diff erent impacts on villages even in
a limit ed geographic area, due to the initial diversity of
village situations. Our fieldwork was performed over the
course of several stays in the villages in 1999 and 2000.
We combined informal interviews with direct observa-
tion. In addition, we conducted an exhaustive survey of
all village households (n ¼ 106), allowing us to system-
atically complete the qualitative data collection.

We analysed the impact of forestland allocation on
farmers’ livelihood systems and trajectories of changes
in land use by eliciting local people’s explanations of
why they had changed agricultural practices, by
examining forest tenure rights on the ground, and by
investigating, and when possible, ruling out other causes
of change in land-use practices. Some evidence on
changes in land use/land cover at the village and district
scales is reported in more detail in Castella et al. (2002).
Allocation of forestland in the mountain areas of northern
Vietnam
The objectives of the allocation
Fixed settlement of shifting cultivators: In the moun-
tain areas of northern Vietnam, the predominant upland
production system is swidden cultivation (Do, 1994).
Swidden cultivation involves burning a section of forest
and then growing crops (usually upland rice) in the rich
soil that is left behind. The quality of the soil
degenerates quickly, so after several years of cultivation
the land is left fallow for a much longer period so that
the forest can regenerate (Husson et al., 2001; Roder,
2001). The particulars of swidden cultivation differ from
one group and from one region to another, but share the
common principle of leaving the land to regenerate. In
this paper, we distinguish rotational cultivation (fixed
house, shifting fields) from migratory cultivation (shift-
ing house, shifting fields), which complies with the need
of swidden cultivators to search for mature forests to
clear.
The desire to settle mountainous populations in fixed

areas is not new in Vietnam. Suc h policies first appeared
in the colonial era, where settled populations were
necessary for political control and taxation. Further,
one of the objectives for the institution of the
cooperatives in the mountainous areas was bringing
nomadic populations down from the mountains, and
encouraging them to participate in lowland ricefield
cultivation ( Dang, 1991). Before the collectiv e period,
certain ethnic groups were the predominant cultivators
of lowland rice, while others engaged in swidden
cultivation. During the collective period, they all
engaged to some extent in swidden cultivation, the
former in the form of rotational cultivation and the
latter as migratory cultivation. Today, such a distinct ion
is no longer possible—swidden cultivators are rather
those who have been excluded from land allocation,
regardless of ethnicity.
Despite its diversity, until recently all swidden
cultivation was considered to be similar and equally
destructive (Do, 1994; Morrison and Dubois, 1998). The
State views migratory cultivators as leading a precarious
existence and harming forest resources (Dang, 1991). By
settling these farmers on allocated pieces of land, the
State hoped to end their slash-and-burn practices and
encourage them to develop stable and fixed production
systems (e.g., perennial plantations) that would allow
them to prod uce more and consequently earn more.
Protection of forest resources and the battle against
deforestation: Forest cover in Vietnam has decreased
dramatically in recent decades, from 45% of the

country’s area in 1943 to only 28% in 1991 (of which
only 10% was primary forest; Vo, 1998). These
percentages correspond to a reduction in natural forest
area by 350,000 ha/year over the last 25 years. Defor-
estation has been accompanied by the appearance of
severely eroded cleared lands, which according to some
estimates covered up to 40% of the country in 1990 (Vo
and Le, 1994).
While the causes of deforestation include war, timber
exploitation (Poffenberger et al., 1997), relative land
scarcity due to increasing population and to regulations
on access to resources, the State placed the blame for
deforestation squarely on the backs of the migratory
cultivators. The State has long perceived swidden
cultivation (ra
ˆ
y in the Kinh language) as an ‘‘irrational’’
technique, a backward system that must eventually
evolve into sedentary cultivation: ‘‘In Vietnam, since
1968, the Government has combined sedentarization and
agricultural cooperation campaigns in search of a rational
use of forest resources, to help mountain populations
assist in mass agricultural production and put an end to
their outdated raˆ y practices’’ (Dang, 1991, p. 26). It is
worth noting that if a burned field is left fallow
for enough time after cropping, then the soil and forest
can regenerate sufficiently to provide for the next cycle
of cultivation. Thus under the right conditions, particu-
larly a combination of low population pressure and
regular migration, swidden cultivation is sustainable (De

Rouw and Van Oers, 1988; Mazoyer and Roudart,
1997). However, in the period following the coopera-
tives, low population pressure was no longer a
characteristic of northern Vietnam. Allocating the
forestland an d obliging nomadic peoples to practice
sedentary production would, the State hoped, result in
an end to the deforestation process begun decades
earlier.
ARTICLE IN P RESS
J C. Castella et al. / Land Use Policy 23 (2006) 147–160 149
Increasing production through regional specialization:
Resolution 10 aimed to transform the economy from a
focus on self-sufficiency to a focus on the national
market: ‘‘The renovation of economic management must
meet the following needs: [y] transform our agriculture,
still autarkic in many regions, into specialized goods
production based on the natural, social and economic
characteristics of each region, thus modernizing socialist
agricultural production. [y]’’ (Resolution 10). For the
agricultural sector, the State envisioned a system of
regional specialization, with intensive rice production in
the delta regions complemented by large-sca le sylvicul-
ture and animal husbandry in the mountainous regions.
These sedentary production systems would require
sedentary populations to manage them.
Fig. 2 summarizes the rationale behind the forestland
allocation. The granting of individual property rights on
sloping land put an end to the free access system. Free
access has often been identified as the primary cause of
the ov er-exploitation of forest resources that had taken

place, as farmers can gain individual profits from the
land, while the costs of such exploitation are shared
collectively (Hardin, 1968). By granting renewable land-
use rights, the State hoped to transfer responsibility to
the individual, who would reap the benefits but a lso pay
the costs associated with the exploitation of forestland.
Individual responsibility would give farmers the needed
incentive to make ‘‘rational’’ use of their land, leading to
the protection of forestland (arrow 1 in Fig. 2). Private
ownership would also encourage individuals to invest in
their forestland and develop perennial plantations (2).
Developing commercial perennial crops would lead to
both an increase in tree cover and additional income for
households, thus improving living conditions (3).
Assuming that poverty is what motivates farmers to
adopt short-term unsustainable survival strategies, the
economic gains from perennial plantations would allow
better protection of the forest (4). Further, the migration
associated with shifting cultivation would no longer be
necessary, tying improved living standards to policy on
sedentary livelihood systems (5). Sedentary livelihood
systems reinforce individual responsibility for resources
and provide an incentive for the development of
perennial plantations (6).
The allocation process
Forestland allocation began in 1992, and was still
underway in 2002 in some remote communes. Resolu-
tion 10 (April 1988) defined the main features of land
allocation, with the 1993 land law and Decree 02-CP
(January 15, 1994) supplying additional details.

