Ecological Economics 179 (2021) 106829

Contents lists available at ScienceDirect

Ecological Economics
journal homepage: www.elsevier.com/locate/ecolecon

Analysis

Impact of oil palm expansion on the provision of private and community T

goods in rural Indonesia
Vijesh V. Krishnaa, , Christoph Kubitzab,c
⁎

a
International Maize and Wheat Improvement Centre (CIMMYT), Hyderabad, India
b
German Institute for Global and Area Studies (GIGA), Hamburg, Germany
c
University of Goettingen, Goettingen, Germany

ARTICLE INFO ABSTRACT

Keywords: Curtailing deforestation and the negative environmental and social externalities of oil palm expansion in the
Land-use change tropics requires a clear understanding of the associated economic benefits. Through analyzing regency-level
Private goods panel data for 2005–2014 from rural Indonesia, we find that oil palm expansion has had a robust and positive
Community goods impact on the provision of private goods (i.e., amenities owned and benefited by individual households), al­
Impact estimation
though not on the provision of most community goods (i.e., amenities owned by the community or the gov­
'Potensi Desa' (PODES)
ernment and benefited by several households in the community). One possible reason for the weaker impact on
Inclusive development
the creation of community goods is the absence of differential government spending in the regencies cultivating
oil palm. Policies that aim to regulate oil palm expansion should take into consideration the positive effects that
the crop has on rural livelihoods through enabling access to private goods. On the other hand, the central and
regional governments should not assume that the development of rural infrastructure can be instigated solely by
the expanding plantation sector.

1. Introduction in roads for better market access). In the long run, the resultant eco­
nomic development could even decrease pressure on forest land
Trade relations between the major palm oil producers – Indonesia (Angelsen, 2010).
and Malaysia – and some of the strategic importers in the international As a necessary precursor to the effective regulation of the ecologi­
commodity markets – the United States of America and the European cally unsustainable land-use changes associated with oil palm expan­
Union – have taken a downward turn in the last decade. The European sion, various economic incentives associated with the crop should be
Union, for instance, is considering reducing imports of palm oil, citing a examined in depth. In this connection, we estimate the effects of oil
large body of evidence showing that oil palm expansion is threatening palm expansion on rural economic development in Indonesia, the cur­
biodiversity hotspots and contributing to climate change (Taheripour rent global leader of palm oil production. Does oil palm's large con­
et al., 2019; Yacob, 2019). On the other hand, the governments of In­ tribution to Indonesia's export tax revenues and inflow of foreign cur­
donesia and Malaysia emphasize the central role that oil palm plays in rency contribute to the economic development of the oil palm
economic development (Ickowitz et al., 2016; Krishna et al., 2017). cultivating regencies? Most empirical studies address this question by
However, to what extent oil palm acts as an effective instrument of considering the sphere of private goods and connecting the changes in
rural economic development, including the reduction of poverty and farm-household income or consumption to oil palm cultivation. By
the development of infrastructure, remains contested. Economic evi­ doing so, these studies overlook the possible welfare gains achieved
dence supports the argument that growth in agricultural productivity through the creation of private and community goods.
augments the rural economy and results in structural transformation as The economic effects of oil palm cultivation depend largely on the
income levels and local demand increase (Byerlee et al., 2009; Dethier regime in which it is taking place. While large companies produce most
and Effenberger, 2012). Moreover, large-scale agricultural plantations of the Indonesian palm oil, smallholders are becoming increasingly
might increase tax revenue and invest in the development of local in­ involved in oil palm cultivation. Data from the Directorate of Estate
frastructure, not only to compensate for the local community's loss of Crops (DJP), Government of Indonesia, show that smallholders' share of
resources but also to increase their own productivity (e.g., by investing the national oil palm acreage has increased gradually over time: from

⁎
Corresponding author.
E-mail address: [email protected] (V.V. Krishna).

https://doi.org/10.1016/j.ecolecon.2020.106829
Received 12 March 2020; Received in revised form 8 July 2020; Accepted 21 August 2020
Available online 02 September 2020
0921-8009/ © 2020 Elsevier B.V. All rights reserved.
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

