This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
 Author's personal copy

Habitat International 41 (2014) 165e175

Contents lists available at ScienceDirect

Habitat International
journal homepage: www.elsevier.com/locate/habitatint

Short-term risk experience of involuntary resettled households in the

Philippines and Indonesia
Melissa Quetulio-Navarra a, *, Anke Niehof a, Hilje Van der Horst b, Wander van der Vaart c
a
Sociology of Consumption and Households Group, Wageningen University, Hollandseweg 1, 6706 KN Wageningen, The Netherlands
b
Wageningen University, The Netherlands
c
Social Research Methodology, University of Humanistic Studies, Kromme Nieuwegracht 29, 3512 HD Utrecht, The Netherlands

a b s t r a c t
Keywords: Involuntary resettlement often impoverishes the displaced households. Cernea argued that impover-
Involuntary resettlement ishment can be avoided with his Involuntary Risks and Reconstruction Model (IRR). The IRR Model has
IRR model
been widely utilized in resettlement studies and identies nine interlinked potential risks inherent to
Resettlement risks
Resettlement in the Philippines
displacement. Nonetheless, in assessing the risks as well as the effectiveness of interventions, most
Resettlement in Indonesia attention is directed at the institutional context. This paper argues that factors that mitigate risks of
impoverishment are much more wide ranging. It covers the period of a year before and a year after
resettlement. It investigates the manifestation of eight risks brought about by involuntary resettlement
episodes in the Philippines and Indonesia as well as the factors that increased or decreased them. The
ndings in the study of two different resettlement contexts show the multi-dimensionality of resettle-
ment risk causes. The study furthermore shows that institutional context alone is not enough to explain
the risk outcome. Culture, physical location, individual and household characteristics should be factored
in during the examination of impoverishment risks.
2013 Elsevier Ltd. All rights reserved.

Introduction infrastructure projects such as construction of railroads, cleaning

Resettlement studies are in agreement that the displacement of
impoverished families pushes them into worsening poverty situa-
tions (Cernea, 1985; Cernea & McDowell, 2000). Yet, involuntary
resettlement still takes place rampantly. A decade-long review of
the World Banks development projects entailing population
movement revealed that 90e100 million individuals were reported
to have entered the cycle of involuntary resettlement (MacDowell,
1996). Estimates reveal that approximately 10 million people per
year enter the cycle of involuntary displacement and relocation due
to dam- and transportation-related development programs alone
(Cernea & McDowell, 2000).
In the latest Indonesian Centre on Housing Rights and Eviction
Report, more than 12,000 people were reportedly evicted in July
and August 2008 to give way to the green space land reclamation
project (COHRE, 2008). In the Philippines, 59,462 households
relocated in the period 2001e2006 (HUDCC, 2008) due to various
* Corresponding author. Tel.: 31(0)619 15 98 63.
E-mail addresses: melissa.navarra@wur.nl, melissaqn@gmail.com (M. Quetulio-
Navarra), Anke.Niehof@wur.nl (A. Niehof), hilje.vanderhorst@wur.nl (H. Van der
Horst), w.vandervaart@uvh.nl (W. van der Vaart).
up of waterways, road widening, etc. Though more recent data are
lacking, there is no evidence that the pace of displacement is
slowing down.
Despite the evidence of negative effects of displacement on
households livelihoods, the governments of the Philippines and
Indonesia view involuntary resettlement as a development op-
portunity for both the poor resettlers and the public. It is believed
to stimulate regional development, economic development,
employment opportunities, and poverty alleviation, among others
(Arndt & Sundrum, 1977; NEDA, 2011; WB, 2012).
The Impoverishment, Risks and Reconstruction (IRR) model
(Cernea, 2000), which identies nine interlinked potential risks
inherent to displacement, has been widely utilized in resettlement
studies and projects. It is an analytical guide in mapping out the
impoverishment risks incurred by involuntary displacement (Bang
& Fewa, 2012; Cernea & Schmidt-Soltau, 2006; Dhakal, Nelson, &
Smith, 2011; Muggah, 2000; Price, 2009) as well as a basis in
generating an anti-poverty approach (Hong, Singh, & Ramic, 2009;
Wilmsen, Webber, & Yuefang, 2011b; Wilmsen, Webber, & Yuefang,
2011a).
In assessing the risks as well as the effectiveness of in-
terventions, most attention is directed at the actions and frame-
works provided by institutional actors. While institutional actors

