Professional Documents
Culture Documents
in Indonesia
Edited by:
Gumilang Aryo Sahadewo
Evi Noor Afifah
Eny Sulistyaningrum
Devanto S. Pratomo
Alin Halimatussadiah
Budy P. Resosudarmo
Penyunting :
Gumilang Aryo Sahadewo,
Evi Noor Afifah,
Eny Sulistyaningrum,
Devanto S. Pratomo,
Alin Halimatussadiah, and
Budy P. Resosudarmo
Desain sampul :
Arief A. Yusuf
Tata letak :
Deni Lesmana
Proofreader :
Lilies Achmadi
Diterbitkan oleh :
Perkumpulan Ilmu Regional Indonesia (IRSA)
Sekretariat IRSA, Jl. Cimandiri No. 6 Bandung 40115
West Java-Indonesia
Email : info@irsa.or.id
ISBN : 978-623-94282-4-2
Cetakan pertama, Juni 2022
CONTRIBUTORS.................................................................................................. v
FOREWORD........................................................................................................... x
Chapter 1 Institutional and Human Capital Aspects of Development in
Indonesia
Gumilang Aryo Sahadewo, Evi Noor Afifah, Eny Sulistyaningrum,
Devanto S. Pratomo, Alin Halimatussadiah, and Budy P.
Resosudarmo................................................................................ 1
PART I. INSTITUTION
Chapter 2 Productivity Impact of Political Dynamics in A Decentralised
Indonesia: The Case of the 2016 Centralisation Policy of Batam
Free Trade Zone*
Bangkit A. Wiryawan..........................................................................15
Chapter 3 Rent Seeking Undercover: Corruption and Public Spending
Composition in Indonesia
Jesita Wida Ajani, I Gede Sthitaprajna Virananda, and Ghifari
Ramadhan Firman................................................................................39
Chapter 4 Political Competition and Deforestation: Evidence from
Indonesia
Harnum Yulia Sari, Raka Rizky Fadilla, and Zihaul Abdi.................65
PART II. HUMAN CAPITAL
Chapter 5 The Role of Socioeconomic Factors in Voting Preferences in
Indonesia’s Local Electionsa
Dharra Mahardhika and Ekki Syamsulhakim......................................89
Chapter 6 Heritability of Cognitive Ability and Its Relation to Economic
Achievement in Adolescence
Siti Munawaroh and Ni Made Sukartini........................................111
iii
Chapter 7 The Effect of Educational Attainment on Female Labour Force
Participation in Indonesia
Rizka Isnaini Husna and Eny Sulistyaningrum...............................131
PART III. DEVELOPMENT
Chapter 8 Forestry Agrarian Reform in Balumpewa Village, Central
Sulawesi Province
Kasmiati, Mohamad Shohibuddin, Ganies Oktaviana, and Ahmad
Jaetuloh...............................................................................................151
Chapter 9 Elements of Maritime Logistics Transaction Costs in Islands
Province of North Maluku, Eastern Part of Indonesia
Chairullah Amin, Heti Mulyati, Eva Anggraini, Tridoyo Kusumastanto,
and D.S. Priyarsono...........................................................................177
Chapter 10 Institutional Access and Poverty of Capture Fisheries Households
in Indonesia
Silviana Taniu, Dyah Wulan Sari, and Dias Satria..........................201
Ahmad JAETULOH
Sajogyo Institute, Bogor, Indonesia
E-mail: Jaetulohahmad@gmail.com
Alin HALIMATUSSADIAH
Department of Economics, Faculty of Economics and Business
University of Indonesia, Depok, Indonesia
E-mail: alin.halimah@gmail.com
Bangkit A. WIRYAWAN
Department of Politics and Government
Diponegoro University, Semarang, Indonesia
E-mail: bwiryawan@lecturer.undip.ac.id
Budy P. RESOSUDARMO
Arndt-Corden Department of Economics
Australian National University, Canberra, Australia
E-mail: budy.resosudarmo@anu.edu.au
Chairullah AMIN
Department of Economics, Faculty of Economics and Business
Khairun University, Ternate, Indonesia
E-mail: chairulamin79@gmail.com
D.S. PRIYARSONO
Department of Economics, Faculty of Economics and Management
Bogor Agricultural University, Bogor, Indonesia