10% in the 1980s to 20% in the 1990s and 40% in the 2010s (Jelsma Critics point out that the wealth generated in the oil palm sector is
et al., 2017; Li, 2018). Given that the total harvested area of oil palm largely appropriated by the large corporations, and does not result in
cultivation increased from 4 million hectares in 2000 to about 12 rural development (Pye, 2019). So far, not many have examined the
million hectares in 2018 in Indonesia (FAOSTAT, 2020), even a stable causal effect of oil palm expansion and associated land-use changes on
share would denote that oil palm cultivation by smallholders is ex­ the prevalence of and access to private and community goods in rural
panding rapidly and affecting more and more rural livelihoods both Indonesia.
directly and indirectly. However, intra-regional differences are sig­ Against this backdrop, we carried out a nation-wide study aiming to
nificant: smallholder oil palm cultivation is more prevalent in Sumatra examine the effects of oil palm expansion on rural households' access to
(contributing to 49% of the provincial oil palm area) and Sulawesi essential private goods (i.e., the direct effect on amenities owned and
(47%), as compared to Kalimantan (23%) and Papua (24%) used by individual households, such as an electricity connection in the
(Anonymous, 2015). house) and on community goods (i.e., the indirect effect on amenities
Oil palm cultivation by smallholder farms and large estates gen­ owned either by the community or the state and used by several
erates employment for millions of people in Indonesia and provides households in a village, such as streetlights). From the available data­
them with higher wages than they could earn by cultivating food crops sets, we could derive three private goods (household access to an
(Burke and Resosudarmo, 2012). For smallholder farm households, oil electricity connection, a piped water connection, and use of non-solid
palm cultivation is associated with high labor productivity, which is the fuel), and 12 community goods (including the prevalence of street
most constraining input in many parts of rural Indonesia (Euler et al., lights, schools, and asphalt roads). Access to private goods is de­
2017; Krishna et al., 2017). Empirical evidence for the welfare effects of termined primarily by household resources, but its availability could be
smallholder oil palm cultivation at the micro-level is mixed. Most ex­ determined by the characteristics of the village (e.g., location factors)
isting quantitative studies demonstrate positive livelihood outcomes, and certain public investment.1 Besides these 15 amenities, we have
such as increased consumption expenditure, household income and studied the effect of oil palm expansion on the number of statement
nutrition (Alwarritzi et al., 2016; Euler et al., 2017; Krishna et al., 2017; letters of poverty issued by the village authorities. Details of these
Sibhatu, 2019; Chrisendo et al., 2020; Qaim et al., 2020). However, variables are provided in the next section.
these studies were conducted in different provinces of Sumatra (e.g.,
Jambi), where the Indonesian Government has vigorously promoted oil 2. Methodology
palm expansion through transmigration and Nuclear Estate Programs
during the 1990s (Gatto et al., 2017; Kebschull, 2020). Hence the The global growth rate of palm oil production during the last dec­
findings of these studies may be extended to other provinces only with ades has been unprecedented – an increase of 525% was registered
caution, especially if no such institutional support existed during the during 1990–2018 – and a major share of this expansion took place in
initial expansion of oil palm cultivation in the region. The studies Indonesia.2 In the region, several studies have shown the overriding
conducted outside Sumatra indicate heterogeneous impacts of oil palm. importance of economic reasoning behind the decision of communities
For example, although the crop was found to be associated with sig­ and farmers to engage in oil palm cultivation and contract with the
nificant economic benefits in Kalimantan, these benefits were restricted plantation companies (Levang et al., 2016; Yuliani et al., 2020). How­
to village communities where farmers had prior knowledge of planta­ ever, we do not know whether the anticipated infrastructure develop­
tion management and exposure to the market economy (Santika et al., ment actually materialized in Indonesia after the oil palm boom. We
2019a; Santika et al., 2019b). created a regency-level panel dataset with private and community
The studies mentioned above focus mostly on smallholder farming goods and oil palm acreage for 2005–2014. The analyses were con­
systems. However, the potential effects of industrial plantations on ducted using instrumental variable fixed-effects models.
rural livelihoods, which include infrastructure development and di­
minishing access to land, are often overlooked. Only a few empirical
2.1. Data
studies have analyzed the impacts of oil palm expansion on livelihoods
at the national level (Kubitza and Gehrke, 2018; Edwards, 2019). An­
The key empirical analyses are based on the data from village
other nationally representative study conducted in Indonesia that ex­
census conducted by the Bureau of Statistics of Government of
amined the expansion of tree crops including oil palm is Ickowitz et al.
Indonesia (Badan Pusan Statistik, BPS), known as the Village Potential
(2016).
Statistics (Potensi Desa or PODES). The village census has been con­
In contrast to most quantitative studies, the existing qualitative case
ducted at intervals of approximately 3 years. Datasets from four rounds
studies emphasize the heightened social tensions and the margin­
of the census from 2005 to 2014 were used for the core analysis.3 The
alization of local communities due to oil palm plantations (Obidzinski
second major dataset is on oil palm expansion. Total oil palm area and
et al., 2014; Li, 2015, 2018), or the lack of interest of village commu­
area under smallholder cultivation are published by the Department of
nities in contracting with large-scale plantation companies in Indonesia
Agriculture annually under the Tree Crop Statistics of Indonesia
(Yuliani et al., 2018).
The existing empirical evidence shows that, despite significant en­
vironmental and social externalities associated with area expansion 1
Access to certain private goods necessitates prior government investment to
(Merten et al., 2016; Dislich et al., 2017; Grass et al., 2020), oil palm establish the complementary infrastructure. For example, the increased dis­
cultivation has the potential to facilitate rural economic development. posable income from oil palm may make electricity connection affordable to
However, the potential indirect effects of oil palm on rural develop­ rural households. However, without the government's decision to invest in
ment, such as creation of infrastructure, are rarely studied. While extending the power grid to and within the village, the causal effect of oil palm
Edwards (2019) and Kubitza and Gehrke (2018) investigated the effect in conferring this benefit may weaken or vanish completely.
2
of oil palm expansion on some private and community goods, they Global production of palm oil increased from 11 million tons in 1990 to
about 71 million tons in 2018 (FAOSTAT, 2020).
examined the effects on a limited range of variables only, and without 3
The data come from the PODES rounds of 2005, 2008, 2010, and 2014.
any robustness check. The lack of research effort in examining the
Although PODES 2000 and 2003 datasets were also acquired, we did not use
changes in rural infrastructure is surprising. After all, the area expan­ them to derive the dependent variables in the impact estimation, as the lagged
sion of oil palm is projected to bring millions of dollars as export tax data on oil palm area were not available. However, we used the PODES 2003
revenues to the Indonesian Government (Anderson et al., 2016; Naylor data to estimate the impact of the expansion of smallholder and industrial oil
et al., 2019). In addition, private investments in infrastructure by large- palm plantations, as three-year lags were available for the key explanatory
scale agricultural estates are expected to promote rural development. variable.

2
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

(Ministry of Agriculture, 2017). The data on total oil palm area dis­ exercised while interpreting the changes in SKTM provision over time.
aggregated at the regency level is only available from 2005. To use We have included several proxy variables for rural economic de­
additional PODES rounds and to apply time lags, we relied on data on velopment that are consistently available across different PODES da­
smallholder oil palm area, available at the regency-level between 2000 tasets. However, we have presented only a few of them in the main text
and 2014.4 for brevity, and the rest can be found in the supplementary materials.
Additionally, we used satellite data on the expansion of industrial
oil palm plantations to compensate for the lack of governmental data on
2.2. Establishing causality
large oil palm estates before 2005. We obtained oil palm area data from
2000, 2005, 2010, and 2015 based on Landsat data from Austin and
To estimate the average effect of oil palm expansion, access to
colleagues (Austin et al., 2017). Mapping was conducted by visually
various community and private goods is modeled as a function of the
interpreting the Landsat imagery. The authors created maps of large-
area share of oil palm at regency level. However, oil palm expansion
scale oil palm plantations in Sumatra, Kalimantan, and Papua at a re­
does not take place across the regencies randomly, and a direct com­
solution of 250 × 250 m. The industrial-scale oil palm data were not
parison of households' access to private and community goods and
available for Sulawesi, Java, and other smaller groups of islands such as
services between regencies with and without the crop could be mis­
Nusa Tenggara and Maluku. The oil palm area in Java was set to zero
leading. If unobserved heterogeneity is strongly correlated with the
based on the Tree Crop Statistics. Nevertheless, without Sulawesi and
outcome variables and oil palm expansion, the estimates will suffer
other smaller islands, we have a regionally more-restricted sample for
from endogeneity bias. Part of this bias could be avoided using a re­
industrial-scale oil palm acreage.
gency-level fixed-effects estimator that will control for the time-in­
The share of the oil palm area in regencies was calculated by di­
variant heterogeneity (Baltagi, 2010). Even after employing fixed-ef­
viding the total oil palm area obtained from the Tree Crop Statistics or
fects, the model might still suffer from reverse causality and bias from
satellite imagery by the total regency land area, obtained from the
the unobserved, time-variant heterogeneity (Kubitza and Krishna,
administrative maps of Indonesia. In order to combine data on oil palm
2020). We opted to use an instrumental variable fixed-effects model to
expansion with that on village development indicators, PODES data
address this issue.
were aggregated to the regency level, weighted by the total number of
Our instrument has both a cross-sectional (time-invariant) compo­
households residing in the village. We only considered rural regencies
nent and a time-variant component. The cross-sectional component is
(Kabupaten) and not urban cities (Kota), because plantations are scant
the average altitude of a regency based on SRTM (Shuttle Radar
in the densely populated cities due to the high opportunity cost of land.
Topography Mission) data. The intuition is that oil palm is most pro­
The spillover effects of oil palm plantations on urban livelihoods (e.g.,
ductive in altitudes below 400 m, while high altitudes and steep slopes
through labor migration) are possible, although one could assume these
decrease productivity as well as increase cultivation costs (Pirker et al.,
effects to manifest over a significantly longer period. After the exclusion
2016). Even within a province, oil palm cultivation by smallholders is
of urban cities, we obtained 248 observations per PODES round, each
found to decline with increasing altitude (Krishna et al., 2017). The
representing a regency. Regency-level government revenues and ex­
time-variant component is Indonesia's annual palm oil export.7 The
penditures were sourced from the Indonesia Database for Policy and
export magnitudes are mainly driven by external demand and national
Economic Research (INDO-DAPOER) database, administered by the
policies, and not by regency-specific factors. The altitude variable is
World Bank.5
interacted with the annual national palm oil export to generate the
For the current study, we captured the economic dimensions of rural
instrument. Our instrument is hence a proxy for how much each re­
development by assessing the changes in a set of proxy variables, such
gency's oil palm area should have grown, facing variation in global
as the share of villages within a regency with access to community
demand based solely on topographic characteristics. This approach to
goods (e.g., the share of villages in a regency with streetlights in the
establishing causality is inspired by the work of Duflo and Pande
main road, junior high schools, and asphalt roads), and the share of
(2007), which has been used by some recent studies (Lipscomb et al.,
households in a regency with private goods (e.g., electricity connection,
2013; Kubitza and Gehrke, 2018). The first stage of the model is as
piped drinking water, and non-solid fuel use at the house). In addition,
follows:
the changes in the number of households with statement letters of
poverty (Surat Keterangan Tidak Mampu; SKTM) were analyzed. These OilPalmjt = 0 + 1 (Altitudej Exportt ) + 2 x jt + rrt + uj + yt + ijrt
letters are issued to families that meet a series of absolute poverty in­ (1)
dicators (Morgans et al., 2018). However, the number of SKTM issued
stands proxy for two mutually conflicting dimensions of rural devel­ where Altitudej is the average altitude in regency j and Exportt is the
opment: access to social assistance programs in the region, as well as national palm oil export volume in year t. Oil Palmjt is the share of oil
the level of rural poverty (Priebe et al., 2014).6 Caution should thus be palm area in the regency and xjt is a vector of time-variant explanatory
variables (i.e., the annual export of palm oil and the number of villages
in each regency). uj presents regency-level fixed-effects, yt the year
4 dummies (representing the year of PODES data collection), rrt presents
There is ongoing discussion on whether the smallholder oil palm area, as
region-specific time trends capturing differing trends common to each
defined by the government, includes only the area managed by smallholder
farmers. Indeed, there exists some evidence for large-scale oil palm estates
region (Java, Sumatra, Kalimantan, Sulawesi and other outer islands).8
being labeled as smallholder farms (Jelsma et al., 2017). Such data irregula­ The second stage of the instrumental variable fixed-effects models is as
rities are neither rare in developing countries nor unique to oil palm. Due to follows:
these data problems, we used three different variables for oil palm expansion,
and compare the estimates to ensure robustness. Amenityijrt = 0 + 1 OilPalmjt + 2 x jt + rrt + uj + yt + ijrt (2)
5
Data source: https://datacatalog.worldbank.org/dataset/indonesia-
database-policy-and-economic-research
6 7
The impact of oil palm on rural poverty has been examined by Edwards Supplementary Figure S1 reports palm oil exports over time. Although ex­
(2015, 2019), and the impact of different tree-dominated land uses (which ports are continuously increasing, there is significant variation in the rate of
includes oil palm) on dietary quality and people's consumption of food groups growth over time.
8
rich in micronutrients was examined by Ickowitz et al. (2016). These studies Even without oil palm expansion, these regions might have followed dif­
indicated a crucial role of oil palm/tree crops in reducing poverty in rural In­ ferent trends in infrastructure development due to their geographic isolation
donesia. To avoid duplication, we are not assessing the poverty impacts of oil and population density. To address this issue, we use regional trends in all our
palm in this study. estimations.