0197-3975/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.habitatint.2013.07.013
 Author's personal copy
166 M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175
are important, the factors that mitigate risks of impoverishment are
much more wide ranging. In addition to the institutional context,
also cultural factors can be crucial. Furthermore, the IRR model
does not adequately explain the differences across households.
Household characteristics may cause great diversity in poverty
outcomes, which are not explained when looking at the level of
communities. In previous studies, differentiation of poverty risks in
households was made by extrapolating from already disadvanta-
geous poverty prole of the households. In other words, poverty
before resettlement led to a higher risk of poverty after resettle-
ment. In this article we evidence that such reasoning does not al-
ways hold true and that other household characteristics need to be
taken into account.
Through a comparative study of risks brought about by invol-
untary resettlement in the Philippines and Indonesia, at both
household and community level this study sheds light on a wider
range of factors that mitigate poverty at the level of communities
and at the level of households during the rst year of resettlement.
If we understand the wider range of factors that mitigate poverty
risks, as well as the underlying dynamics, we are in a better position
to gear policies towards preventing impoverishment after the
resettlement. Furthermore, the analytical strength of IRR model can
be greatly improved with such knowledge.
The present study uses the IRR lens to investigate the nature
and magnitude of risks experienced by poor households a year
after the displacement and resettlement episodes in two different
cases, one in the Philippines and one in Indonesia, and tries to
identify the factors (individual and household related character-
istics and institutional interventions) that are instrumental in
alleviating the resettlement impact and raise the resettlement
risk level in each country. Moreover, the research looks into the
factors that explain the differences in the risk experience between
the two sites.
Approaching involuntary displacement and resettlement
The most widely-adopted approach in resettling involuntary
displaced populations is Cerneas (2000) Impoverishment, Risks
and Reconstruction (IRR) Model. The model highlights the intrinsic
risks for impoverishment through displacement as well as the ways
to alleviate the risks. The IRR model was conceptualized and
developed in 1996, in a series of studies. Based on empirical nd-
ings, Cernea (2000) has identied the following nine interlinked
potential risks that are inherent to displacement:
1. Landlessness
2. Joblessness
3. Homelessness
4. Marginalization
5. Food insecurity
6. Increased morbidity and mortality
7. Loss of access to common property
8. Social disarticulation
9. Educational loss
Although these risks are inherent in a resettlement episode, the
intensity of their manifestation can vary among individuals and
households, and can be site-specic (Cernea & McDowell, 2000).
Forced evictions affect invariably the poorest and most disadvan-
taged groups in any displacement context (Leckie, 1994; Stanley,
2000), while resettled individuals who have social, educational,
and economic advantages can cope with and recover from the
debilitating effects of the displacement faster than others without
such benets (Parasuraman & Cernea, 1999).
Parasuraman and Cernea (1999) stressed that the outcome of
the resettlement is signicantly inuenced by the institutions
approach to the displacement and resettlement activity through
their resettlement policies and programs. In a dam-related reset-
tlement in China, the inputs from the government resulted into
maintaining or raising the income of the Chinese resettlers
(McDonald, Webber, & Duan, 2008). The opposite transpired in a
development-induced resettlement case in Indonesia (Nakayama,
Yoshida, & Gunawan, 1999), in Xiaolangdi, China (Webber &
McDonald, 2004), and in the Philippines (Quetulio-Navarra, 2007)
where the inadequacy in the interventions (income restoration
programs, resettlement policies, strategies, etc.) of the project im-
plementers led to failure.
Manifestation of the risks in the study
Although Cernea (2000) cited nine risks in his model, in this
paper only eight risks will be dealt with. We did not investigate the
risk of loss of access to common property and the risk of
increased morbidity and mortality is reduced to morbidity risk
only. The manifestation of these eight risks (nature and extent) in
the Philippines and Indonesia was examined by rst establishing
indicators for the eight risks followed by designing a risk index
based on these identied indicators. Data on a year before the
resettlement and a year after the resettlement were compared and
a score system was applied: 1 if there is a positive change (less
risk), 0 if there is no change, and 1 if the change is negative (more
risk). Eventually all the scores (for each risk) were summed up to
get the total risk score. Each of these eight risks was measured
using the established indicators below:
Landlessness. Landlessness was analysed by looking at the land-
ownership situation of the households before and after the
resettlement.
Homelessness. The homelessness risk was looked into using
housing-related data (housing ownership, house materials, oor
size, number of bedrooms) before and after the resettlement.
Joblessness. Based on the preceding qualitative investigation, the
magnitude of joblessness that affected the household was deter-
mined based on the data regarding the employment status of only
the household heads before and after the relocation.
Marginalization. The characteristics and extent of marginalization
were determined by utilizing data on household income, number of
factors that divide the community, number of social services the
households did not have or had limited access to, reasons for de-
nied social services, and whether the resettlement site was
perceived as peaceful and harmonious.
Food insecurity. Level of food insecurity was looked into by
comparing the before and after resettlement data on the percent-
age of low household income spent on food.
Increased morbidity. Data on the number of adults and kids who got
ill before and after the resettlement were compared along with the
data on the number of basic services available in the previous and
present communities.
Social disarticulation. Social disarticulation in the two communities
was measured by examining reports on the number of support ties,
memberships in organizations, and relationship of the community
with the local government, central government, NGOs and inter-
national organizations, and availability of public places before and
after the involuntary displacement.
 Author's personal copy

M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175 167

Educational loss. The gravity of loss of education in the households
after the relocation was based on the reports on the number of kids
who stopped going to school after the resettlement.
Methodological design
In the Philippines, a household survey was undertaken from
April to June 2011 in Kasiglahan Village 1 (KV1). KV1 is a
government-managed urban resettlement community situated
in Rizal Province, municipality of Rodriguez (see Fig. 1). It has
a total land area of 85.70 ha with 9915 housing structures of
32 m2 each. The resettlement community was built for poor
households that were evicted due to development projects,
natural and man-made disasters. Resettlement of families started
in 1999.
In Indonesia, a survey was conducted from April to June 2012 in
a rural community in Bantarpanjang Translok (BT). Bantarpanjang
Translok is also a government-managed resettlement community
situated in the district of Cilacap, in Central Java (see Fig. 2). Ban-
tarpanjang Translok has a total land area of 3.1 ha with 97 housing
structures of around 45 m2 each. The resettlement community
was built for poor households that were displaced by landslides
in nearby communities. It has been accommodating households
since 2001.

Fig. 1. Philippine resettlement site.

 Author's personal copy

168 M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175

Fig. 2. Indonesia resettlement site.