E-mail: priyarsono@apps.ipb.ac.id
v
Devanto S. PRATOMO
Department of Economics, Faculty of Economics and Business
Brawijaya University, Malang, Indonesia
E-mail: devanto@ub.ac.id
Dharra MAHARDIKA
Department of Economics, Faculty of Economics and Business,
Padjadjaran University, Bandung, Indonesia
E-mail: dharra.mahardhika@gmail.com
Dias SATRIA
Department of Economics, Faculty of Economics and Business
Brawijaya University, Malang, Indonesia
E-mail: dias.satria@ub.ac.id
Ekki SYAMSULHAKIM
Department of Economics, Faculty of Economics and Business
Padjadjaran University, Bandung, Indonesia
E-mail: ekki.syamsulhakim@fe.unpad.ac.id
Eny SULISTYANINGRUM
Department of Economics, Faculty of Economics and Business
Universitas Gadjah Mada, Yogyakarta, Indonesia
E-mail: eny@ugm.ac.id
Eva ANGGRAINI
Department of Resource and Environmental Economics
Bogor Agricultural University, Bogor, Indonesia
E-mail: evaanggraini@apps.ipb.ac.id
Ganies OKTAVIANA
Sajogyo Institute, Bogor, Indonesia
E-mail: ganies@sajogyo-institute.org
Heti MULYATI
Department of Management, Faculty of Economics and Management
IPB University, Bogor, Indonesia
E-mail: heti@apps.ipb.ac.id
vii
Jesita Wida AJANI
Department of Economics, Faculty of Economics and Business
Universita Gadjah Mada, Yogyakarta, Indonesia
E-mail: jesita.wida.ajani@mail.ugm.ac.id
KASMIATI
Faculty of Agriculture and Forestry
Universitas Sulawesi Barat, Majene, Indonesia
Email: kasmiatis35@gmail.com
Mohamad SHOHIBUDDIN
Faculty of Human Ecology
Bogor Agricultural University, Bogor, Indonesia
E-mail: m-shohib@apps.ipb.ac.id
Ni Made SUKARTINI
Department of Economics, Faculty of Economics and Business
Universitas Airlangga, Surabaya, Indonesia.
E-mail: ni-made-s@feb.unair.ac.id
Silviana TANIU
Department of Development Studies, Faculty of Economics
University of Gorontalo, Gorontalo, Indonesia
E-mail: silvianayunus@gmail.com
Tridoyo KUSUMASTANTO
Department of Resource and Environmental Economics
Bogor Agricultural University, Bogor, Indonesia
E-mail: prof.kusumastanto@gmail.com
Zihaul ABDI
Department of Economics, Faculty of Economics and Business
University of Indonesia, Depok, Indonesia
E-mail: abdizihaul@gmail.com
ix
FOREWORD
Regards,
Arief A. Yusuf
President, Indonesian Regional Science Association
xi
xii Institution, Human Capital and Development in Indonesia
Chapter 1
INTRODUCTION
Initial policies to supress the outbreaks of the COVID-19 pandemic were
mainly strict restriction of human movements and social distances (Baldwin
and Evenett, 2020). These policies, although have been compensated with
fiscal expansion policies, negatively affected the economies around the
world. The global economy was contracted in 2020. Vaccines, fortunately,
were invented approximately a year after the start of the pandemic and
were able to be widely distributed since then. This situation allowed many
governments to relax restriction on human movements and social distances.
The economy, hence, has been able to bounce back since 2021.
Such a situation also happened in Indonesia when the economy
experienced a contraction in 2020 (Lewis and Witoelar 2021; Resosudarmo
et al. 2021). Nevertheless, a year later, as vaccines have been relatively
widely distributed and government’s expansion fiscal policy have been fully
implemented, the economy bounced back (Halimatussadiah et al., 2022).
Figure 1 shows growths of gross domestic product (GDP) per capita, as a
welfare measure of the country, during several years before the COVID-19
till 2021. It can be seen that several years before the pandemic, the country’s
welfare has been growing with a rate of almost 5 percent. This performance
has been relative better than many developing countries. Furthermore,
1
Figure 1 also shows that welfare in all regions in Indonesia has been growing
as fast as the levels before the pandemic.