3
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

where Amenityijrt is the share of villages or households with access to 0.50

Share of villages (0-1) where plantation crops are the major source of
amenity i in regency j in year t. We conduct a robustness check, al­
lowing for differing time trends between regencies based on a large
number of variables, which are likely to correlate with remoteness. 0.40
We also test the hypothesis that oil palm expansion has differential
effects on rural economic development depending on the operational
scale of cultivation – either at smallholder scale (independently or 0.30
within the Nucleus Estate Scheme or as farmer cooperatives) or at in­
dustrial scale. Unlike the data on total oil palm acreage, data on

income
smallholder and industrial scale oil palm expansion reach back to 2000,
which allows us to use time lags for oil palm expansion. We use three- 0.20

year lags in the analysis.9

As the final step, the reasons for the observed impacts (or the lack
thereof) of oil palm expansion are explained by analyzing the govern­ 0.10
ment revenue and expenditure datasets. The empirical model used was
similar to that of Eq. (2). However, for the models with government
revenue and expenditure as the dependent variables, lags of oil palm 0.00
area were not used. Considerable tax revenues are obtained in the very 2000 2003 2005 2008 2011 2014
first year of plantation establishment, due to timber extraction in PODES rounds (years)

clearing land for oil palm (e.g., Forest Resource Provision Fee, PSDH) Sumatra Java, Bali and smaller islands
(Obidzinski et al., 2013). Kalimantan Sulawesi
Papua and Maluku Indonesia

3. Results and discussion Fig. 1. Share of Indonesian villages where cultivation of tree crops is the most
important economic activity.
3.1. Descriptive statistics Source: PODES (different rounds).

Tree plantation (estate) crops contribute significantly to Indonesia's 2005–2014, while the share of regencies with oil palm plantations in­
rural economy (Edwards, 2019).10 Despite the significant regional creased from 39% to 47%.
heterogeneity, the cultivation of these crops remains the major eco­ A comparison of private and community goods between regencies
nomic activity in about 17% of rural villages (Fig. 1). The share of such with and without oil palm in 2005 and 2014 (Appendix Table A1)
villages has increased over time due to the high profitability of plan­ shows the following patterns. As per the PODES 2005 data, regencies
tation crops in contrast to food crops, and to monetary and non- with oil palm were lagging behind those without with respect to five
monetary incentives from the Indonesian government (Gatto et al., community goods (streetlights, asphalt roads, kindergartens, primary
2017). Oil palm is the most important plantation crop of Indonesia, schools, and maternity facilities). Regencies without oil palm lagged
covering more than 50% of the national plantation area, and it is the behind in six cases (junior high schools, post offices, police posts, sports
only plantation crop with consistent area expansion over time (Sup­ facilities, hospitals, and markets without a permanent structure). Re­
plementary Fig. S2). In the islands where the economic relevance of gencies with oil palm continued to lag behind those without in 2014 for
plantation crops to the local economy is high – e.g., Sumatra and Ka­ the same five community goods as in 2005. However, the situation had
limantan – oil palm cultivation is the major land-use system (Supple­ improved for the non-oil palm regencies during this period. Concerning
mentary Fig. S3). In 2015, oil palm was the main source of income in the three private goods examined, there were no significant differences
about 44% of villages with plantation crops in Sumatra and 39% of between regencies with and without oil palm in either 2005 or 2014.
villages in Kalimantan. Natural rubber is the main competing crop in We found significant increases over time in the prevalence of all private
these islands. In other parts of Indonesia, such as Sulawesi and Papua, goods and some community goods in most regencies.
coconut and cocoa are the competing crops of oil palm. All these differences and changes are highly significant from the
Significant spatial and temporal variation in the prevalence of oil perspective of rural development. Nevertheless, one cannot attribute
palm cultivation is evident at the provincial level, as shown in Table 1. them solely to the cultivation of oil palm. Oil palm cultivation has been
Oil palm cultivation has been minimal in Java and Bali, and in the outer increasing mainly in Sumatra and Kalimantan, and these regions differ
islands like Maluku and Papua. In Sulawesi, however, oil palm has been in several ways (e.g., a lower population density) from Java, where oil
cultivated in several regencies, although the land area coverage re­ palm is not cultivated widely. The pathway of oil palm expansion was
mains low (2–3%), in comparison with Kalimantan (5–13%) and Su­ also dissimilar, even between Kalimantan and Sumatra. In Sumatra, oil
matra (6–23%). In all the provinces that had oil palm in 2005, a sig­ palm plantations have expanded recently onto non-forest land, while in
nificant increase in the plantation area was registered over the Kalimantan, the expansion mostly took place at the expense of forest
subsequent decade. At the same time, the crop was introduced in some land (Austin et al., 2017; Jelsma et al., 2017).
provinces, such as Sulawesi Uttara. As a result, the oil palm area in­
creased from 3.1% to 5.7% of the total land area of Indonesia during
3.2. Impact of oil palm on rural economic development
9
For this analysis, we use data on oil palm acreage from years 2000, 2005 The causal effects of oil palm expansion on community and private
and 2011 (2010 in the case of the satellite data) and PODES data from 2003, goods are estimated using instrumental variable fixed-effects regression
2008 and 2014. We have data on 193 regencies across the three waves due to
models for 2005–2014. In general, access to private goods is directly
the discussed data limitations from the satellite data with two missing values,
linked to increases in the disposable income of rural households.
which results in a total of 577 observations.
10
The categories of tree plantations or estate crops used commonly by the However, the provision of community goods is also based on the ca­
Government of Indonesia are oil palm, natural rubber, coconut, cacao, and pacities and willingness of local and central governments to tax the
clove. The production data are available at the Tree Crop Statistics of Indonesia increased income and eventually invest in the public infrastructure.
yearbooks, published by the Directorate General of Estate Crops at the Besides, private investment by companies might also increase the pro­
Department of Agriculture, Government of Indonesia. vision of infrastructure. The effect of oil palm expansion – combining