Sampling in the Philippines and Indonesia group interviews, participant and non-participant observations, in-
depth interviews and focus group discussions.
The 150 respondents in the Philippines were chosen through The calendar tool is a graphical grid with domains that served as
random sampling from a sampling frame of 6144 households who a visual aid as well as a data entry tool for both the interviewer and
are either original house and lot owners or rights buyers (bought the respondent. The calendar collected histories (for this paper,
the house and lot rights from the original owner). The Project Ofce only the data on a year before and a year after the resettlement)
did not have the exact numbers of these two types of residents, but pertaining to the following six major life domains of the
qualitative data yield that around 30e40% of the 6144 households respondent:
are rights buyers. The 150 respondents are all original owners. In
Indonesia, all 76 legitimate beneciaries of the resettlement project 1. Respondents prole
in the Indonesia community were interviewed. 2. Childrens information
Similar data gathering methods (quantitative and qualitative) 3. Household-related information
were applied in both sites. A household survey was conducted that 4. Physical features of the community
included a household composition sheet and a tailored-calendar 5. Respondents social engagement
tool. Qualitative methods included key informant interviews, 6. Respondents perception on the community
 Author's personal copy
M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175
Analysis of quantitative and qualitative data
The quantitative data were entered into Excel and were ana-
lysed using SPSS version 19. The data on a year before and a year
after resettlement were utilized and risk indexing was done for
each potential risk. Multiple linear regression analysis was con-
ducted on identied dependent and independent variables (see
Table A.1). The dependent variable is the total risk score of a
household and the independent variables are grouped into two-
those that reect the respondents socio-economic and de-
mographic prole as well as the characteristics of his or her
household, and the institutional context variables. The institu-
tional context variables pertain to the factors that represent the
interventions and inuence of the government both local and
central government level to the resettlement community. The
qualitative data were recorded, transcribed and content analysis
was applied. They were used to validate and complement the
survey data.
Field results
Involuntary displacees and the resettlement prole
In the Philippines, 68% of the respondents were female and 32%
male (Table B.1). More than half of the respondents were within the
age bracket of 25e45, the ages ranging between 20 and 85. Only
27% of the respondents reached college or studied in technical
school after high school, 47% nished or reached high school, and
25% only studied until elementary level. The average household
size was 5.58 and the average yearly household income is Php
88,103.00 (US$2065.72). Thirty-three percent were housewives or
husbands staying at home, 22% were labourers, 16% had a business
in the community, while another 16% were said to be unemployed.
The resettlers in KV1 were victims of development projects, natural
disasters (like ooding along Pasig river), man-made disaster
(garbage slide in Payatas) and wide-scale re. The resettlement
started in 1999.
The resettlement program was a component of the Pasig River
Environmental Management and Rehabilitation Sector Develop-
ment Program funded by the Asian Development Bank amounting
to US$100 million USD (PRRC, 2006). The program aimed to reha-
bilitate the Pasig River and part of it was relocating the poor fam-
ilies living along the river. The National Housing Authority, a
government agency mandated to implement socialized housing
program in the Philippines, undertook the resettlement program.
The resettlement package includes a house and lot in a resettle-
ment community that has basic services and primary public facil-
ities. The house and lot is payable in 20 years at Php 250 (US$5.9)
per month. The package was supposed to be prepared in accor-
dance with the Urban Development Housing Act, which provides a
systematic program for land use planning towards the allocation of
lands for social housing for the underprivileged and homeless city
dwellers. It covers a wide range of provisions the resettlers.
In Indonesia, the majority of the respondents in Indonesia was
male (92.1%) and 7.1% female (Table C.1). Most respondents (64.6%)
belonged to the age bracket of 41e60 years old. High school was the
highest education level reported by the respondents, while 71.1% of
the respondents had only elementary-level education. More than
half of the respondents were either doing elementary jobs (31.6%)
or were farmers (23.7%). The average household size was 3.96
and the average household income was IDR 10,975,006.58
(US$1141.45). All were Muslims and were ethnic Sundanese. The
community is divided into blocks called RT (rukun tetangga).
Although the landslide took place in 2000, it took a year before the
169
housing structures were in place. All households were victims of
landslides. Their occupancy in Translok (the resettlement site) is in
a lease-like agreement.
The agencies primarily involved in the resettlement project
were the Cilacap Provincial Government and the Department of
Transmigration of Cilacap in particular. At national level the Min-
istry of Transmigration is a government agency tasked to undertake
transmigration programs that aim at decongesting the densely
populated area of Central Java. Mostly, but not in this case, families
are relocated to other islands. Documents pertaining to the Ban-
tarpanjang Translok project were difcult to get. According to the
ofcials in Cilacap they were either lost or misplaced during the
times that they transferred ofces. An in-depth interview with a
Transmigration ofcer in Cilacap who was directly involved in the
supervision of the project, revealed that there was actually no
budget earmarked for the Translok families affected by the 2000
landslide. However, given the urgency of the case, the Department
reallocated some of their funds for the regular transmigration ac-
tivities to the development of the Bantarpanjang Translok com-
munity. Subsequently, negotiations started about the land where
the new community would be built. The land was originally under
the administration of Perhutani, an Indonesian state-owned en-
terprises that has the duty and authority to administer the plan-
ning, management, exploitation and protection of forests. In order
to acquire the land for the resettlement project, the Cilacap Pro-
vincial Government and Perhutani agreed to enter into a land
switching contract in which the Cilacap Provincial Government
compensates for the Perhutani land with a property twice the size
of the target land.
The land and the houses are still owned by the provincial gov-
ernment and there is no option for the resettlers to acquire it. Up
until now, the Indonesian government still has no national policy
for resettlement. The Transmigration Ministry follows the Keppres
55/1993 in the transmigration of families but the law only covers
the aspect of land acquisition, not the provision of basic services
and public facilities (Zaman, 2002). Hence, resettlement projects in
Indonesia have been handled on an ad hoc basis.