Figure 1. Growth GDP per Capita (%)
20
%
15
10 Sumatra
Java-Bali
Kalimantan
5
Sulawesi
Eastern Indonesia
0 Indonesia
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
-5
-10
INSTITUTION IN INDONESIA
The importance of institution quality of economic growth has been argued
in the literature for at least since 1990s. Examples would be the works by
Barro (1996), by Perotti (1996), and by Alesina et al. (1996). Since mid 2000s,
the most prominent scholars arguing this relationship are Acemoglu and
Robinson through many of their seminal works (Acemoglu and Robinson,
2008 and 2012; Acemoglu, et al., 2019). They show both theoretically and
empirically that better quality of institution would produce higher economic
growth in a country. This section, hence, will provide slight picture on the
quality of Institution in Indonesia.
In general, it can be confirmed that the quality of institutions in Indonesia
has been improving. However, it can also be argued that these institutions
have not yet having quality as good as those in developed countries. One
indicator of institutional quality that is publicly available is the World
Governance Indicators (WGI) produced by the World Bank. Figure 2
provides information on for several countries published by the World Bank.
WGI is an index consisting of several measures of institutional quality
in a country. These measures are (1) voice and accountability, (2) political
stability and absence of violence/terrorism, (3) government effectiveness,
(4) regulatory quality, (5) rule of law, and (6) control of corruption. Figure 2
shows that Indonesia’s WGIs from 2000 till 2020 have been improving. It was
-0.9 in 2000 and became 0.1 in in 2020, which is a major progress.
Australia
Brazil
Germany
India
Indonesia
Mexico
Netherlands
Philippines
United Kingdom
United States
Source : http://info.worldbank.org/governance/wgi/
Note that the World Governance Indicator consists of six dimensions of
governance, which are (1) voice and accountability, (2) political stability and
absence of violence/terrorism, (3) government effectiveness, (4) regulatory
quality, (5) rule of law and (6) control of corruption. The worse situation would
be -2.5 and the best would be 2.5.
Comparing to several developed countries, however, the quality of
institution in Indonesia is still relatively low. WGIs for Australia, Germany,
Netherlands, and UK are all above 1. Though institutional quality in India, as
another large democratic country in Asia, has been indicated to be declined
in the last 10 years or so, still it is better than Indonesia. The key question on
how Indonesia could improve the performance as its institutions is, hence,
a valid issue.
To better understand in what aspects of institutional quality that
Indonesia is lagging, we could compare Indonesia and Germany in each
aspect of governance indicators. Figure 3 shows these comparisons.
Control of Corruption
Rule of Law
Regulatory Quality
Government Effectiveness
Political Stability and Absence of
Violence/Terrorism
Voice and Accountability
Source: http://info.worldbank.org/governance/wgi/
As indicated in Figure 3, while gaps of all governance dimensions
between Indonesia and Germany have been steadily narrowing from
2000 to 2020, Indonesia has been progressing in all issues of governance—
particularly in terms of its institutional quality that has been improving.
Large progresses have been taking place with regard to the issues of political
stability, absence of violence/terrorism, and government effectiveness. The
two key issues—which Indonesia particularly needs to pay attention—are
control of corruption and rule of law. This fact indicates that corruption is
still rampant in Indonesia and enforcement of laws is relatively weak.
Since early 2000s, Indonesia has conducted activities to reduce
corruption in the country. In 2002, Indonesia set up a special anti-corruption
committee, known as Komisi Pemberantasan Korupsi (KPK). This committee
has so far managed to bring many important corruption cases to the court
that punished those involved in such crime. However, cases of corruption
have been argued not to significantly decline. One argument is that law
enforcements have not been that strong so far. Indonesia’s efforts so far
might not be sufficient to resolve the issues of corruption and enforcement
of laws. Better measures to combat corruption and to better enforce the laws
are needed. Should it be successful, Indonesia could expect having a higher
economic growth rate accordingly.
120
100
80
60
40
20
0
REFERENCE
Acemoglu, D. and J. Robinson. 2008. The Role of Institutions in Growth and
Development. Commission on Growth and Development Working
Paper No. 10. World Bank, Washington, DC. https://openknowledge.
worldbank.org/handle/10986/28045 License: CC BY 3.0 IGO.
Acemoglu, D. and J. Robinson. 2012. Why Nations Fail: Origins of Power,
Poverty and Prosperity. New York: Crown Publishers.
Acemoglu, D., S. Naidu, P. Restrepo, and J. Robinson. 2019. Democracy Does
Cause Growth. Journal of Political Economy, 127(1), pp. 47-100.