The allocation policy is also closely associated with the
forest protection and de velopment code implemented in
1991.
The 1993 land law (Article 43) defines forestland as
‘‘all land identified as being destined for sylviculture,
natural forest regeneration, reforestation, timber, nur-
series, forestry research and experimentation’’. Because
forestland was defined according to planned future use
rather than present use, the policy resulted in classifica-
tion as ‘‘forests’’ of some lands that were currently being
cultivated with annual crops and had been for dozens of
years. This categorization included the majority of
sloping land in the mountainous regions.
Forests were classified into three distinct types:
 Protected forest, for the preservation of water
resources; the prevention of erosion, natural disasters,
and climatic risks; and the overall protection of the
environment.
 Special-use forest, intended for the conservation of
nature and of plant and animal species; scientific
research; and the protection of historic, cultural and
tourist sites.
 Production forest, used primarily for timber and
other forest products, and associated with the other
types of forest to protect the environment.
ARTICLE IN P RESS

Fig. 2. Rationale for forestland allocation.
J C. Castella et al. / Land Use Policy 23 (2006) 147–160150
The land-allocation process consisted of a series of

meetings, beginning at the district People’s Committee
and pro gressing down the administrativ e hierarchy to
each individual village. The district developed an
allocation plan, which delineated the areas to be
classified as one of three types of forest (protection
forest, production forest, special-use forest). The plan
was then disseminated through the communes to the
villages, where each household who wanted to receive a
plot of forestland had to fill in a request form, which was
sent on to the Forest Service. The Forest Service then
measured and classified each individual plot. In many
cases, the process only ratified the informal rights that
already existed, the clear er’s rights, for households often
requested the plots that they wer e currently developing
or cultivating. The allocations thus legitimized tradi -
tional land-use rights. That said, a number of conflicts
also arose from the allocations, with households
occasionally evicted from the process by a village
head trying to privilege his own lineage. Such conflicts
are reminiscent of those that arose with the allocation
of irrigated ricefields at the end of the collective
period and the ancestral reclamation movement
(Nguyen and La, 1999). Once all conflicts had been
cooperatively resol ved at the village level, the Forest
Service integrated the information into a land map and
gave certificates of land-use rights to households.
Following the granting of land-use certificates, a meet-
ing was held in each village to address the issue of forest
protection. At this meeting, the Forest Service explained
the policy and regulations concerning forest protection

and development, and each household possessing forest-
land had to sign an agreement to treat their land
accordingly. Each village was then able to develop its
own system of forest management, protection and
development based on its own particular circumstances.
However, the rules implemented by all villages
were based on model regulations supplied by the
Forest Service.
The rule development system appeared to be par-
ticipatory and flexible, but in reality it was a
top-down procedure. Instructions were passed down
from one hierarchical level to the next with minimal
changes. It was not uncommon to find that the
management rules of one village were nothing
more than a photocopy of those of the ne igh-
bouring village, with the names of the villa ge and its
leaders changed. The top-down implementation of
the system favoured neither local participation nor
ownership.
The effects of the forestland allocation
We examined the effects of the forestland allocation
through a case study of four villages in Bac Kan
Province (Fig. 1 and Box 1). In Na Ri District as a
whole, the allocation process began in 1992 and ended in
2000. Spe cifically, the allocation of land-use rights in
Lang San and Luong Thuong communes took place from
1997 to 2000.
The villages studied were all founded relatively
recently, the oldest being Nam Ca, settled just
100 years ago. All of the villages with the exception

of Na Hiu were settled for their abundant forest
resources,
2
and even Na Hiu now relies on the forest
more than ever, as two-thirds of its ricefields
were repossessed during the Tay land reclamation
movement. Although paddy rice has gained popularity
as a means of subsistence, there is almost no remaining
land in the region suitable for terracing for new
paddyfields. The majority of households cannot produce
enough paddyland rice to feed themselves, and have
adopted production strategies based on rotational
swidden cultivation.
Effects on sedentary livelihood systems
Clarification of individual and village land boundaries:
The nature of the forestland allocation process necessi-
tated the clear demarcation of each individual’s property
rights. In doing this, the Forest Service implicitly defined
all village and commune boundaries. ‘‘Under the old
system, individuals of the neighbouring villages could clear
lands in the territory of our village for swidden cultivation,
and the village authorities were powerless to stop them.
This often led farmers to blame neighbouring villages for
forest deterioration’’. Under the new land allocation
system, farmers are fully responsible for their own
land, and should be satisfied that no household other
than their own can harm this resource. What was
designated as ‘‘village territory’’ now comprised the set
of land plots allocated to the households of a village,
together with plots placed under direct village owner-

ship. Each village territory defined clear boundaries for
agricultural and forestry activities. By definition, the
existence of individual land-use rights excluded indivi-
duals from use of land to which they did not have
formal land-use rights. Thus, the new land policy
abruptly ended the traditional free access rules that
had existed before. Each village and household now
possessed land that was protected from outsiders, but
that same protection now also limited the scope of their
actions.
The end of free access was of critical importance
to those groups with production systems based on
ARTICLE IN P RESS
2
Although settlers chose the locations for these villages for their
abundance of ‘‘primary’’ forest, the settlers did not realize that the
forest was not more than 20 years old. Aerial photographs from 1954,
1977 and 1998 show that although the areas were densely forested in
1954, they had been cleared almost entirely at least once by 1977.
J C. Castella et al. / Land Use Policy 23 (2006) 147–160 151
migratory cultivation, particularly the H’mong and
Dao people in the study area. Under the new land
policy, these groups could no longer migrate in search
of mature forests to clear, as other individuals or
villages now owned the right to use all forests. At first
glance, one of the goals seemed to have been accom-
plished: previously nomadic populations were settled
permanently. How ever, it is necessary to distinguish
between sedentary housing and sedentary production
systems.