4
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

Table 1
Presence and share of land area under oil palm in Indonesian provinces.
Province name Number of regencies PODES 2005 PODES 2014
(Kabupaten)
Share of regencies with oil Share of land area under oil Share of regencies with oil Share of land area under
palm palm palm oil palm

Java and Bali

Bali 8 0.000 – 0.000 –
Java Barat 18 0.167 0.005 0.389 0.007
Java Tengah 29 0.000 – 0.000 –
Java Timur 29 0.000 – 0.000 –
D.I. Yogyakarta 4 0.000 – 0.000 –
Sumatra
Bengkulu 3 1.000 0.075 1.000 0.149
Aceh 11 0.909 0.046 1.000 0.070
Jambi 5 0.800 0.082 1.000 0.141
Lampung 8 1.000 0.045 1.000 0.057
Riau 9 1.000 0.137 1.000 0.244
Sumatera Barat 9 0.889 0.071 0.778 0.095
Sumatera Selatan 8 1.000 0.067 1.000 0.111
Sumatera Utara 13 0.923 0.129 1.000 0.180
Kalimantan
Kalimantan Barat 7 1.000 0.026 1.000 0.054
Kalimantan Selatan 9 0.667 0.037 1.000 0.134
Kalimantan Tengah 5 0.800 0.029 1.000 0.074
Kalimantan Timur 4 1.000 0.010 1.000 0.046
Sulawesi
Sulawesi Selatan 22 0.182 0.012 0.364 0.026
Sulawesi Tengah 4 0.600 0.008 0.750 0.023
Sulawesi Tenggara 4 0.250 < 0.001 0.750 0.013
Sulawesi Utara 4 0.000 – 0.250 0.002
Outer islands
Maluku 4 0.000 – 0.250 0.003
Nusa Tenggara Barat 6 0.000 – 0.000 –
Nusa Tenggara Timur 12 0.000 – 0.000 –
Papua 12 0.250 0.001 0.333 0.002
Indonesia 247 0.393 0.031 0.470 0.057

Notes: In total, there are 248 eligible regencies, but there is missing information for one regency in both 2005 and 2015. Due to multiple splits of regencies and
provinces over time, we merged our data based on the boundaries of the 26 provinces existent before 1999. We omitted the Special Capital Region of Jakarta since
this region does not include any rural regencies.

acreage under smallholder management with private and government PODES data, and focused on the differences between private and
estates – is shown as Model 1 (Table 2). The first stage Kleibergen–Paap community goods.
rk Wald F-statistic > 10.0 indicates that our instrument is strong.11 In the next step, we test if our results are sensitive to (a) the type of
We find that oil palm expansion has consistently increased rural producer by comparing smallholder farmers with large-scale estates, and (b)
households' access to private goods such as piped drinking water and the use of time lags in our explanatory variable. We opted for a three-year
non-solid fuel use while reducing the number of poverty letters issued. lag for oil palm and applied the same instrumental variable fixed-effects
Increased disposable income of farm households and wages of planta­ models as for Model (1). We find that both smallholder and industrial-scale
tion workers are the likely cause of this positive trend. However, the oil palm expansion significantly increased the provision of private goods
provision of certain public goods is a necessary prerequisite for in­ (Model 2 and 3, Table 2). While we observe that the magnitude of the
creasing the access to these private goods (e.g., an electricity grid re­ effects is larger for smallholder oil palm expansion, a direct comparison
quired in the village for household electricity connection, cf. Footnote between these two regimes is not possible because of the overlaps in the two
1). Due to the lack of information on these supplementary community datasets. Furthermore, there could be significant spillover effects. Moreover,
goods in the PODES datasets, further analyses could not be conducted. smallholder oil palm acreage could be spatially correlated with that of in­
On the other hand, the effects of oil palm cultivation are insignif­ dustrial estates.12 It is surprising that we still could not observe a consistent
icant for most community goods. For brevity, we have included a se­ positive effect of industrial-scale oil palm expansion on community goods,
lected subset of three community goods in Table 2. The others are which is one of the often-cited promises of the oil palm industry.
presented in Supplementary Table S2. The oil palm effects are found The expansion of oil palm cultivation between 2000 and 2014 did not
statistically insignificant for the establishment of most community increase the expenditure of regency governments (Table 3), which is the key
goods, including streetlights, junior high schools, asphalt roads, post determinant of the creation of public infrastructure since decentralization in
offices, police posts, sports facilities, kindergartens, markets without a 2001. The dearth of targeted investment in rural development is surprising,
permanent structure, and maternity facilities. A positive effect above as there has been a significant positive effect of oil palm expansion on tax
the 5% significance level was observed for only two out of twelve revenues during this period. Only a small share of palm oil tax revenues –
community goods – the share of villages with hospitals, and markets less than 15% in 2012/2013, according to Falconer et al. (2015) – is allo­
with a permanent structure. Although some of these effects are similar cated to the regional governments. Between the center and the provinces,
to that of Edwards (2019), our conclusions differ because we examined there is also no formal arrangement for sharing resource revenues for the
the impact on a larger set of commodities, analyzed more rounds of
12
Smallholder plantations depend on the availability of palm oil mills that
are often placed adjacent to industrial plantations. The spillover effects might
11
First-stage results are reported in Supplementary Materials (Table S1). even be the reason for the observed similarity in the effect magnitudes.