Involuntary resettlement risk index
Table D.1 presents the resettlement risk indexes in the Philippines
and Indonesia.
Land ownership
As shown on the mean score (1.00) for land ownership risk in
Table D.1, KV1 resettlers did not experience any risk in losing land,
rather the opposite happened. Before the involuntary resettle-
ment, the respondents did not own the land where they were
occupying (legally or illegally), which qualied them for the so-
cialized housing program in the Philippines. After the relocation,
each household was given a house and lot by virtue of an Award
Certicate that is payable in 20 years at Php 250 (US$5.9) per
month. After they have paid the full amount, they will be given a
land title.
The opposite was the case for the Indonesian respondents.
Having owned their land, at least a small plot for their house and for
farming, after the landslide the Translok relocatees became land-
less (mean risk score of 1.00). The resettlement project funds could
not afford to replace their lost lands and they cannot purchase the
land where they are presently residing. Series of negotiations be-
tween the Cilacap Transmigration Department and the Provincial
Government did not result in a positive outcome. One leader in
Translok said:
 Author's personal copy
170 M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175
We have been working out with the RW leader and the village
head for the possibility of acquiring the land where we are staying
right now. But nothing is happening. This is really worse than when
we were on the mountains. Before we owned the land where our
houses were built and also had land for farming. But the landslide
swallowed everything. We are scared now because anytime the
Bupati (head of the district) wants to take back the land where
were staying now, we have to leave.
(Interview with a male community leader, Translok Indonesia,
June 2012)
Housing situation
The Philippine resettled households are also in a better-off sit-
uation when it comes to housing, as evidenced by 2.03 risk score
mean. Before the displacement, the households were renters, living
with parents or relatives, or living in shanties under the bridge, on
sidewalks, or along the railroad tracks. All of them, did not own a
house but after their transfer to KV1 they were awarded with a
concrete house structure with a toilet and gained security of tenure.
When asked whether they were happy with their housing situa-
tion, one respondent said:
Now I dont worry about strong rain or even storm because we
now live in a house that has stone walls and strong roof. Previ-
ously, I was scared every time there was a storm or even a strong
rain with strong wind because our roof and walls were made of
very light materials and they could be just blown away at any
moment.
(Survey interview with a female respondent in KV1 Plains,
Philippines, June 2011)
In contrast, the Indonesian resettlers are in a worse housing
situation (risk score mean of 1.75). Before, they owned their houses,
but when they transferred to Translok they were only given a
permission certicate from the village head which gives them the
right to use the housing structure but not own it. If before their
houses were made of concrete, after the landslide they transferred
to a wooden house with no divisions for bedroom and kitchen and
no provision for toilet inside the house. Eight public toilets were
built for the resettlers.
Employment situation
Employment risk score in the Philippines is very low (0.01),
with only 5.4% of the respondents becoming jobless after the
transfer. While most were able to keep their jobs or source of
livelihood by re-establishing their small business like a small store
or junk shop in KV1, some could not keep their jobs. Those who
experienced a worse-off condition in terms of employment
attributed this to the distance (50e70 km) of KV1 from their
previous job, unrecoverable loss brought about by the garbage
slide in Payatas, and the unavailability of employment opportu-
nities in the new site.
The Indonesian resettlers also scored very low on this risk
(0.04), with 5.3% of the household heads losing their jobs. Since
most of the resettlers are farm labourers it was not hard for them
to nd work in the rice elds after they transferred. Some even
revealed that the resettlement got them closer to the rice elds; if
previously they would walk for 2e3 h to reach their rice elds,
now they only had to walk for 1 h from Translok. Some of them
were also able to forge a special arrangement with Perhutani that
allows them to plant some crops for own household consumption
or for selling on Perhutani land in the forest at a very minimal
payment.
Marginalization
In aggregate, the KV1 relocatees in the Philippines are a little
better off (0.03 mean risk score) after the transfer in terms of their
level of economic, social, and psychological marginalization. The
Indonesian Translok relocatees in Indonesia also scored positively
in this respect (mean risk score of 0.28). However, in terms of
access to social services during their rst year in Translok, almost
everybody expressed the same sentiments regarding their suffering
from the absence or very poor basic services inside the resettle-
ment site and the transition to public toilet system, one household
head uttered:
On the mountains we did not have problems with the toilet. We
just needed to nd our own toilet space along the river and we
didnt need to keep it clean. When we came here we had a new
system of using the toilet- we needed to line up, wait for our turn to
use the toilet, and keep it clean which is difcult because you share
the toilets with everybody here.
(Survey interview with a male household head in RT 2, Translok,
Indonesia, May 2011)
Food security
The food security risk index for the Philippine case registers a
very low score of 0.15 (mean), with 24.7% of the respondents
adversely affected and 9.3% being better off. The risk score reects
the increase in the proportion of household income spent on food
for the household, which in KV1 may be attributed to higher price
of food items due to an increase in transactions costs coupled with
the reported growth in number of children in the household after
the resettlement.
The food security situation of the Indonesian resettlers
improved a little after the relocation as evidenced by the food se-
curity risk score of 0.09 (mean), with 11.8% of the households at
risk and 9.3% of households doing better. While the resettlers lost
their gardens to the landslide, after transferring to Translok some
were able to arrange (with Perhutani or friends and relatives) a new
garden lot where they could plant their crops, such as cassava,
peanuts, and eggplants, again. Others who also had an increase in
household income found food in Translok more accessible because
of the stores around and the peddlers who come to the place to sell
food.
Morbidity
Based on the morbidity index, the Philippine households are
slightly at risk (0.53 risk score). However, this score is largely due to
the reduction in the number of basic services provided in the
resettlement site and does not reect a higher incidence of illness
in the household. Although the houses were already constructed
when the families came to KV1, the basic services were provided
only gradually, depending on the release of the budget from the
central government. In the Indonesian case resettlement did not
increase morbidity at all (mean risk score 0.33).