Sedentary housing versus sedentary production systems:
Effective transformation from a migratory to a seden-
tary society has at least two different levels of indicators:
 sedentary housing, indicated by kinds of homes and
the materials used to construct them; and
 sedentary production systems, indicated by an agri-
cultural system that is ecologically and economically
sustainable within a fixed territory (without requiring
regular migration).
ARTICLE IN P RESS
Box 1
Main characteristics of the four villages under study.
In 2000, Na Hiu village comprised 9 households spread out over a large area. The village population
was made up of 62 inhabitants of diverse ethnic backgrounds, including Dao Cooc Mun, Red Dao and
Nung. The villagers shared a history of coming to Na Hiu in search of cultivable land. The first arrived
in 1985 and redeveloped a group of paddyfields that had been abandoned for more than 20 years. Most
of these paddyfields have since been repossessed by the original owners. The remaining inhabitants
(the later arrivals) were not able to acquire paddyfields and had to rely on the forested slopes.
Although Na Hiu village was founded because of its abandoned ricefields, its inhabitants have
increasingly turned to swidden cultivation and hunting and gathering as a mean of survival. Even the
inhabitants who arrived in the hope of developing new paddyfields now depend on the forest for
survival, as terracing requires considerable investment in labour. Many paddyfield terraces have been
under construction for years and remain unfinished.
Nam Ca village was founded over 100 years ago by a Tay, whose descendants now make up the 11
households and 57 inhabitants of the village. He chose a location favourable for the development of
irrigated paddyfields, the basis of Tay production systems in northern Vietnam (Castella and Erout,
2002).
However, the lowlands are now saturat ed, and no more paddyfields can be constructed. This has led
the village to develop new inheritance rules forcing all but the eldest son in each family to emigrate.
Households rely primarily on paddy-rice cultivation, with upland glutinous rice as a supplement. The

forest is also used for free-grazing buffaloes in the winter season.
In the spring of 2000, most households attempted a spring-season cycle of rice. The trials were not a
complete failure, but produced mediocre yields due to cold weather.
Khuoi Sap is the village farthest from the road, the administrative centre, and the marketplace (about
an hour and a half away by foot). In 2000, Khuoi Sap comprised 17 households and 95 people, all Red
Dao with the exception of one Tay family. The village encompasses four hamlets settled by two
different family groups.
The continuous spread of the village was motivated by the abundance of old-growth forestland that
could be cleared and used for the production of upland rice. Village inhabitants initially opened
swidden fields very close to their houses, but with time villagers cleared areas farther and farther from
the settlements, now reaching distances of up to one or two hours on foot. The village production
system is based on swidden cultivation, beginning with several years of upland rice follow ed by
cassava and maize. This is complemented by hunting and gathering. Only two families own
paddyfields, and these only cover a small area (1000 m
2
each).
Khuoi Noc village occup ies the entire northern area of Luong Thuong Commune. The village is
composed of 394 H’mong inhabi tants forming 70 households spread among six hamlets. When the
first households (about 22) arrived from Cao Bang Province in 1983, the land was unoccupied.
Founded because of the area’s abundant primary forests, the village now has an economy based on
the swidden cultivation of upland rice and cassava, narrow fields of maize on the flat lands along the
river, and hunting and gathering. Only 5 of the 70 families own irrigated ricefields, which they built
themselves; the combined surface area is only 8100 m
2
. Some families engage in gold mining to
complement their income.
J C. Castella et al. / Land Use Policy 23 (2006) 147–160152
The Tay village of Nam Ca offers an example of
sedentary housing. The people of the village have lived
in the same location for over 100 years and have built

houses on wooden stilts with tile roofs. Such houses
require substantial investmen ts in both materials and
labour. Migrating populations tend not to build such
houses, but usually live instead in houses made from
bamboo, built on the ground or on simple stilts. This
type of house is mostly found in Dao and H’mong
villages. For a family to build a wooden-stilt house with
a tile roof is both a sign of relative wealth and an
indicator that they intend to stay in one place for several
years at least, even if such structures can occasionally be
sold upon departure.
The second, more fundamental change in the liveli-
hood systems is the transition from a system based on
swidden cultivation to a system based on irrigated
ricefields, whether in flatlands or terraces on sloping
land surrounded by intensive (long cultivation, short
fallow period) upland fields. Almost all farmers in the
study area had this transition as one of their goals, even
before the new land policy. Even though the construc-
tion of paddyfields or terraces is extre mely labour-
intensive, farmers consider paddy rice cultivation to be
easier than sloping-la nd cultivation. This is particularly
the case when the swidden fields have to be cleared from
degraded forests, requiring increased time investment in
weeding for lower and highly uncertain yields (Husson
et al., 2001 ; Roder, 2001). The village headman of
Khuoi Sap stated that: ‘‘Everyone in the village would be
ready to buy a paddyfield like I did, or expand the paddy
area as I did, but we do not have any more suitable land
for terracing’’. Although the majority of households

have identified paddyfields as a major objective, few
have achieved this objective. Reasons put forward by
the farmers surveyed include inadequate access to
flatlands, inadequate water sources for irrigating ter-
races on sloping lands, and insufficient labour for the
construction of paddyfields or terraces.
The villages with economies based on irrigated rice
(Nam Ca and part of Na Hiu) had both sedentary
housing and sedentary production systems even before
the forestland allocation. Consequently, transformation
to a sedentary society was not an issue for them. The
remaining villages (Khuoi Sap, Khuoi Noc and the rest of
Na Hiu) have been converted to a sedentary existence
only in terms of their housing. Even after the
implementation of the new land policy, the majority of
production systems in the studied Dao and H’mong
villages continue to be based on swidden cultivation, a
system that is sustainable only when accompanied by
regular migration and under the condition of low
demographic pressure (Mazoyer and Roudart, 1997).
These populations now find themselves in a very delicate
situation, with a production system that is poorly
adapted to their institutional environment. At the same
time, their sedentary housing reveals a de facto
transformation to a sedentary lifestyle. This is not a
true transformation of society based on a new livelihood
system that no longer requires migration. Rather, it is
merely a transitional stage induced by the loss of the
possibility to migrate but not yet truly sustainable.
Effects on forest protection