5
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

plantation sector, unlike for forestry and for mining.13 The statistically in­ palm expansion created large private benefits (Euler et al., 2017; Edwards,
significant effect of oil palm cultivation on community goods during 2019; Chrisendo et al., 2020). However, rural economic development is
2005–2014 is hence understandable and conforms with the evidence pre­ more than the creation of private wealth. Oil palm expansion does not seem
sented by some of the published case studies (Li, 2015). to have contributed significantly to the development of shared community
goods, with a few exceptions, during the studied period. The lack of access
3.3. Robustness of the estimates to community goods on the fringes of oil palm plantations could result in an
increasing gap between the wealthier and the poorer segments of commu­
The robustness of the impact estimates of the models mentioned nities or villages. The rural income-poor rely heavily on the provision of
above was inspected through different statistical tests and alternative communal resources to sustain their livelihoods, not only on material public
model estimation. Supplementary Tables S3 and S4 report fixed-effects infrastructure but also on ecosystem goods and services (Kabubo-Mariara,
estimates without employing an instrumental variable to correct time- 2013; Nguyen et al., 2020). Therefore, the lack of targeted government
variant heterogeneity. The results again show significant increases for spending on the plantation fringes could indirectly enhance not only inter-
private goods but not for community goods. However, the magnitudes regency inequality but also the overexploitation of natural resources.
are considerably smaller compared to the estimates in Table 2 and
Table 3, which we attribute to (a) estimating the local average treat­ 4.1. Limitations of the study
ment effects (LATE) in the instrumental variable regression model, (b)
the inability to account for the time-variant heterogeneity in the simple The present study has certain limitations, especially concerning the
fixed-effects estimates, and (c) the possible measurement error in the analytical approach. To establish causality, we have primarily used the
variable – oil palm area share in the regency – which is adjusted only in fixed-effects instrumental variable framework. True, this approach ac­
the instrumental variable framework. Furthermore, without an instru­ counts for a large number of confounding factors that cannot be in­
mental variable, the fixed-effects models do not facilitate any compar­ corporated in the cross-sectional regression OLS models. However, our
ison with regencies where oil palm has never been cultivated during the estimates are the local average treatment effects (LATE). Hence, al­
study period. In Supplementary Table S3, Model 2 uses satellite data to though we are confident regarding the direction of causality, the effect
estimate the oil palm area. These data are likely to be less affected by size should be interpreted with caution; the LATE could diverge from
measurement error than those using government data, and hence less the average population effects (Imbens and Angrist, 1994). Further­
likely to be biased toward zero even in simple fixed-effect models. In­ more, the analysis conducted at the regency level leaves us with a ra­
deed, simple fixed-effects estimates and instrumental variable fixed- ther small sample of observations, which, in the end, limits our ability
effects estimates are more similar for Model 2 than for the other models. to fully examine the temporal and spatial heterogeneities of impact.14
While it is not possible to test the exogeneity restriction for our in­
strumental variable, we can conduct a falsification test by checking whether
our instrument is significantly affecting the provision of goods in regions 4.2. The policy implications
without oil palm. The estimates are reported in Supplementary Table S5.
We observe that in almost all regression models, the coefficient of the in­ The oil palm expansion in Indonesia has been highly controversial and is
strument is statistically insignificant. This falsification test supports the se­ one of the most contested agrarian changes in the world due to its detri­
lection of the instrumental variable. mental social and environmental effects in regencies with a large forest
While we control for fixed characteristics in our main results, we do not cover. Potential ways to regulate unsustainable oil palm cultivation and
control for diverging trends. In particular, our instrument may correlate associated land-use changes include a trade ban on unsustainable palm oil
with remoteness. It is hence necessary to verify that our instrument is not and associated products, the direct transfer of benefits to producer countries
picking up diverging trends between remote and less remote villages. As an for not establishing plantations within forest land, and certification for
additional robustness check, we allow for different regency time trends sustainable palm oil production (Venter et al., 2009; Carlson et al., 2018).
based on the initial conditions that correlate with remoteness. We include Feasibility analyses of any of these mechanisms require a clear under­
time trends based on the initial values from PODES 2000, such as the share standing of how oil palm cultivation benefits rural communities in produ­
of the labor force employed in agriculture, or the share of villages that cing countries. A complete ban on the import and use of palm oil sourced
border the forest. The results, shown in Supplementary Table S6, are similar from recently deforested areas and peatlands may have significant im­
to Model 1 (Table 2). We also examine the robustness of our results in the plications on rural livelihoods and poverty in the tropics. Several environ­
case of the other community goods. Again, including additional time trends mental policies have advocated financial payments as compensation for the
based on the initial values does not alter our results, which reiterates that oil foregone benefits from oil palm cultivation to deter rural communities from
palm expansion did not have a significant impact on the provision of most further deforestation and conversion of peatlands. Many of these benefit
community goods (Supplementary Table S7). A significant (and positive) transfers, although promising in theory, face difficulties in the field. The
impact was observed only for hospitals and markets with a permanent sustainability standards are also criticized, for example, for not encouraging
structure. We also test if our results are sensitive to longer time differences independent unions or independent smallholders to develop autonomous
in the estimation (instead of using the three-year intervals). Even when we production on their own land (Li, 2018). The lack of rigorous enforcement
estimated the oil palm impacts using only the PODES rounds of 2005 and of the principles and criteria of the Roundtable on Sustainable Palm Oil
2014, our results remained stable (Supplementary Table S8). (RSPO), a market-based sustainability certification mechanism, does not
lead to significant improvements in sustainability metrices (Morgans et al.,
2018). These criticisms open a passionate debate about the incentive
4. Conclusion and policy implications
structure and implementation institutions that would maximize the welfare
In this study, we have tried to answer the question, ‘does oil palm
cultivation function as the engine of economic development in Indonesia?’. 14
The effect size could have been drastically different in the initial stages of
We used nation-wide panel data on oil palm expansion and household ac­ oil palm expansion in Indonesia (i.e., the 1990s), as compared to the study
cess to community and private goods and showed that the expansion had period. However, delineating the effect of oil palm from that of the initial
positive effects on some but not all aspects of rural economic development. government support programs (e.g., Nuclear Estate Scheme) is extremely
Our findings are in tune with previous studies, which had shown that oil challenging. One might also surmise that the impact of the crop could be dif­
ferent in the same period across the different islands of Indonesia, especially
between Sumatra and Kalimantan. These islands differ concerning the nature
13
We thank an anonymous reviewer for pointing out this aspect. and timing of oil palm expansion (Jelsma et al., 2017; Krishna et al., 2017).

6
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

Table 2
Effect of oil palm on community and private goods.
Prevalence of community goods Access to private goods Poverty letters
(SKTM)
Street light Junior high Asphalt road Electricity Piped drinking Nonsolid fuel
school connection water use

(a) Model 1 (n = 980)

Share of total oil palm area in the regency 0.729 0.596 −0.558 2.676⁎ 2.550⁎⁎⁎ 1.780⁎ −0.669⁎
(0.886) (0.400) (0.618) (1.588) (0.719) (1.039) (0.348)
F-Stat 11.935 44.974 26.733 106.991 12.422 120.678 34.238

(b) Model 2 (n = 575)

Share of industrial-scale oil palm area in the 1.980 1.282 0.696 4.002⁎⁎⁎ 3.719⁎⁎⁎ 3.249⁎⁎ −1.954⁎⁎⁎
regency (1.524) (0.839) (1.176) (1.408) (1.153) (1.440) (0.538)
F-Stat 20.373 43.051 33.028 133.129 18.437 138.391 27.475

(c) Model 3 (n = 575)

Share of smallholder oil palm area in the 2.950 1.910 1.037 5.963⁎⁎ 5.541⁎⁎ 4.841⁎⁎ −2.911⁎⁎⁎
regency (2.448) (1.305) (1.794) (2.706) (2.440) (2.432) (1.097)
F-Stat 16.479 46.709 32.020 174.944 13.355 121.459 25.482

Notes: For Model 1, data from 2005 to 2014 are analyzed, as data on aggregate oil palm area at the regency level are not available for 2003, and data on the industrial
scale and smallholder oil palm are not compatible to aggregate. For Models 2 and 3, PODES data from 2003 to 2014 are analyzed, and we use a three-year lag for the
oil palm variable. Instrumental variable regression estimates are reported using regency's average altitude multiplied by the annual Indonesian exports of palm oil as
the instrument. In these models, two observations are dropped due to singleton groups. Standard errors clustered at the regency level are shown in parentheses.
Regional time trends, year and regency fixed-effects, annual export of palm oil, and the number of villages in each regency are included as the additional explanatory
variables. ⁎p < 0.10, ⁎⁎p < 0.05, ⁎⁎⁎p < 0.01. The Kleibergen–Paap rk Wald F-statistic is 22.94 for Model 1, 24.43 for Model 2, and 8.60 for Model 3.