INTRODUCTION
During his presidential visit to Singapore in July 2015, President Joko Widodo
attempted to lure more investment to the neighbouring island of Batam, the
capital of Riau Island province. The island hosts Indonesia’s first successful
Free Trade Zone (FTZ). It showed stellar performance in the first half of the
1990s with the majority of investments coming from Singapore (Damuri et
al. 2015; Negara and Hutchinson 2020; Wulandari 2012) . In the meeting with
Singaporean business leaders during the visit, he laid out a plan to restructure
the FTZ governance in order to better improve its business climate. The
purpose of this effort is two folds. First, he tried to resolve the long-lasting
conflict between the central and local governments in the zone which was
detrimental towards industrialisation attempt in the region (Jati 2016).
Second, considering the regional context the measure was seemingly done to
anticipate a major upcoming change, i.e., the implementation of the ASEAN
Economic Community (AEC) at the end of that year. AEC could potentially
lessen the policy incentive provided in the zone. The utilisation of placed-
based policies such as the FTZ to boost industrial output has been a growing
practice in many parts of the world (Farole & Akinci 2011). Its instrumentation
in the Asian Tiger economies, and particularly in promoting export-oriented
* This manuscript is part of author’s dissertation research and the findings have been presented at
the JASID Tokai Conference for Young Researcher, Nagoya 21 August 2021.
15
activities, has managed to sustain unprecedented high growth spanning for
more than two decades. The remarkable success of China’s industrialisation
from the 1980s to the 2010s was also largely contributed by their location-
based policy. China has been successfully harnessed foreign investment by
establishing Special Economic Zones in Shenzhen, Zhuhai, Shantou, and
Xiamen in the early days of the open-door policy. The success has led to
the larger experimentation of this approach, leading to the open coastal and
open cities policy between 1988 and 1992 (Ge 1999).
This chapter aims to estimate the impact of vertical institutional change
in the management of Batam FTZ on firms’ labour productivity. One of the
latest changes, following the president’s aforementioned visit to Singapore,
was started in 2015 where the central government attempted to revitalise
the zone through centralising its administration. This move was finally
formalised in February 2016 with the Presidential Decree (Keppres) no. 8.
Source: databasin.org
The initiative to make Batam a free trade zone began to emerge in October
1971 with Presidential Decree No. 74. It was then followed by another decree
in 1973 that assigned Batam Island Industrial Area Development Authority
(BIDA) as the agency responsible for the zone’s development (BP Batam
2021). This placed the zone under Jakarta’s direct influence with BIDA serves
as de facto government of the island. This is in contrast to most other areas in
Indonesia where regional and local governments are the ones responsible for
the development in their respected jurisdiction, even when they were merely
an extended arm of the central government. President Suharto’s Minister of
Research and Technology, BJ Habibie, was appointed as the head of BIDA,
in the hope to propagate not just industry but also promoting industrial
upgrading (Hutchinson 2017). However, for nearly two decades Batam still
encountered difficulties in attracting investments, which soon had changed
with the establishment of a growth triangle plan better known as SIJORI.
In 1989, the Indonesian government was one part of the tripartite SIJORI
Growth Triangle plan. SIJORI stands for Singapore, Johor, and Riau, where
the latter two regions are the state/province of Malaysia and Indonesia,
respectively. The triangle comprised a development plan with Singapore
at the center and both Johor and Riau as its periphery areas. The plan was
announced by then Deputy Prime Minister of Singapore Goh Cok Tong and
was responded with much enthusiasm by all parties involved.
Following the announcement of the SIJORI plan, massive investment
flocked into Batam FTZ. Between 1991 and 1997 total foreign investment was
In the year after, the central government in Jakarta issued the long-
awaited FTZ Law (44/2007) which was accompanied by a presidential
decree (Perpres) effective in 2008. Under the new regulations, both Governor
of Riau Island and the Mayor of Batam served as the chair and vice-chair of
the board, respectively (see Figure 3). The membership of the board under
the regulation has also consisted of local officials. The dynamics had swung
the zone’s political pendulum from a centralised to a decentralised one.
The revitalisation plan dramatically increased FDI flow by 45% whilst
also sustaining GDP growth at around 7% between 2008 and 2013 as shown
in Figure 2. It was only interrupted in 2009 due to the global financial crisis.