Individual responsibility for forests: Farmers fre-
quently mentioned the rapid decrease in old-growth
forest that has taken place in recent years, and state that
the over-exploitation of the forest could have serio us
effects on their families. They stated that ‘‘four or five
years ago farmers in Khuoi Noc reached the last forested
areas suitable for swidden cultivation within their village
boundaries’’. This statement was confirmed by our
analysis of a chronological series of land-use maps
(1954, 1977, 1998) that revealed the rapid decrease in
forest cover between 1977 and 1998 in the study area
(see land-use maps of Luong Thuong and Lang San
communes in the digital atlas in Castella et al., 2003).
Deforestation was associated with a sharp increase in
upland crop areas and shrub land, whi ch is typical of the
rapid expansion of swidden cultivation that occurred
everywhere in Bac Kan Province before the allocation of
forestland (Castella et al., 2002). As a consequence,
forest wildlife, another important component of peo-
ple’s livelihood, has also rapidly disappeared.
The populations of the study villages wer e mostly
satisfied with the individual land allocations (Table 1),
though not necessarily for the same reasons. Most
farmers received the land that they requested, that which
would have been theirs by traditional rights. Some were
happy to own land near their home, as this either
reduced the walking time to swidden fields or allowed
for the continual surveillanc e required for the develop-
ment of perennial plantations. Others were happy with
the quality of the land (generally when old-growth forest

was involved). The reasons for individual satisfaction
tended to reflect people’s planned use of the forestland.
Consequently, many were less enthusiastic about the
accompanying forest protection policy as an outright
ban on forest exploi tation would jeopardize their food
security. Farmers in the study area are particularly
conscious of the need for forest protection, as their
livelihoods depend on the survival of this resource base.
The announcement of the forest protection policy,
particularly the ban on the clearing and burning of
forestland, was met with apprehension by a large
number of farmers who had no short-term alternative
ways to feed their families. Given their reliance on the
forest, farmers are eager to learn about initiatives to
protect it. New circumstances will inevitably force them
to develop new production systems to compensate for
the ban on swidden production.
ARTICLE IN P RESS
J C. Castella et al. / Land Use Policy 23 (2006) 147–160 153
The end of swidden cultivation: Farmers are experien-
cing declining yields from swidden fields and increasing
labour requirements for weeding. Fig. 3 shows that the
number of newly opened swidden fields has decreased
over the last 2 years. Our field survey indicates that a
large number of households have continued to cultivate
swidden fields opened in previous years. Farmers
explained that this change is a direct consequence of
the forestland allocation and forest protection policy as
it is no longer possible to open new fields. The
cultivation of old swidden fields is parti cularly striking

in Khuoi Noc, where some swidden fields have been
cultivated continuously since 1990, while in Na Hiu and
Khuoi Sap the oldest fields date from 1994 and 1995,
respectively. In addition, the yields of newly cleared
swidden fields (defined as the rice harvest in proportion
to the quantity of seed sown) have been declining in
Khuoi Noc. For the village head of Khuoi Noc ‘‘the low
and still diminishing yields in my village are the result of a
forest that is becoming thinner and thinner while the
population of swidden cultivators, swollen by immigration,
continues to rise’’. The surprising (and alarming) result
is that fields in their first year of cultivation produce
yields that are even lower than fields that have been
cultivated for 7 years (Fig. 4). We checked with the
farmer that this result was not imputable to an
exceptional bad climatic season or unfavourable field
topography but to the ‘‘repeated use of the same pieces of
land, which exhausts the soil’’ as an older farmer in Khuoi
Sap explained.
The average surface area of swidden fields
3
opened 3
years earlier has not changed substantially (Fig. 5).
Taken with declining yields, this is a clear sign that,
confined to fixed terr itories under the new land policy ,
production systems based on swidden agriculture are
slowly failing.
ARTICLE IN P RESS
Table 1
Farmers’ satisfaction with forest land allocations (expressed answers only)

Village
Na Hiu Nam Ca Khuoi Sap Khuoi Noc Total
Expressed satisfaction Very satisfied 2 6 8 16 32
Fairly satisfied 3 3 3 31 40
Not satisfied 1 0 1 13 15
Reason for satisfaction Proximity 5 1 5 23 34
Land quality 0 5 5 16 26
Other 1 3 1 6 11
Fig. 3. Number of upland fields under cultivation in 2000, plotted
against year of opening. * Swidden fields opened in 2000 under
cultivation in 2001.
Fig. 4. Yield of upland rice crops cultivated in 2000, plotted against
year of opening. * Yields are expressed as multiples of the quantity of
seeds sown. For example, if a farmer sowed 10 kg/ha and harvested
150 kg/ha, then the yield would be 15 (150/10=15).
3
Because it is difficult to measure actual plot sizes on sloping lands,
we estimated the surface area of swidden fields by the kilograms of
seeds sown per plot, using a conversion factor of 10 kg seed sown per
hactare.
J C. Castella et al. / Land Use Policy 23 (2006) 147–160154
Effects on agricultural production and economic
development
The new land policy was intended to result in both
intensification of paddyland production and the use of
the uplands for perennial plantations and livestock
development. Both outcomes were intended to increase
the income and thus the quality of life of farmers. In
reality, the new system exacerbated the problems of
swidden cultivation without providing farmers with an

adequate alternative source of income.
Migratory cultivators facing crisis: The reduction in
upland rice yields implies a reduction in the living
standards of the populations who rely on swidden
cultivation. In 1999, in two out of the three villages that
formerly practiced migratory cultivation, the average
rice yield was under the 250 kg/person self-sufficiency
threshold (Table 2). Indeed, a substantial number of
households faced rice deficits in 1999 (Table 3). This
does not necessarily mean that they went hungry, but
instead purchased rice, or relied on non-rice staples.
However, as most of these complementary sources of
food are traditionally harvested from the forest ecosys-
tem, their quantity and quality were deeply affected by
deforestation and access to these resources was affected
by the forestland allocation to individuals. The Dao
(which means ‘‘people of the forest’’) people recognize
that ‘‘beyond the negative impact of the imposed
sedentary life on their traditional cropping systems the
new land policy has transformed other key activities such
as animal husbandry, hunting, gathering, and collection of
timber and firewood’’. The multiplicity of land uses and
local rules over natural resource management that
overlap on the same forestland creates a very complex
situation in the relationships between people and their
environment. Considerable variability was observed
between locations with respect to specific village
regulations over individual and or collective manage-
ment of forest resources (Pandey and Dang, 1998;
Castella and Dang, 2002).