Social articulation
The KV1 residents did not suffer from social disarticulation risk
after the move as shown by a 0.39 mean score. Similarly, the
Translok residents seemed to have improved in terms of social
 Author's personal copy
M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175
articulation after the resettlement (mean score 1.25). A respon-
dent remarked during an informal conversations:
.one of the good things I like in Translok is that I can now borrow
money from a money lender to buy some goods that I like or buy
some stuff from a seller and pay them later with or without
additional interest. We did not have this kind of arrangement
before on the mountains.
(Participant observation with women farm labourers in Translok,
Indonesia, June 2012)
Education
A low mean risk score (0.01) was found in the Philippine
resettlement case for educational loss. In-depth interviews and
desk research revealed that when the families relocated to the site,
there were no schools. They had to send their children to schools
located in nearby communities. However, they found this very
inconvenient and costly (transport). They lodged strong complaints
at the Project Management Ofce (National Housing Authority).
Several months later (in 2000) some unoccupied housing units
were converted into classrooms. In the Indonesian case the mean
score is also low (0.07). The small number of children who stopped
going to school may be due to the fact that the day care centre was
only built much later.
Total post resettlement risks
As can be gleaned from Table D.1, the Philippine households are
in a better-off condition (mean risk score of 2.75) compared with
the Indonesian resettlers (mean risk score of 0.90). This is due to
the fact that KV1 households gained house and lot assets after the
displacement, while it is the opposite in the Indonesia case. In
order to ascertain which among these dimensions yield the results
shown in Table D.1, the total risk score was regressed against a
number of variables that are grouped into three models presented
in Table A.1.
Regression results are shown in Table D.2 for the Indonesia case
and Table E.1 for the Translok resettlement. Table D.2 shows that
in Model 1 only the variables husbands employment status and
the number of membership in organizations can account for the
level of involuntary resettlement risks experienced by the
households in KV1 with an R-square of 12.8 percent. Model 2
(R2 28%) which regressed the institutional context variables
with the total risk scores yield four signicant variables e the
number of public places in KV1, community relationships with the
local government, NGOs, and international organizations. How-
ever, the combination of these two dimensions (model 1 and 2
variables) explains best (at 38.9%) the variance in the post-
resettlement risks in KV1. Three of the variables that reect
socio-demographic and household characteristics are signicant.
Education of the household head, memberships in organizations
after the resettlement, and the husbands employment status, are
the strongest predictor for risks.
Surprisingly, in this resettlement case education is negatively
related to risk, which challenges the usual claim of resettlement
studies that education alleviates the harmful effects of involun-
tary resettlement. The result tells us that the higher the education
level of the household head has the higher risks the household will
face during resettlement. Compared to the lower educated, more
educated household heads lost more after resettlement. Previously
they had better jobs, better salaries and, therefore, could afford to
rent a better house and had access to basic services. Some were able
to keep their jobs in the city for a while, but the rent or the huge
171
transportation expenses took toll on their food and health budget.
One respondent said:
Life was so hard here on our rst year. We lost our source of in-
come and my family had to rely on food rations and relief goods
such as rice, noodle soup, canned sardines from the government,
church, and NGOs. We would line up all the time for the food dis-
tribution and stock up food rations in our house which we will eat
little by little.
(In-depth interview with a mother in KV1 Plains, Philippines, May
2011)
An unemployed husband and a decrease of membership in
organizations (voluntary and involuntary) increase the vulnera-
bility level of a household after relocation. While the husbands
unemployment is self-explanatory, membership in organizations
is about access to resources which may cushion an individual
against risks.
Six institutional context variables turn out to inuence the
households susceptibility to the risks inherent to involuntary
resettlement. These are: living in Plains, number of public places
in KV1, the communitys relationship with the local government
and international organization (all reducing risk) and the com-
munitys relationship with NGOs and church (both, surprisingly
enhancing risk). Kasiglahan Village 1 is physically divided into two
parts, the Plains and the Suburban. The Plains was the part
developed rst in the resettlement project, where the rst groups
of resettlers transferred. Although the Plains is more developed
now than the Suburban, the situation was dismal during the early
years, physically as well as socially. A community leader from
Plains shared:
They (referring to NHA) resettled us all here to die. Housing was
terrible, basic services were very minimal, and no jobs available. I
dont think they are really concerned with our welfare.
(In depth-interview with a KV1 community male leader, KV1
Plains, May 2011)
The outcome on the number of public places (not the basic
services) as inuencing the level of risk a household might expe-
rience underscores the social and economic value of public places
to the KV1 relocatees during their rst year in the community.
Public places like sidewalks, and markets functioned as venues for
meeting new friends and acquaintances and as physical spaces for
economic activities (e.g. in the market place).
The risk level outcome in KV1 was likewise a result of the nature
of the community relationship with the local, NGOs, international
organizations and church. However, while a better relationship
between the community and the local government as well as with
the international organizations would mitigate the risks brought
about by the resettlement, it is entirely the opposite in the case of
relationships with the NGOs and church. A project manager
admitted that basic services and basic facilities in KV1 were built
late due to budgetary constraints. Households complained non-
stop about the nature of service delivery in KV1. The NGOs inter-
vened, but their lobbying efforts only exacerbated the social issues
in KV1, eventually negatively inuencing the cooperation of the
households. Similarly, when the community or the households
forge a good relationship with the church, most of the time this
results in a patroneclient like relationship that limits connecting
with other people who do not share the same religion or faith,
thereby preventing them from expanding their social networks
that are a potential well of resources.
Table E.1 shows the regression results for the risk score of the
Indonesian households. Three variables turned out to be signicant
 Author's personal copy
172 M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175
in Model 1 with an R square of 18.60 percent: the education of the