These numbers suggest that the decentralised policy works, accommodating
local government’s role in the FTZ administration helped in improving
climate change which eventually results in higher investment and growth.
However, as we can see in Figure 2, nearing the implementation of AEC in
2015 growth begins to falter.
Data Description
The main data source was taken from Statistics of Indonesia’s Large and
Medium Industrial survey that is conducted yearly. However, there are some
gap years in accessing the data, leaving with only odd years data selected.
In choosing the data to be used, first the identification of the recent policy
change in Batam’s FTZ and its parent province Riau Islands is conducted.
From there the data in between its implementation were selected, since
the targeted policy intervention was the 2016 FTZ centralisation policy. It
was then followed by selecting the industrial data of 2015 and 2017 for the
estimation.
Independent var.
Labor cost (log current IDR 1394 3.22 0.50 1126 3.14 0.48 268 3.51 0.49
mil.)
Mac. purchase (log current 1394 1.58 2.90 1126 0.84 2.13 268 4.36 3.86
IDR mil.)
Material cost (log current 1394 10.01 3.27 1126 10.42 2.68 268 8.29 4.68
IDR mil.)
Foreign ownership (%) 1394 19.59 38.45 1126 9.87 28.20 268 60.47 47.83
Independent var.
Labor cost (log current IDR 1392 3.42 0.52 1125 3.34 0.49 267 3.75 0.48
mil.)
Mac. purchase (log current 1392 2.39 3.06 1125 2.03 2.87 267 3.92 3.33
IDR mil.)
Material cost (log current 1392 10.26 3.09 1125 10.62 2.66 267 8.75 4.15
IDR mil.)
Foreign ownership (%) 1392 19.32 38.43 1125 9.18 27.45 267 62.03 47.65
Source : Author, calculated based on the SI (Industrial Survey) data of 2015 and 2017
Estimation Strategy
The treated region (treat) is a binary variable mark with 1. The period before
the policy came into effect (post) is marked as 0, and 1 is given for the period
following the policy implementation which is 2016. The initial equation set-
up in predicting the impact of policy on productivity is written as follows:
… (1)
resembles output or per worker productivity change in firm i at the
year t. All the beta coefficients ( ) are unknown parameters. Hence, the
parameter of interest is , which resembles the DID coefficient. Menahwile,
is the error term, expected to be uncorrelated with the other variables.
When the expected condition is met, then it can be confirmed that the parallel
trend assumption is fulfilled (Albouy 2004).
Using the OLS, the model specification is formulated as follows:
… (2)
X is a vector of endogenous variables that includes (1) unit labor cost,
(2) machinery purchase, (3) raw material cost, and (4) foreign ownership
(% share); whereas and are region-specific and individual firm fixed
effects, respectively, which control unobservable heterogeneity that might
occur due to culture, language, and supply-chain network. Lastly, is an
error term not correlated to the main independent variable.
A. Output B. Productivity
Note : Batam FTZ is set as the treatment group while the neighboring provinces (North
Sumatera, Riau, South Sumatera, and Lampung) are set as the control group.
Vertical thick line denotes beginning of policy change; the centralization of FTZ
in February 2016. Dashed black line marks the counterfactual line should the
policy was not applied.
Source : Author’s own calculation based on the SI (Industrial Survey) data of 2013, 2015,
2017.
The parallel trend graph was confirmed by testing it using lead (post)
and lag (pre) data. Using the firm output as the dependent variable does
not allow it to pass the lead-trend test, but it does manage to pass the time-
trend one (see Table 2, Column 1). The time-trend result does not suffice the
DID assumption that treatment coefficient is zero in the pre-period (Pischke
2005). Our main variable of interest, productivity per worker, managed to
pass both the lead-trend as well as the time-trend tests (see Table 2, Column
3). The F-test score probability indicates whether the variables pass or do not
pass the test.
Baseline Result
Table 3 presents the baseline result, without controlling for endogenous
variables. Columns (1) and (2) describe the result using output as a dependent
variable. First, it can be seen that the DID coefficients are significantly
negative, either calculated using the standard OLS or with Fixed-Effects
OLS. Second, using the main dependent variable, productivity, in Columns
(3) and (4), the coefficients are consistently negative and significant for both
the output and productivity, suggesting that the result could be robust.