In sum, a general feature of the villages predomi-
nantly populated by swidden cultivators is a high
percentage of households facing food deficits, which is
indicative of a production system in crisis, a system that
is no longer adapted to its environment. We will now
discuss how farmers have adapted their production
systems in response to the crisis, and more generally how
they have adapted their systems to the new land policy.
Intensification of lowland production: The ban on
opening new upland fields was intende d to encourage
households to focus their energies on the lowlands.
Lowland rice production can be increased either by
creating new pa ddyfields; or intensifying production in
existing paddyfields (e.g., by introducing a spring-season
rice crop); or both. New paddyfields have recently been
constructed in Na Hiu (eight households) and Khuoi Noc
(six households), while two households from Khuoi Sap
purchased paddyfields. However, most of this increase
happened between 1987 and 1993, before the new
forestland policy was implemented. Our interview s
ARTICLE IN P RESS
Fig. 5. Surface area of upland ricefields cultivated in 2000 plotted
against year of opening. * Plot size is expressed in kilograms of seeds
sown per plot (the sowing density commonly used as a reference is
10 kg seeds/ha).
Table 2
Indices of upland rice production in 1999, by village
Index of upland rice
production
Village

Khuoi Sap Na Hiu Khuoi Noc
‘A’=quantity seed
sown/person (kg)
16 12 13
‘B’=average yield 18 18 14
‘A Â B’=average
production/person (kg)
288 216 182
N.B. Yields are expressed as multiples of the quantity of seeds sown.
Table 3
Indices of rice deficit in 1999, by village
Index of rice
deficit
Village
Khuoi Sap Nam Ca Na Hiu Khuoi Noc
Percentage of
households
experiencing
deficit (food
requirements
exceeded rice
production)
59% 20% 87% 67%
Average
length of the
deficit
(months)
2253
J C. Castella et al. / Land Use Policy 23 (2006) 147–160 155
indicated that the decreasing availability of old-growth

forest was already encouraging individuals to move to
the lowlands; land allocations were only an added
incentive. As we mentioned earlier, the paddyland
within the study area is already almost saturated; little
flatland remains that could still be developed. In spite of
a growing population and the progressive subdivision of
fields (because of inheritance), no new ricefields have
been created in Nam Ca since 1971. With the lowlands
already saturated with ricefields, the only possibility for
increased production in Nam Ca is to find ways to
intensify rice production in existing ricefields. In the
spring of 2000 Nam Ca farmers began to experiment
with increasing production by adding a second, spring -
season rice crop. Farmers in Nam Ca justified this
intensification of paddy-rice production as a conse-
quence of the increased demographic pressure within the
village co mbined with the restrictions imposed by the
village authorities on clearing new lands on the hillsides.
This illustrates the interdependence of paddylands and
sloping lands, demonstrated by Castella and Ero ut
(2002).
Perennial plantations: State policy suggests a future of
uplands brimming with fruit tree plantations, bringing
wealth or at least self-sufficiency even to farmers who
lack paddyfields. With swidden cultivation no longer
feasible, upland rice production will decline or cease,
and perennial plantations are expected to fill the gap.
Plantations are indeed being developed in the study
area, but they are far from an all-purpose solution to the
difficulties faced by farmers. Cinnamon and anise

plantations account for 70% and 15%, respectively, of
the trees planted in the study area, while fruit trees (e.g.
plum, orange, apricot) make up the remaining 15%
(Fig. 6). For most farmers, tree-crop production is still
experimental, as revealed by the fact that the majority of
plantations are very small—the trees are located in the
vegetable garden or very near to swidden fields, and the
harvest is mostly consumed by the household. ‘‘Planta-
tions represent a highly uncertain source of income
because of village remoteness and market uncertainty’’
mentioned the few household heads who engaged in
large scale cinnamon plantations in Na Hui with the
help of the World Food Programme thanks to their
proximity to the main road as compared to the three
other villages.
The new land policy successfully destabilized migra-
tory cultivation, but perennial plantations and lowland
intensification have not been sufficient to meet the needs
of swidden farmers. Struggling to feed their families,
they are rapidly experimenting with new possibilities,
from aquaculture and animal husbandry (mainly
poultry, pigs, and cattle) to the testing of any cropping
innovation that shows promise. Another option is to
move to a new location in search of better environ-
mental conditions for agriculture or opportunities for
non-agricultural income.
New kinds of migration: Migrating to new communes
or districts within the province is no longer possible
because all the land has already been allocated to
villagers. However, households without opportunities in

their current locations can move their houses closer to
forestland they own within a village. In Khuoi Sap and
Khuoi Noc, some households have created new hamlets,
and others have moved their houses to their more
remote upland plots to develop irrigated paddyfields or
fishponds. A few other families have moved closer to the
road to engage in non-farming activities such as small-
scale commerce (e.g., selling consumer goods or
agricultural inputs) or motorcycle taxi services, as well
as to give their children better access to schools.
The most dramatic option is migration to southern
Vietnam, where New Economic Zones are being
developed. Supported by State subsidi es, families can
relocate to work in the new industrial plantations
(mostly coffee and rubber) in these zones, located in
the Central Highlands region or in the south of
Vietnam. The New Economic Zones hold a particular
appeal for highland peoples; the prospect of having
sufficient land to grow commercial crops that can be
readily sold leads many to dream of a better future. Ten
households from Khuoi Sap have already made the long
journey. Between 1991 and 1996 nearly 1.5 million
people migrated to the Central Highlands (Vo et al.,
1999). The journey south is costly; farmers interviewed
estimated that such an undertaking would cost at least
ten milli on dong. Migration is thus only possible for
households who have been able to accumulate capital,
but who nonetheless find themselves in sufficiently
difficult circumstances to justify leaving everything
behind for the great unknown. Households who have