household head (by far with the strongest impact), the husbands
employment status, and membership in organizations. In Model 2
(R2 19.50%), the variables number of public places and the rela-
tionship of the community with the local government turned out as
predictors of risk level in Translok.
Similar with the KV1 regression results, Model 3 provides the
best explanation for the resettlement risk score in Translok
(R2 35.9%). Signicant variables are education, husbands
employment status, and relationship with the local government.
Like in the KV1 case, the higher the level of education of the
household head the higher the resettlement risks score will be and
the same explanation applies to this outcome. A good relationship
between the Translok community and the local government de-
creases the risks since social programs that can mitigate impover-
ishment in the locality usually emanate from the level of the local
government.
Discussion and conclusion
This study about two different resettlement contexts in South-
east Asia presents robust evidence on the multi-dimensionality of
resettlement risk causes. The study shows that, while the institu-
tional context is important, also culture, physical location, and in-
dividual and household characteristics add to the explanation of
risks experience during rst year in the resettlement.
The resettlement program policy of the country concerned
proved to greatly inuence the manifestation of eight risk factors
identied by Cernea (2000). The resettlement program showed
large differences between the two sites. The Philippine KV1 reset-
tlers gained assets in the form of housing and land because there is
a national policy on resettlement and socialized housing in the
Philippines that protects right to housing and aims for humane
relocation. The presence of the national policy made the KV1 res-
idents aware of their rights, as apparent in the case of their lobbying
for building a school in the site. On the other hand, the Indonesian
Translok relocatees lost their houses and land to landslides and
were never compensated for their loss due to the absence of a
country policy on resettling natural disaster victims. Hence, they
had no grounds to demand basic facilities and just compensations.
While differences in level of educational level and age be-
tween the Philippine and Indonesian households could explain
these dissimilar situations, the multivariate analyses showed that
in both sites there is no signicant impact of the respondents
age, gender, income and household size on the total risk score.
Both the Philippine and Indonesian households experienced four
risks with two risks emerging in both cases (employment and
education). Nonetheless, it is apparent that the risks manifested
differently in the two cases. The role of the physical environment
both in the previous and the resettlement residence, along with
the households cultural background, inuenced the emergence
of risks related to food security and morbidity. The Indonesian
Translok relocatees were able to prevent food insecurity after the
resettlement because they continued their practice of planting
crops in the mountain area through new arrangements with
Perhutani, a relative, or a friend. It also helped that most men
quickly found jobs and continued to work as farm labourers in
the nearby rice elds after the transfer. In the Philippine KV1
case, however, households were unable to avoid food insecurity
during their rst year in the site. The KV1 resettlers were not
accustomed to growing their own crops and had no access to
land for farming. They relied on their income from non-
agricultural activities to secure high-priced food items from
peddlers coming to the site.
The multivariate analyses showed that in both the Philippine
and Indonesian case the individual and household variables add
substantively to the institutional ones in explaining risk. Moreover,
in spite of the large differences, in both cases the same socio-
demographic variables are inuential (employment, membership
in organizations and the unexpected negative impact of education)
and their weight as relative to the institutional variables is about
similar (12e18% versus 39e36% R-square). Regarding the institu-
tional factors, the outcomes did show important differences be-
tween the Philippine and Indonesian case. Most consist of the
absence of signicant relationships in the Indonesian case. The
ndings also show that irrespective of the level of signicance, the
sign of all effects is similar (positive or negative) in both cases,
except for the inuence of NGOs and church which, unexpectedly,
enhanced risk in the Philippines.
The results of the research convey a new perspective on the
concept of vulnerability in a forced resettlement context. Unlike the
usual nding that education shields a resettler from risk, this study
yielded a contradictory result. This may be due to the fact that a
more educated household head tends to lose more during
displacement and that resettlement in an impoverished commu-
nity actually misplaces him or her in an employment pool, thereby
eventually aggravating the risk level in other areas.
In the Philippines, a good relationship with the church and
NGOs during the rst year in the resettlement site generated
negative effects. We can infer from this that in an urban reset-
tlement context like the Philippines, wherein you have a national
policy on resettlement and community leaders who are aware of
their rights, the intervention of the NGOs and perhaps the church
may result in delays in service delivery and trigger disagree-
ments among the project stakeholders. What should be nurtured
rst is the community relationship with the host local govern-
ment (negative regression weight) who is in the best position to
deliver the social services and interventions needed by the
households and the project managers. This study demonstrates
that features of the resettlement program, the institutional
context, and individuals and households together cause and
mitigate risk during the rst year of resettlement. Taking these
factors together proved to add substantially to the strength of the
IRR model, even when applied to two quite divergent resettle-
ment populations.
Acknowledgement
The authors are grateful to the Netherlands Fellowship Pro-
gramme that funded the eldwork in the Philippines and the
Neys Van Hoogstraten Foundation that funded the eldwork in
Indonesia. Also, to the households and leaders in Kasiglahan
Village Community and the Bantarpanjang Translok for their
participation in the research, Engr. Abeth Matipo of the National
Housing Authority, Mr. Candra Musi and Mr. Narso of Perum
Perhutani for the logistical assistance, Dr. Ana Marie Karaos and
Ms. Vangie Serrano of Action Group for the assistance at the
community level in the Philippines, Dr. Sri Sunarti Purwaningsih
and Dr. Herry Yogoswara of Indonesian Institute for Sciences
(LIPI), Dr. Tyas Wulan and Dr. Suliyanto Jayakrama of Universitas
Jenderal Soedirman, Yayasan Kusuma Buana led by Dr. Firman
Lubis, Presidential Commission for the Urban Poor led by
Chairman Hernani Panganiban, Dennis Sorino and Dulce Joy
Sorino for database design and data processing, the Philippine
Research Team led by Elizabeth Avila, the Indonesian Research
Team led by Fanny Dwipoyanthi, Eden Frunt, Victor Allan C. Ila-
gan, Alicia S. Diaz, Martina Robles, Daniel Vincent, and in WUR
Dr. Hester Moerbeek and Feng Zhao for statistical analysis.
 Author's personal copy