Table 3
Effect of oil palm on regency governments' revenue and expenditure.
Total government revenue per capita (log) Total government expenditure per capita (log)

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

Share of total oil palm area in the regency 1.309 −2.522

(2.703) (1.861)
Share of smallholder oil palm area in the regency 6.139⁎ 2.631
(3.264) (2.236)
Share of industrial oil palm area in the regency 4.750 ⁎
2.182
(2.595) (1.866)
F-Stat 123.898 600.869 620.582 460.940 619.815 624.868
Kleibergen–Paap rk Wald F-statistic 16.57 20.43 18.68 21.97 24.00 21.47
Number of observations 644 678 678 726 746 746

Notes: Instrumental variable regression estimates are reported using a regency's average altitude times the annual Indonesian exports of palm oil as the instrument.
Standard errors clustered at the regency level are shown in parentheses. Regional time trends, year and regency fixed-effects, and annual export of palm oil are
included as the additional control variables. Model (1) uses data from PODES 2005, 2010, and 2014 rounds, while Models (2) and (3) use 2000, 2005, 2010, 2014
rounds of data. Data for several regencies are missing, reducing the total number of observations. ⁎p < 0.10, ⁎⁎p < 0.05, ⁎⁎⁎p < 0.01.

impacts of oil palm cultivation among smallholders. governments to carry out the production and maintenance of communal
Many researchers have pointed out that one of the reasons for cus­ goods, could enhance the developmental effects of oil palm cultivation.
tomary landholders to engage in contracts with the plantation companies
was that the contract would result in increased access to schools, markets, Declaration of competing interest
health services, and jobs (Rist et al., 2010; Li, 2015). Besides, large-scale
agricultural investments are widely regarded as vehicles for the develop­
The authors declare that they have no known competing financial
ment of communal infrastructure in the face of the limited capability of
governments in developing countries (Deininger and Byerlee, 2011). The interests or personal relationships that could have appeared to influ­
ence the work reported in this paper.
insignificant effect of oil palm expansion on the provision of most com­
munity goods is hence a sobering note. Our findings complement that of
Santika et al. (2019a) and Santika et al. (2019b), which illustrated that the Acknowledgment
developmental effects of oil palm need not always be positive. The central
and regional governments should not assume that rural infrastructure de­ This study was funded by the Deutsche Forschungsgemeinschaft
velopment will be instigated de facto by the private plantation sector in the (German Research Foundation/DFG; project number 192626868) in the
regions where oil palm is being cultivated. Increased government invest­ framework of the collaborative German-Indonesian research project
ment at the fringes of oil palm plantations is critical in creating and CRC990. We thank the four anonymous reviewers of the journal for
maintaining community goods and services in rural Indonesia. Certain fiscal their valuable comments on the previous manuscript draft. We are also
reforms, such as increasing the share of tax revenues from oil palm for local grateful to Matin Qaim for valuable suggestions and support, and Ms.
governments, and earmarking revenues in order to encourage local Liz Lucas for language editing.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ecolecon.2020.106829.

7
 Appendix

Table A1
Prevalence and access to community and private goods in Indonesia.

Indicator variables of rural economic development Definition Status as of 2005 in regencies Status as of 2014 in regencies
V.V. Krishna and C. Kubitza

With oil palm (n = 97) Without oil palm (n = 147) With oil palm (n = 116) Without oil palm (n = 131)

Community goods: The share of villages per regency with -

Streetlight Streetlights at the main road 0.553*** 0.713# 0.667### 0.779***
(0.212) (0.311) (0.193) (0.299)
Asphalt road An asphalt main-road 0.629*** 0.712### 0.685## 0.844***
(0.172) (0.204) (0.161) (0.191)
Kindergarten At least one kindergarten 0.520*** 0.732### 0.732### 0.862***
(0.223) (0.243) (0.214) (0.204)
Primary school At least one primary school 0.938*** 0.979 0.935 0.973***
(0.113) (0.058) (0.102) (0.093)
Junior high school At least one junior high school 0.496* 0.460### 0.574### 0.583
(0.185) (0.147) (0.177) (0.165)
Post office At least one post office 0.127*** 0.098 0.100## 0.099
(0.088) (0.058) (0.069) (0.068)
Police post At least one police (security) post 0.242*** 0.159 0.212 0.172***
(0.149) (0.073) (0.117) (0.084)
Sports facility Any sports facility (e.g., football grounds) 0.410** 0.374### 0.360### 0.335
(0.117) (0.107) (0.123) (0.114)
Hospitals At least one hospital 0.043* 0.035### 0.044 0.052
(0.038) (0.031) (0.040) (0.061)
Maternity facility At least one maternity clinic 0.493* 0.557## 0.542 0.639**

8
(0.204) (0.305) (0.253) (0.352)
Market with a permanent structure At least one market that has a permanent structure 0.267 0.260### 0.311## 0.312
(0.159) (0.136) (0.160) (0.154)
Market without a permanent structure At least one market that does not have a permanent structure 0.221*** 0.140 0.156### 0.131
(0.118) (0.079) (0.130) (0.125)

Private goods: Share of households per regency -

Electricity connection Having electricity connection in the house 0.673 0.675### 0.933### 0.966
(0.117) (0.136) (0.069) (0.116)
Piped drinking water Having access to piped water in the house 0.106 0.121### 0.207### 0.221
(0.103) (0.106) (0.145) (0.162)
Nonsolid fuel use Using non-solid fuel for cooking 0.456 0.449### 0.720### 0.761
(0.224) (0.272) (0.210) (0.233)
SKTM Issued with statement letters of poverty (SKTM) 0.066 0.056### 0.122### 0.100*
(0.036) (0.030) (0.060) (0.048)

Note: Figures in parentheses show the standard deviation of the sample mean values. Household indicators are weighted by the number of households per village. We use a t-test to determine if there is a significant
difference between the means of regency with and without oil palm. The subscripts * and # denote the statistical significance of the difference in the two-tailed t-test. Here *, **, *** show that the difference between with
and without oil palm categories in a single year is significant at 0.10, 0.05, and 0.01 levels, respectively. #, ##, ### show that the difference between 2005 and 2014 data within the same category is significant at 0.10,
0.05, and 0.01 levels, respectively. The details of the PODES datasets from which these variables were derived are available online at https://mikrodata.bps.go.id/mikrodata/index.php/catalog/PODES.
Ecological Economics 179 (2021) 106829
V.V. Krishna and C. Kubitza Ecological Economics 179 (2021) 106829

References 990 - EFForTS, Ecological and Socioeconomic Functions of Tropical Lowland