Main Result
Let us continue the baseline result above by adding endogenous control
variables necessary to explain productivity. Table 4 shows the main estimation
results. Column (1) shows that the negative coefficient for firm output
stays (-0.368) after controlling for unit labor cost, machinery purchase, raw
materials purchase, and foreign ownership share. In terms of productivity,
the same trend can be observed as shown in Column (5). The coefficient
has smaller magnitude (-0.279) compared to previous specification, but the
negative trend remains.
Unit labor 0.307*** 0.340*** 0.250** 0.338*** 0.324*** 0.362*** 0.211* 0.285**
cost (0.059) (0.080) (0.122) (0.129) (0.058) (0.075) (0.110) (0.117)
Raw material 0.168*** 0.144*** 0.097*** 0.098*** 0.159*** 0.137*** 0.102*** 0.107***
(0.021) (0.023) (0.023) (0.023) (0.018) (0.019) (0.021) (0.021)
Foreign own. 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.003**
(%) (0.001) (0.002) (0.002) (0.003) (0.001) (0.002) (0.001) (0.001)
Extension Analysis
For all the estimations presented in this sub-section, the same control variables
used in Table 4 (Column 3) in the previous section are applied. However, this
part is focused on firm productivity, which is the main dependent variable. To
begin with, the extension analysis is started by dividing firms based on their
size. The size definition follows national standards where large firm status is
assigned for a firm with 100 or more workers. Meanwhile, the medium-size
firm is given to a firm with 20 to 99 workers. Table 5 presents the estimation
result based on firm size with additional legal ownership status.
Discussion
Through multiple exercises in the previous section, we have observed that
there is a convincingly negative productivity impact of the 2016 centralisation
policy of Batam FTZ. From the main result, the -0.279 coefficient magnitude
corresponds to around 32% drop in productivity. The same negative result
(45% drop) is also observed for firm growth. This section will be focused on
(1) the issue of central-local relation in the currently decentralised Indonesia,
and (2) possible determinants of industrial resilience against policy change.
The central government, through its affiliated agency, has been the
dominant agent in the development and industrialisation of Batam FTZ
during the new order era (1967-1998). The centralisation strategy worked
well given the centralised nature of the institution at that time. However,
under the currently contrasting political landscape since decentralisation
reform in 2001, applying a centralised approach is expected to bring a
counter productive consequence as shown in Table 3. Hence, the finding is
similar to the previous studies (Aritenang 2017; Aritenang and Chandramidi
2020; Negara and Hutchinson 2020) that failed to find positive impact of
political and policy change in the Batam FTZ. Poor institutions that come
with the policy change have also been pointed out to contribute to the
industrial decline in the case of electronic sector in the island (van Grunsven
and Hutchinson 2017).
Despite its ambitious attempt to strengthen national coordination, the 2016
policy was still not able to resolve the dual authority issue, if not rather
CONCLUSION
After it was perceived to fail to perform well under the local government
leadership, president Jokowi attempted to recentralise Batam FTZ
administration in 2016 by placing it under the Coordinating Minister of
Economic Affairs. However, this does not resolve the still existing dual
authorities issue and rather increasing the tension between the central and
local government.
As mentioned previously, this research aims to investigate the impact
of that policy change on firms’ productivity. The difference-in-difference
estimation strategy is employed to exploit the impact, using Batam’s
neighbouring provinces as the control group. Prior to such analysis, the data
passing the parallel trend test for growth (passing the time trend test) and
productivity (passing both lead and time trend tests) were set. Main results
suggest that the policy has caused the output growth to decline by 45% and
32% drop in terms of productivity growth. The impact was serious that led
the government to take drastic action to revert the policy decision by 2019.
There are two important things that stand out from this study. First,
firms’ choice of location can determine their severity against policy change.
Those located in the Industrial Park are not necessarily affected by the
negative impact of the policy. Second, foreign and export oriented firms are
more resilient to the change. The result of the analysis has explained that
those types of firms are able to record positive growth relative to the control
group in the period after the policy implementation. Regardless of future
policy options decided by the government, maintaining the operations of
these firms is important to prevent deeper productivity loss.