M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175 173

Appendices Table B.1 (continued )

Variables Percentage Frequency

7e9 22.0 33
Table A.1
10 or more 5.3 8
Dependent and independent variables.
Household income quartile
Dependent variable Explanation <48,000 25.0 36
>48,000 and<90,000 25.0 35
Total risk score Sum of all scores on the eight
>90,000 and<130,656 25.0 36
different areas of potential risks
>130,656 25.0 35
Independent Variables
Present address
Demographic, socio-economic, & household characteristics
Plains 60.7 91
Age
Suburban 39.3 59
Gender Dummy variable, female 1 (Philippines)
Dummy variable, male 1 (Indonesia)
(after resettlement)
Husbands employment status Dummy variable, where employed 1
Income
(after resettlement)
Household size
Membership in organizations
Number of support ties Table C.1
Institutional context Descriptive statistics of respondents in Indonesia N 76.
(after resettlement)
Variables Percentage Frequency
Living in plains (KV1) Dummy variable, Plains 1
Living in RT1 and RT 3 (Translok) Gender
Number of basic services Male 92.1 70
Number of public places Female 7.9 6
Number of denied social services Age
Rate of community relationship Dummy variable, Very low 1; 20e30 4.0 3
with: Low 2; Moderate 3; High 4; 31e40 23.7 18
Very High 5 41e50 34.2 26
Central government 51e60 30.3 23
Local government 61 e more 7.9 6
NGOs Civil status
International organizations Single 2.6 2
Church/mosque Married 93.4 71
Widowed 4.0 3
Education
Never been to school 4.0 3
Elementary school or less 71.1 54
Junior high school 19.7 15
High school 5.3 4
Occupation
Table B.1 Farmer 23.7 18
Descriptive statistics of respondents in the Philippines N 150. Entrepreneur 13.2 10
Elementary occupation 31.6 24
Variables Percentage Frequency Govt/private employee 7.9 6
Housewife 5.3 4
Gender
Others 15.8 12
Male 32.0 48
Household size
Female 68.0 102
1e3 34.2 26
Age
4e6 63.2 48
25e35 0.2 30
7e9 2.6 2
36e45 0.3 47
Household income quartile
45e55 0.3 43
<4,560,000 25.0 19
56e65 0.1 21
>4,560,000 and<9,000,000 27.6 21
66 e more 0.1 9
>9,000,000 and<14,850,000 22.4 17
Civil status
>14,850,000 25.0 19
Single 5.3 8
Present address
Married 75.3 113
RT 1 26.3 20
Separated 3.3 5
RT 2 36.8 28
Widowed 6.7 10
RT 3 32.9 25
Co-habiting 9.3 14
RT 5 4.0 3
Education
Elementary school or less 0.3 37
High school 0.5 71
More than high school 0.3 41
Occupation
Entrepreneur 16.0 24
Labourer 22.0 33
Govt/private employee 14.0 21
Housewife 32.7 49
Retired 2.0 3
Unemployed 16.0 24
Household size
1e3 18.7 28
4e6 54.0 81
 Author's personal copy

174 M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175

Table D.1
Risk index in the Philippines

Philippines N 150 Philippines Indonesia Philippines Indonesia Difference Philippines Indonesia Difference Philippines Indonesia Difference
Indonesia N 76

Potential risks areas Mean SD Mean SD Worse off% Worse off% Same% Same% Better off% Better off%

Land ownership 1.00 0.00 1.00 0.00 0 100.0 100.0 0 0.0 0.0 100 0.0 100.0
Housing situation 2.03 1.16 1.75 1.13 1.3 88.2 86.9 10.7 9.2 1.5 88 2.6 85.4
Employment 0.01 0.46 0.04 0.26 5.4 5.3 0.1 88 92.1 4.1 6 1.3 4.7
Marginalization 0.03 1.51 0.28 1.09 32.7 19.7 13.0 34.7 34.2 0.5 32.6 46.1 13.5
Food security 0.15 0.56 0.09 0.57 24.7 11.8 12.9 66 67.1 1.1 9.3 21.1 11.8
Morbidity 0.53 1.03 0.33 0.57 57.3 5.3 52.0 25.3 73.7 48.4 17.3 21.1 3.8
Social Articulation 0.39 2.09 1.25 1.29 34 5.3 28.7 18.7 26.3 7.6 47.2 68.4 21.2
Education 0.01 0.16 0.07 0.25 2 6.6 4.6 96.7 93.4 3.3 0.07 0.0 0.1
Total risks score 2.75 3.55 0.90 2.28 15.3 53.9 38.6 9.3 19.7 10.4 75.3 26.3 49.0

Table D.2
Effects of socio-demographic, household characteristics and institutional context on the involuntary resettlement risks in the Philippines (pairwise deletion of missing values)
N 150.

MODEL 1 MODEL 2 MODEL 3

B SE B SE B SE

Age 0.030 0.023 0.132 0.022

Female 0.120 0.532 0.116 0.480
Education 0.086 0.144 0.185** 0.139
Husbands employment status 0.207** 0.603 0.215*** 0.538
Income 0.030 0.557 0.089 0.503
Household size 0.087 0.128 0.040 0.116
Membership in organizations 0.215** 0.383 0.014* 0.031
Number of support ties 0.031 0.034 0.150 0.356
Living in Plains 0.043 0.471 0.237*** 0.519
Number of basic services 0.047 0.073 0.005 0.078
Number of public places 0.177** 0.092 0.179** 0.096
Number of denied social services 0.158 0.080 0.125 0.082
Rate of community relationship with central government 0.043 0.305 0.032 0.306
Rate of community relationship with local government 0.327*** 0.300 0.259** 0.316
Rate of community relationship with NGOs 0.317*** 0.249 0.255*** 0.256
Rate of community relationship with international organizations 0.282*** 0.267 0.349*** 0.282
Rate of community relationship with church 0.106 0.216 0.183** 0.230
R2 12.8% 28.0% 38.9%

Signicant at *p < 0.10,**p < 05 &, ***p < 0.01.