Rainforest Transformation Systems (Sumatra, Indonesia). Discussion Paper 22.
University of Goettingen, Goettingen.
Alwarritzi, W., Nanseki, T., Chomei, Y., 2016. Impact of oil palm expansion on farmers’ Kubitza, C., Krishna, V.V., 2020. Instrumental variables and the claim of causality: evi­
crop income and poverty reduction in Indonesia: an application of propensity score dence from impact studies in maize systems. Glob. Food Secur. 26, 100383. https://
matching. J. Agric. Sci. 8 (1), 119–131. https://doi.org/10.5539/jas.v8n1p119. doi.org/10.1016/j.gfs.2020.100383.
Anderson, Z.R., Kusters, K., McCarthy, J., Obidzinski, K., 2016. Green growth rhetoric Levang, P., Riva, W.P., Orth, M.G., 2016. Oil palm plantations and conflict in Indonesia:
versus reality: insights from Indonesia. Glob. Environ. Chang. 38, 30–40. https://doi. evidence from West Kalimantan. In: Cramb, R., McCarthy, J.F. (Eds.), The Oil Palm
org/10.1016/j.gloenvcha.2016.02.008. Complex: Smallholders, Agribusiness, and the State in Indonesia and Malaysia. NUS
Angelsen, A., 2010. Policies for reduced deforestation and their impact on agricultural Press, Singapore, pp. 283–300.
production. Proc. Natl. Acad. Sci. U. S. A. 107 (46), 19639–19644. https://doi.org/ Li, T.M., 2015. Social impacts of oil palm in Indonesia: a gendered perspective from West
10.1073/pnas.0912014107. Kalimantan. In: Occasional Paper 124. Center for International Forestry Research
Anonymous, 2015. Tree Crop Estate Statistics of Indonesia 2014: Oil Palm. Directorate (CIFOR), Bogor, Indonesia.
General of Estate Crops, Jakarta, Indonesia. Li, T.M., 2018. After the land grab: infrastructural violence and the “Mafia System” in
Austin, K.G., Mosnier, A., Pirker, J., McCallum, I., Fritz, S., Kasibhatla, P.S., 2017. Shifting Indonesia’s oil palm plantation zones. Geoforum 96, 328–337. https://doi.org/10.
patterns of oil palm driven deforestation in Indonesia and implications for zero-de­ 1016/j.geoforum.2017.10.012.
forestation commitments. Land Use Policy 69, 41–48. https://doi.org/10.1016/j. Lipscomb, M., Mobarak, A.M., Barham, T., 2013. Development effects of electrification:
landusepol.2017.08.036. evidence from the topographic placement of hydropower plants in Brazil. Am. Econ.
Baltagi, B.H., 2010. Econometric Analysis of Panel Data, 4th ed. Wiley, Chichester. J. Appl. Econ. 5 (2), 200–231. https://doi.org/10.1257/app.5.2.200.
Burke, P.J., Resosudarmo, B.P., 2012. Survey of recent developments. Bull. Indones. Econ. Merten, J., Röll, A., Guillaume, T., Meijide, A., Tarigan, S., Agusta, H., Dislich, C., Dittrich, C.,
Stud. 48 (3), 299–324. https://doi.org/10.1080/00074918.2012.728620. Faust, H., Gunawan, D., Hein, J., Hendrayanto, Knohl, A., Kuzyakov, Y., Wiegand, K.,
Byerlee, D., de Janvry, A., Sadoulet, E., 2009. Agriculture for development: toward a new Hölscher, D., 2016. Water scarcity and oil palm expansion: social views and environmental
paradigm. Ann. Rev. Resour. Econ. 1 (1), 15–31. https://doi.org/10.1146/annurev. processes. Ecol. Soc. 21 (2). https://doi.org/10.5751/ES-08214-210205.
resource.050708.144239. Ministry of Agriculture, 2017. Basis Data Statistik Pertanian (BDSP). Government of
Carlson, K.M., Heilmayr, R., Gibbs, H.K., Noojipady, P., Burns, D.N., Morton, D.C., Indonesia. http://aplikasi.pertanian.go.id/bdsp/newdata.asp Accessed 27 July 2017.
Walker, N.F., Paoli, G.D., Kremen, C., 2018. Effect of oil palm sustainability certifi­ Morgans, C.L., Meijaard, E., Santika, T., Law, E., Budiharta, S., Ancrenaz, M., Wilson,
cation on deforestation and fire in Indonesia. Proc. Natl. Acad. Sci. U. S. A. 115 (1), K.A., 2018. Evaluating the effectiveness of palm oil certification in delivering mul­
121–126. https://doi.org/10.1073/pnas.1704728114. tiple sustainability objectives. Environ. Res. Lett. 13, 064032.
Chrisendo, D., Krishna, V.V., Siregar, H., Qaim, M., 2020. Land-use change, nutrition, and Naylor, R.L., Higgins, M.M., Edwards, R.B., Falcon, W.P., 2019. Decentralization and the
gender roles in Indonesian farm households. Forest Policy Econ. 118, 102245. environment: assessing smallholder oil palm development in Indonesia. Ambio 48
https://doi.org/10.1016/j.forpol.2020.102245. (in press). (10), 1195–1208. https://doi.org/10.1007/s13280-018-1135-7.
Deininger, K., Byerlee, D., 2011. Rising Global Interest in Farmland: Can it Yield Nguyen, T.T., Do, T.L., Halkos, G., Wilson, C., 2020. Health shocks and natural resource
Sustainable and Equitable Benefits? The World Bank, Washington, DC, USA. extraction: a Cambodian case study. Ecol. Econ. 169, 106517. https://doi.org/10.
Dethier, J.-J., Effenberger, A., 2012. Agriculture and development: a brief review of the 1016/j.ecolecon.2019.106517.
literature. Econ. Syst. 36 (2), 175–205. https://doi.org/10.1016/j.ecosys.2011.09. Obidzinski, K., Takahashi, I., Dermawan, A., Komarudin, H., Andrianto, A., 2013. Can
003. large scale land acquisition for agro-development in Indonesia be managed sustain­
Dislich, C., Keyel, A.C., Salecker, J., Kisel, Y., Meyer, K.M., Auliya, M., Barnes, A.D., ably? Land Use Policy 30 (1), 952–965. https://doi.org/10.1016/j.landusepol.2012.
Corre, M.D., Darras, K., Faust, H., Hess, B., Klasen, S., Knohl, A., Kreft, H., Meijide, A., 06.018.
Nurdiansyah, F., Otten, F., Pe’er, G., Steinebach, S., Tarigan, S., Tölle, M.H., Obidzinski, K., Dermawan, A., Hadianto, A., 2014. Oil palm plantation investments in
Tscharntke, T., Wiegand, K., 2017. A review of the ecosystem functions in oil palm Indonesia’s forest frontiers: limited economic multipliers and uncertain benefits for
plantations, using forests as a reference system. Biol. Rev. Camb. Philos. Soc. 92 (3), local communities. Environ. Dev. Sustain. 16 (6), 1177–1196. https://doi.org/10.
1539–1569. https://doi.org/10.1111/brv.12295. 1007/s10668-014-9519-8.
Duflo, E., Pande, R., 2007. Dams. Q. J. Econ. 122 (2), 601–646. https://doi.org/10.1162/ Pirker, J., Mosnier, A., Kraxner, F., Havlík, P., Obersteiner, M., 2016. What are the limits