difficulty meeting their most basic needs are more likely
to consider small-scale migration (often within a village)
to improve their situation.
ARTICLE IN P RESS
Fig. 6. Plantation composition per village.
J C. Castella et al. / Land Use Policy 23 (2006) 147–160156
The departure of some households for South Vietnam
occasionally gives other households the valuable oppor-
tunity to buy paddyfields from the departing families. In
2000, four households in Khuoi Sap purchased ricefields
in a neighbouring commune from Tay families who were
leaving for a New Economic Zone.
Discussion
Land allocation and trajectories of change in land use
Our study of four remote mountain villages in
northern Vietnam revealed many of the consequences
of individ ual forestland allocations. The biggest distinc-
tion in effects is between two kinds of farmers: those
who have their food needs met by paddyland rice, and
those who do not.
The allocation of forestland and the associated
protection policy did not endanger the land-use systems
of farmers belonging to the first group. The focus of
their production continues to be paddylands. With
irrigated rice forming the basis of their household
consumption, these families, often with houses and
fields established decades earlier, had already ceased to
rely on sloping lands to feed themselves. They still clear
sloping lands on the edges of ricefields both to allow
more sunshine on the ricefields in the narrow valley

bottom and to produce complementary crops (upland
rice varieties, cassava, maize, etc.). Raising pigs and
selling surplus rice provides enough money for house-
hold expenses and investment in the form of water
buffaloes. The pressure exerted on these sloping lands
depends on local circumstances, primarily on paddy rice
production. During the times of cooperative crisis,
sloping lands were often subjected to an increase in
use, leading to significant clearing and deforestation.
Such lands were also used for private gain under a
political system that sought to make individual
accumulation of wealth impossible. Uplands were
privately owned according to customary rights that
granted temporary ownership to the individual
who cleared any given plot. The land allocation
law served only to permanently confirm informal
land-use rights that already existed. With new obliga-
tions to protect and reforest upland areas, households
returned their focus to the ricefields. Paddy-rice growers
are now mostly concerned with develop ment issues in
the lowlands, for example, paddyfield intensification,
new rice varieties, and mechanization. Although their
rice needs are met, households in this category continue
to seek new opportunities to accumulate income
(Castella and Erout, 2002). Plantation projects offer a
highly unc ertain future, but as they are supported by
outside development agencies, they cost little to imple-
ment. Furthermore, poultry, fish, pig and cattle raising
are successive stages in a process of progressive
accumulation of investments through animal husban-

dry, though market access and demand remain un-
certain.
On the other hand, households who do not own
irrigated ricefields, or whose food needs are not met by
their irrigated ricefields, were significantly and severely
affected by the changes in the forestland use and
protection policy. The most immediate consequence of
the new land policy was the clear demarcation of
individual land-use rights in the uplands, which pre-
vented migration and thus increased the population
pressure on the land owned by these individuals on the
hillsides. The implicitly enforced sedentarization
has made swidden cultivation unsustainable. The
upland cultivation that is currently continuing is
characterized by decreasing yields and deteriorating
forest and soil quality—the signs of land degradation
are already apparent in the villages wher e this type of
household is in the majority (Castella et al., 2002). No
longer able to migrate to new areas while leaving old
areas to regenerate, these farmers find themselves in an
environment where their traditional land-use system is
no longer an option. Unlike paddy-rice growers,
swidden cultivat ors did not have a backup system to
turn to in times of crisis. From a rather homogeneous
system, these farmers have moved to small-scale
exploration of income generating activities, for example,
the development of terraces, small animal husbandry
and off-farm activities. Individual innovations are
multiplying as farmers experiment with new techniques
to meet their needs and readjust their systems as new

problems develop.
As a result of the history of human settlement in Bac
Kan Province, these two groups of households are often
distributed in distinct villages (Mellac, 2000; Castella et
al., 2002). As a consequence, the differentiation pattern
observed between households is also valid between
villages. ‘‘Paddy rice cultivation’’ villages (e.g. Nam Ca)
as opposed to ‘‘swidden cultivation’’ villages (e.g. Khuoi
Sap, Khuoi Noc) can be described as extremes in a
continuum with inter mediary situations such as Na Hui.
Following land allocation, the forest regenerated in
the former villages, usually occupied by the Tay
ethnic group, while it had almost disappeared from the
latter, predominantly inhabited by Dao and H’Mong
ethnic groups. As the Tay represe nt 75% of the
population in Bac Kan Province, this trend resulted
in a major change in land cover, with an overall
trend towards forest recovery at the district and
province scale since 1995 and spots of increased
deforestation limited to a few villages of swidden
cultivators (Castella et al., 2002). The land allocation
policy was thus successful in stopping galloping
deforestation but some groups were excluded from the
path to economic development.
ARTICLE IN P RESS
J C. Castella et al. / Land Use Policy 23 (2006) 147–160 157
Lessons for development interventions/policy implications
However, the new land policy has created a new
opportunity for de velopment intervention. Swidden
cultivators have long been the poorest of the rural

people in northern Vietnam. Frequent migration has
made them a difficult target for development projects.
Settled in a fixed location, these people are now in more
dire straits than ever before, but are also now a clear and
willing target for new innovations. These households do
not oppose the notion of a settled existence; after all,
farmers of all ethnic backgrounds share the strategy of
purchasing paddyfields wherever possible. Farmers
associate paddyland rice with a feeling of food security,
in contrast with swidden cultivation and its extremely
unstable yields. Thus, the possible technical innovations
are the same for the two groups of households/villages, a
tough policy posed in most urgent terms for those who
still rely on swidden cultivation for their survival.
Farmers in the study area have already initiated a
wide range of innovations aimed at lessening agricultur-
al pressure on forestland while meeting the basic needs
of their families. Research can help these populations in
their solution-finding process, not only by introdu cing
technical innovations but also by identifying the
necessary institutional environments to make these
innovations economically feasible. Two levels are
distinguished in Table 4: the paddies and the hillsides,
where innovations may improve the sustainability of
land-use systems as a whole.
Conclusions
At first glance, the forestland allocation policy seemed
to be effective in transforming mountainous populations
to a sedentary livelihood system. The new context of
stable land rights for individuals also made it feasible to