Table E.1
Effects of Socio-demographic, household characteristics and institutional context on the involuntary resettlement risks in Indonesia (pairwise deletion of missing values)
N 76

MODEL 1 MODEL 2 MODEL 3

B SE B SE B SE

Age 0.045 0.256 0.101 0.274

Male 0.069 1.094 0.016 1.201
Education 0.262** 0.432 0.247** 0.451
Husbands employment status 0.246** 1.361 0.250* 1.454
Income 0.059 0.276 0.003 0.347
Household size 0.008 0.200 0.060 0.216
Membership in organizations 0.208* 0.279 0.133 0.301
Number of support ties 0.107 0.045 0.162 0.051
Living in RT1 0.045 0.755 0.044 0.782
Living in RT3 0.086 0.697 0.010 0.715
Number of basic services 0.032 0.227 0.051 0.236
Number of public places 0.234* 0.267 0.166 0.281
Number of denied social services 0.096 0.146 0.167 0.168
Rate of community relationship with central government 0.134 0.484 0.116 0.524
Rate of community relationship with local government 0.349*** 0.295 0.324** 0.296
Rate of community relationship with NGOs 0.215 1.071 0.217 1.053
Rate of community relationship with international organizations 0.071 1.256 0.019 1.316
Rate of community relationship with mosque 0.021 0.394 0.012 0.391
R2 18.6% 19.5% 35.9%

Signicant at *p < 0.10,**p < 05 &, ***p < 0.01.

 Author's personal copy
M. Quetulio-Navarra et al. / Habitat International 41 (2014) 165e175
References
Arndt, H. W., & Sundrum, R. M. (1977). Transmigration: land settlement or regional
development? Bulletin of Indonesian Economic Studies, 13, 72e90.
Bang, H. N., & Fewa, R. (2012). Social risks and challenges in post-disaster resettle-
ment: the case of Lake Nyos, Cameroon. Journal of Risk Research, 15, 1141e1157.
Cernea, M., & McDowell, C. (2000). Risks and reconstruction: Experiences of resettlers
and refugees. World Bank.
Cernea, M. M., & Schmidt-Soltau, K. (2006). Poverty risks and national parks: policy
issues in conservation and resettlement. World Development, 34, 1808e1830.
Cernea, M. (1985). Land tenure and the social units sustaining alternative forestry
development strategies. World Bank.
Cernea, M. (2000). Risks, safeguards and reconstruction: a model for population
displacement and resettlement. Economic and Political Weekly (Vol. 35, 3659e3678.
COHRE, Indonesia, Center on Housing Rights and Eviction. http://www.cohre.org/
Indonesia Accessed 25.05.10.
Dhakal, N., Nelson, K., & Smith, J. L. D. (2011). Resident well-being in conservation
resettlement: the case of Padampur in the Royal Chitwan National Park, Nepal.
Society and Natural Resources, 24, 597e615 (519).
Hong, P. Y. P., Singh, S., & Ramic, J. (2009). Development-induced impoverishment
among involuntarily displaced populations. Journal of Comparative Social Wel-
fare, 25, 221e238.
HUDCC. Accomplishment report (2001e2006) housing and secure tenure. http://www.
hudcc.gov.ph/hudcc/www/SiteFiles/File/Housing%20Databases/Arroyo,%
202001-2006.pdf Accessed 25.05.10.
Leckie, S. (1994). Forced evictions. Environment and Urbanization, 6, 131e146.
MacDowell, C. (1996). Understanding impoverishment: The consequences of
development-induced displacement. Berghahn Books.
McDonald, B., Webber, M., & Duan, Y. F. (2008). Involuntary resettlement as an
opportunity for development: the case of urban resettlers of the Three Gorges
Project, China. Journal of Refugee Studies, 21, 82e102.
175
Muggah, H. (2000). Conict-induced displacement and involuntary resettlement in
Colombia: putting Cerneas IRLR model to the test. Disasters, 198e216.
Nakayama, M., Yoshida, T., & Gunawan, B. (1999). Compensation schemes for reset-
tlers in Indonesian dam construction projects. Water International, 24, 348e355.
NEDA. (2011). Philippine development plan 2011e2016. Pasig City: National Economic
Development Authority Accessed 30.03.13 http://www.neda.gov.ph/PDP/2011-
2016/FRONT%20MATTER.pdf.
Parasuraman, S., & Cernea, M. M. (1999). The development dilemma: Displacement in
India. Macmillan [etc.].
Price, S. (2009). Prologue: victims or partners? The social perspective in
development-induced displacement and resettlement. The Asia Pacic Journal of
Anthropology, 10, 266e282.
PRRC. (2006). Adb loan nos. 1745/1746-Phi: Pasig river environmental management
and rehabilitation sector development program: JanuaryeDecember 2006 report.
Quezon City: Pasig River Rehabilitation Commission.
Quetulio-Navarra, M. (2007). The Kamanava ood control resettlement project: As
perceived by primary stakeholders (Unpublished Masters Thesis). Japan: Grad-
uate School of International Cooperation Studies, Kobe University.
Stanley, J. (2000). Development-induced displacement and resettlement.
forcedmigration.org.
WB, & IEG. (2012). Indonesia transmigration program: A review of ve bank-supported
projects. Washington: World Bank.
Webber, M., & McDonald, B. (2004). Involuntary resettlement, production and in-
come: evidence from Xiaolangdi, PRC. World Development, 32, 673e690.
Wilmsen, B., Webber, M., & Yuefang, D. (2011a). Development for whom? Rural to urban
resettlement at the three Gorges Dam, China. Asian Studies Review, 35, 21e42 (22).
Wilmsen, B., Webber, M., & Yuefang, D. (2011b). Involuntary rural resettlement:
resources, strategies, and outcomes at the three Gorges Dam, China. The Journal
of Environment & Development, 20, 355e380.
Zaman, M. (2002). Resettlement and development in Indonesia. Journal of Contem-
porary Asia, 32, 255e266.