qjec.122.2.601. to oil palm expansion? Glob. Environ. Chang. 40, 73–81. https://doi.org/10.1016/j.
Edwards, R., 2015. Is plantation agriculture good for the poor? Evidence from Indonesia's gloenvcha.2016.06.007.
palm oil expansion. In: Working Papers in Trade and Development 12. Arndt-Corden Priebe, J., Howell, F., Pankowska, P., 2014. Determinants of access to social assistance
Department of Economics, Australian National University, Canberra. programmes in Indonesia: empirical evidence from the Indonesian family life survey
Edwards, R., 2019. Export Agriculture and Rural Poverty: Evidence from Indonesian Palm (IFLS) East 2012. In: TNP2K Working Paper 11b-2014. Tim Nasional Percepatan
Oil. QLD, Brisbane, 73 pp. Penanggulangan Kemiskinan (TNP2K), Jakarta, Indonesia.
Euler, M., Krishna, V., Schwarze, S., Siregar, H., Qaim, M., 2017. Oil palm adoption, Pye, O., 2019. Commodifying sustainability: development, nature and politics in the palm oil
household welfare, and nutrition among smallholder farmers in Indonesia. World industry. World Dev. 121, 218–228. https://doi.org/10.1016/j.worlddev.2018.02.014.
Dev. 93, 219–235. https://doi.org/10.1016/j.worlddev.2016.12.019. Qaim, M., Sibhatu, K.T., Siregar, H., Grass, I., 2020. Environmental, economic, and social
Falconer, A., Mafira, T., Sutiyono, G., 2015. Improving Land Productivity through Fiscal consequences of the oil palm boom. Ann. Rev. Resour. Econ. 12 (1). https://doi.org/
Policy: Early Insights on Taxation in the Palm Oil Supply Chain. Climate Policy 10.1146/annurev-resource-110119-024922.
Initiative, Jakarta, Indonesia. Rist, L., Feintrenie, L., Levang, P., 2010. The livelihood impacts of oil palm: smallholders
FAOSTAT, 2020. Crops. Food and Agriculture Organization of the United Nations (FAO), in Indonesia. Biodivers. Conserv. 19 (4), 1009–1024. https://doi.org/10.1007/
Rome. s10531-010-9815-z.
Gatto, M., Wollni, M., Asnawi, R., Qaim, M., 2017. Oil palm boom, contract farming, and Santika, T., Wilson, K.A., Budiharta, S., Law, E.A., Poh, T.M., Ancrenaz, M., Struebig,
rural economic development: village-level evidence from Indonesia. World Dev. 95, M.J., Meijaard, E., 2019a. Does oil palm agriculture help alleviate poverty? A mul­
127–140. https://doi.org/10.1016/j.worlddev.2017.02.013. tidimensional counterfactual assessment of oil palm development in Indonesia. World
Grass, I., Kubitza, C., Krishna, V.V., Corre, M.D., Mußhoff, O., Pütz, P., Drescher, J., Dev. 120, 105–117. https://doi.org/10.1016/j.worlddev.2019.04.012.
Rembold, K., Ariyanti, E.S., Barnes, A.D., Brinkmann, N., Brose, U., Brümmer, B., Santika, T., Wilson, K.A., Meijaard, E., Budiharta, S., Law, E.E., Sabri, M., Struebig, M.,
Buchori, D., Daniel, R., Darras, K.F.A., Faust, H., Fehrmann, L., Hein, J., Hennings, N., Ancrenaz, M., Poh, T.-M., 2019b. Changing landscapes, livelihoods and village wel­
Hidayat, P., Hölscher, D., Jochum, M., Knohl, A., Kotowska, M.M., Krashevska, V., fare in the context of oil palm development. Land Use Policy 87, 104073. https://doi.
Kreft, H., Leuschner, C., Lobite, N.J.S., Panjaitan, R., Polle, A., Potapov, A.M., org/10.1016/j.landusepol.2019.104073.
Purnama, E., Qaim, M., Röll, A., Scheu, S., Schneider, D., Tjoa, A., Tscharntke, T., Sibhatu, K.T., 2019. Oil palm boom and farm household diets in the tropics. Front.
Veldkamp, E., Wollni, M., 2020. Trade-offs between multifunctionality and profit in Sustain. Food Syst. 3 (75). https://doi.org/10.3389/fsufs.2019.00075.
tropical smallholder landscapes. Nat. Commun. 11 (1), 1186. https://doi.org/10. Taheripour, F., Hertel, T.W., Ramankutty, N., 2019. Market-mediated responses confound
1038/s41467-020-15013-5. policies to limit deforestation from oil palm expansion in Malaysia and Indonesia.
Ickowitz, A., Rowland, D., Powell, B., Salim, M.A., Sunderland, T., 2016. Forests, trees, Proc. Natl. Acad. Sci. U. S. A. 116 (38), 19193–19199. https://doi.org/10.1073/pnas.
and micronutrient-rich food consumption in Indonesia. PLoS One 11 (5), e0154139. 1903476116.
https://doi.org/10.1371/journal.pone.0154139. Venter, O., Meijaard, E., Possingham, H., Dennis, R., Sheil, D., Wich, S., Hovani, L.,
Imbens, G.W., Angrist, J.D., 1994. Identification and estimation of local average treat­ Wilson, K., 2009. Carbon payments as a safeguard for threatened tropical mammals.
ment effects. Econometrica 62, 467–475. Conserv. Lett. 2 (3), 123–129. https://doi.org/10.1111/j.1755-263X.2009.00059.x.
Jelsma, I., Schoneveld, G.C., Zoomers, A., van Westen, A.C.M., 2017. Unpacking Yacob, S., 2019. Government, business and lobbyists: the politics of palm oil in US-
Indonesia’s independent oil palm smallholders: an actor-disaggregated approach to Malaysia relations. Int. Hist. Rev. 41 (4), 909–930. https://doi.org/10.1080/
identifying environmental and social performance challenges. Land Use Policy 69, 07075332.2018.1457556.
281–297. https://doi.org/10.1016/j.landusepol.2017.08.012. Yuliani, E.L., de Jong, E.B.P., Knippenberg, L., Bakara, D.O., Salim, M.A., Sunderland, T.,
Kabubo-Mariara, J., 2013. Forestpoverty nexus: exploring the contribution of forests to 2018. Keeping the land: indigenous communities’ struggle over land use and sus­
rural livelihoods in Kenya. Nat. Res. Forum 37, 177–188. tainable forest management in Kalimantan, Indonesia. Ecol. Soc. 23 (4). https://doi.
Kebschull, D., 2020. Transmigration in Indonesia. Routledge, London, New York. org/10.5751/ES-10640-230449.
Krishna, V., Euler, M., Siregar, H., Qaim, M., 2017. Differential livelihood impacts of oil Yuliani, E.L., de Groot, W.T., Knippenberg, L., Bakara, D.O., 2020. Forest or oil palm planta­
palm expansion in Indonesia. Agric. Econ. 48 (5), 639–653. https://doi.org/10.1111/ tion? Interpretation of local responses to the oil palm promises in Kalimantan, Indonesia.
agec.12363. Land Use Policy 96, 104616. https://doi.org/10.1016/j.landusepol.2020.104616.
Kubitza, C., Gehrke, E., 2018. Why Does a Labor-Saving Technology Decrease Fertility
Rates? Evidence from the Oil Palm Boom in Indonesia: Collaborative Research Centre