grow some perennial plantations. Further, in the period
since the forestland allocation there has been a marked
increase in overall forest cover (Sikor, 2001; Tachibana
et al., 2001; Castella et al., 2002). Our findings support
the generally advanced argument in favour of individual
property rights that the transfer of responsibility to an
individual, also gives the farmer an incentive to make
‘‘rational’’ use of the land and to protect the resource
because only he/she will reap the benefits but also pay
the costs (associated with the degradation of the
resource). Indeed, the land allocation policy was
concomitant with a general trend towards paddyland
intensification (introduction of a spring season rice crop,
Castella and Erout, 2002) and forest regeneration
(Castella et al., 2002). However, the causal relationship
had not been not fully demonstrated. In some cases, the
trend to pad dyland intensification and forest protection
preceded the implementation of the land allocation
ARTICLE IN P RESS
Table 4
Potential innovations to relieve agricultural pressure on forestland
Type of innovation Potential constraints
Low-lands Intensification by Increasing the number of
crop cycles per year (Spring rice or maize,
Winter potato, sweet potato, etc.)
Biophysical conditions of the plot (irrigation, micro-climate, etc.)
Access to cultivars and inputs
Market for product
Mechanization Limited so far because of the lack of capital. Only farms with
large paddies and limited labour force are concerned (Sadoulet et

al., 2002).
Extension of existing paddies or construction
of new ones
Suitable land very limited; this option is becoming marginal
because of the historical settlement of the lowlands
Hill-sides Plantation of perennial crops: fruit trees, etc. Access to market, transportation infrastructure, Initial investment
supported by extension system or projects
Stabilization of rotational cropping systems on
the existing agricultural land on the slopes:
improved fallow management, mulching, direct
seeding on cover crops, etc. (Husson et al.,
2001)
Requires major change in the cropping practices: research still
needed to develop technologies adapted to local circumstances,
Accompanying extension/training activities are necessary
Terracing to develop irrigated paddies Limited by access to water
Terracing dry paddies, would allow
considerable expansion of paddies
Research prospects on ‘‘aerobic rice’’ that grows on flat dry soil,
several technical problems still to be overcome before this
innovation can be proposed to farmers (FEER, 2001; Lafitte et
al., 2002)
N.B.: Most of these innovations have been tested with farmers in the framework of the Mountain Agrarian Systems project between 1998 and 2002
and research is still on-going (Castella et al., 2003).
J C. Castella et al. / Land Use Policy 23 (2006) 147–160158
policy. Our results suggest that forestland allocation was
perceived by farmers as one more incentive to stop
clearing new forest plots, a tough policy with a variable
time frame and with an impact on forest regeneration
that varied with local circumstances: availability of

ricefields to intensify, or suitable land for terracing,
access to market and a road network for the develop-
ment of perennial crops, the existence of effective
control over land clearing and law reinforcement by
local authorities, etc.
However, the relative successes of the land policy
conceal the deeply destabilizing effect it had on swidden
cultivators. Accustomed to livelihood systems based on
regular migration, these farmers now find themselves in
a critical situation, with an unsustainable sedentary
land-use system with rapidly decreasing yields (Jamieson
et al., 1998; Pandey and Dang, 1998; Le and Rambo,
2001). By eliminating free access to sloping lands, the
State has eliminated the possibility of migration, and
thereby put an end to migrator y cultivation. The
problem is that the State has not yet successfully
implemented a feasible alternative. New paddyfield
development is not currently possible in the study area,
and perennial plantations have only been implemented
on a very small scale, if at all. Meanwhile, swidden
cultivation continues, in a Malthusian spiral of impov-
erishment associated with environmental degradation.
Finally, depending on local circumstances and on the
history of population settlement, the same policy can
have completely opposite effects and lead to a Mal-
thusian situation of resource depletion as well as a
Boserupian trajectory marked by agricultural intensifi-
cation and lessened pressure on the forest. The challenge
for sustainable developm ent is to reduce the human and
environmental costs of the necessary transitions.

References
Bergeret, P., 1995. Land policy in Vietnam. Vietnamese Studies 45 (1),
36–37.
Castella, J.C., Dang, D.Q. (Eds.), 2002, Doi Moi in the Mountains.
Land use Changes and Farmers’ Livelihood Strategies in Bac Kan
Province, Vietnam. The Agricultural Publishing House, Hanoi,
Vietnam.
Castella, J.C., Erout, A., 2002. Montane paddy rice: the cornerstone of
agricultural production systems in Bac Kan Province, Vietnam. In:
Castella, J.C., Dang, D.Q. (Eds.), Doi Moi in the Mountains. Land
Use Changes and Farmers’ Livelihood Strategies in Bac Kan
Province, Vietnam. The Agricultural Publishing House, Hanoi,
Vietnam, pp. 175–195.
Castella, J.C., Tronche, N.R., Vu, N., 2002. Landscape changes in Cho
Don District during the doi moi era (1990–2000) and their
implications for sustainable natural resource management in Viet
Nam’s mountainous provinces. In: Castella, J.C., Dang, D.Q.
(Eds.), Doi Moi in the Mountains. Land Use Changes and
Farmers’ Livelihood Strategies in Bac Kan Province, Vietnam.
The Agricultural Publishing House, Hanoi, Vietnam, pp. 149–173.
Castella, J.C., Chabanne, A., Dang D.Q., Ha, D.T., Husson, O., Le,
Q.D., Tran, D.L. (Eds.), 2003. Towards Sustainable Agricultural
Development in Mountain Areas of Northern Vietnam. Experience
from the Mountain Agrarian Systems (SAM) Program in Bac Kan
Province, 1998–2002. Vietnam Agricultural Science Institute,
Hanoi. Multimedia CD-Rom and Web site: wl-
edgebank.irri.org/sam/home_en.html.
Dang, N.V., 1991. La culture sur bru
ˆ
lis et le nomadisme. Etudes

Vietnamiennes 1, 16–28.
De Koninck, R., 1999. Deforestation in Vietnam. International
Development Research Centre, Ottawa, Canada.
De Rouw, A., Van Oers, C., 1988. Seeds in a rainforest soil and their
relation to shifting cultivation in the Ivory Coast. Weed Research
28, 373–381.
Do, D.S., 1994. Shifting Cultivation in Vietnam: Its Social, Economic
and Environmental Values Relative to Alternative Land Use.
International Institute for Environment and Development, Lon-
don, UK.
Donovan, D., Rambo, A.T., Fox, J., Le, T.C., Tran, D.V., 1997.
Development trends in Vietnam’s Northern Mountain Region,
Vol. 1. An Overview and Analysis. National Political Publishing
House, Hanoi, Vietnam.
FEER, 2001. Aerobic rice, a solution to Asia’s irrigation challenge.
Far Eastern Economic Review, September 6, 2001, /http://
www.feer.com/breaking_news/010829.htmlS.
Fforde, A., de Vylder, S., 1996. From Plan to Market: Economic
Transition in Vietnam 1979–1994. Westview Press, Boulder.
Government of the Socialist Republic of Vietnam, 1994. Allocation of
Forest Land to Organizations, Households, and Individuals for
Stable and Long-term Use for Forestry Purposes. Decree No. 02-
CP, 15.1.1994. Official Gazette, Hanoi.
Hardin, G., 1968. The tragedy of the commons. Science 162,
1243–1248.
Husson, O., Ha Dinh T., Lienhard, P., Le, Q.D., Seguy, L., 2001.
Agronomic diagnosis and development of direct sowing techniques.
Upland rice-based cropping systems in mountainous areas of
northern Vietnam. Contribution to the First World Congress on
Conservation Agriculture, 1–5 October 2001, ECAF/FAO, Ma-

drid, Spain ( />2001_PosterMadrid_E.pdf).
Jamieson, N., Le, T.C., Rambo, A.T., 1998. The Development Crisis in
Vietnam’s Mountains. East-West Center, Honolulu, Hawaii, USA.
Kerkvliet, B.J., Porter, D.J. (Eds.), 1995, Vietnam’s Rural Transfor-
mation. Westview Press, Boulder.
Lafitte, H.R., Courtois, B., Arraudeau, M., 2002. Genetic improve-
ment of rice in aerobic systems: progress from yield to genes. Field
Crop Research 75, 171–190.
Le, T.C., Rambo, A.T., 2001. Bright Peaks, Dark Valleys: A
Comparative Analysis of Environmental and Social Conditions
and Development Trends in Five Communities in Vietnam’s
Northern Mountain Region. National Publishing House, Hanoi,
Vietnam.
Mazoyer, M., Roudart, L., 1997. Histoire des Agricultures du Monde.
Editions du Seuil, Paris.
Mellac, G.M., 2000. Des foreˆ ts sans partage—Dynamique de l’espace
et utilisation des ressources dans un district de montagne au Nord
Vietnam. Ph.D. Dissertation, Universite
´
Michel de Montaigne—
Bordeaux III, Bordeaux, France.
Morrison, E., Dubois, O., 1998. Sustainable livelihoods in Upland
Vietnam: land allocation and beyond. Forestry and Land Use
Series 14, International Institute for Environment and Develop-
ment (IIED), London.
Nguyen, V.H., La C.Y., 1999. Cac Quan He Dat Dai va Nhung Mau
Thuan Ruong Dat o Vung Nong Thon Mien Nui Mien Bac Hien
Nay (Land Relations and Conflicts in Rural Mountain Areas of
Northern Vietnam). Uy Ban Dan Toc va Mien Nui, Hanoi.
Nguyen, T.L., Vu V.M., Nguyen V.T., 1995. Land classification and

land allocation of forestland in Vietnam—a meeting of the national
ARTICLE IN P RESS
J C. Castella et al. / Land Use Policy 23 (2006) 147–160 159
and local perspective. Forest, Trees and People Newsletter 25,
31–36.
Pandey, S., Dang, V.M., 1998. A socio-economic analysis of rice
production systems in the uplands of northern Vietnam. Agricul-
ture Ecosystems and Environment 70, 249–258.
Pillot, D., 1995. La fin des cooperatives: la de
´
collectivisation
agricole au Nord Vietnam. Les Cahiers d’Outre Mer 190,
107–130.
Poffenberger, M., Walpole, P., D’Silva, E., Lawrence, K., Khare, A.,
1997. Linking government and community resource management:
what’s working and what’s not. Report of the 5th Asia Forest
Network Meeting, Surajkund, India, 2–6 December 1996. Research
Network Report No 9, Asia Forest Network,
Rambo, A.T., Reed, R.R., Le, T.C., DiGregorio, M.R., 1995. The
Challenges of highland Development in Vietnam. CRES, Hanoi
University, University of California at Berkeley, East-West Center,
Honolulu, Hawaii.
Roder, W., 2001. Slash-and-Burn Rice Systems in the Hills of
Northern Lao PDR: Description, Challenges, and Opportunities.
International Rice Research Institute, Los Banos, Philippines.
Sadoulet, D., Castella, J.C., Vu H.N., Dang D.Q., 2002. A short
history of land use changes and farming system differentiation in
Xuat Hoa Commune, Bac Kan Province, Vietnam. In: Castella,
J.C., Dang D.Q. (Eds.), Doi Moi in the Mountains. Land Use
Changes and Farmers’ Livelihood Strategies in Bac Kan Province,

Vietnam. The Agricultural Publishing House, Hanoi, Vietnam,
pp. 21–46.
Sikor, T., 1999. The political economy of decollectivization: A study of
differentiation in and among Black Thai Villages of Northern
Vietnam. Ph.D. Dissertation, University of California, Berkeley,
USA.
Sikor, T., 2001. The allocation of forestry land in Vietnam: did it cause
the expansion of forests in the northwest? Forest Policy and
Economics 2, 1–11.
Tachibana, T., Nguyen, T., Otsuka, K., 2001. Agricultural intensifica-
tion versus extensification: a case study of deforestation in the
northern-hill region of Vietnam. Journal of Environmental
Economics and Management 41, 44–69.
Vo, Q., 1998. Ge
´
ne
´
ralite
´
s sur les proble
`
mes de l’environnement au
Vietnam. Etudes Vietnamiennes 3, 7–34.
Vo, Q., Le, T.C., 1994. Conservation of forest resources and the
greater biodiversity of Vietnam. Asian Journal of Environmental
Management 2, 55–59.
Vo, T.X., Nguyen, V.S., Nguyen, H.H., 1999. From farming systems
research to a community-based natural resource management
research agenda: current activities of the Vietnam Farming Systems
Network in northern Vietnam. In: Nguyen N.K., et al. (Eds.),

Towards an Ecoregional Approach for Natural Resource Manage-
ment in the Red River Basin of Vietnam. The Agricultural
Publishing House, Hanoi, Vietnam, pp. 187–195.
ARTICLE IN P RESS
J C. Castella et al. / Land Use Policy 23 (2006) 147–160160