INTRODUCTION
Public officials committing corruption might skew spending towards
categories of activities that provide the most lucrative opportunities to extract
rents (Tanzi and Davoodi 1997; Gupta, deMello, and Sharan 2001; Hessami
2010). Studies by Mauro (1998), Delavallade (2006) and Baraldi (2008), for
example, found that corruption decreases public spending particularly on
education. Corruption might also be committed by marking up spending
predominantly in sectors such as public construction and public health. These
deemed most corruption-prone sectors are driven by the aforementioned
behaviour of the bureaucrats, increasing their shares in the government’s
fiscal spending, which may result in a decrease in other sectors allocation
and in turn in the welfare loss on the part of the whole society.
This chapter aims to investigate whether the rent-seeking activities
are prominent in the local public spending composition. To this end, it
is focused on the sub-national level data because there are variances and
heterogeneities that much more happened at the local level rather than at
the aggregate national level. To accommodate our research intentions and
design, we use a novel dataset of corruption cases in Indonesia from “Cegah
Korupsi” set up by the Faculty of Economics and Business, Universitas
Gadjah Mada. The data comprises individual level corruption history in
39
which we aggregate it into provincial level. We also combined it with the
data from the Central Bureau of Statistics (BPS) of Indonesia to gather other
complementary covariates.
We ran regressions to find the relationship between corruption cases
and public spending allocation. We divided the public spending into
several major categories: public service, health, social protection programs,
economic programs, education and housing. Our main independent variable
is corruption cases per capita, which is the number of corruption cases per
province divided by the number of population. To control heterogeneity
across provinces, we employ fixed effect regressions.
Based on the data provided by Cegah Korupsi, Papua and South
Kalimantan rank the first, as they occupy the highest level of corruption cases
per capita by far (Figure 1). Meanwhile, Banten, West Jawa, Central Jawa and
East Jawa, all have the lowest corruption cases per capita, clearly affected
by their substantially large population sizes as denominators. Finally, our
fixed-effects regression finds three spending categories that are significantly
associated with corruption per capita, namely social protection, health, and
housing. The sign is positive for the first two while negative for housing.
This result could imply that social protection and health present lucrative
rent-seeking opportunities, affirming the urgency to reform institutions in
these spending areas.
Figure 1. Mean of Corruption per Capita and Population from 2007-2016
DATA
The main data we used in this chapter comes from corruption database
named Cegah Korupsi-5 (5th wave) produced by Faculty of Economics and
Business Universitas Gadjah Mada. The data originated from a web scraping
process based on the supreme court website. It comprises individual level
corruption history, such as personal data, the charges, verdict in every level
of judgeship, loss caused by corruption and the name of judges involved.
The data supposedly covers all the corruption cases that have already taken
place from 2000 to 2019, since our data were collected from the verdict
documents at district, provincial and supreme court level.
has not been easily detected. Resosudarmo and Subiman (2006), for example, could not find the
correlation between firm bribery behaviour and the use of illegal logging among wood-based
industries in Indonesia.
Nonetheless, there are several problems with the data. First, not all
cases are recorded on the website of the supreme court. Figure 2 shows how
the recorded cases are normally distributed. We suspect that for early years,
the number is relatively small because there were not many corruption cases
exposed since the corruption eradication commission was formed only
recently. For recent years, most cases that are still ongoing. These cases do
not have any verdict documents, hence not recorded in the dataset. Notably
there are some heterogeneities across courts in Indonesia.
As shown in Table 1, the differences are not that high at glance. A
simple OLS regression is conducted—as presented in Table 2—to figure
out whether there is an association between province and corruption cases.
Here, Java island was set as the base variable, of which the results show that
on average the number of corruption cases is heterogenous across provinces
and statistically different with Java. This might be caused by endogenous
trait of the corruption itself where it is more prominent in region with
more economic activity, or maybe because of the limitation of corruption
eradication commission in exposing cases on some regions. Unfortunately,
no further evidence was found to answer this question.
METHODOLOGY
In general, our regression model is constructed as follows:
Robustness Check
To check whether the results are robust, particularly not suffering from a
Heterogeneity Analysis
Table 6 illustrates heterogeneities across province in Indonesia. For provinces
in Java, housing has a negative association and positive relationship for
public service and social protection on corruption cases per capita. For Bali-
NT and Maluku-Papua, the negative and positive association was shown
by housing and public service spending, respectively. Meanwhile, different
cases take place for Kalimantan where education has positive relationship
with corruption per capita, and Sulawesi in the case of health. Interestingly,
no evidence on the relationship between corruption case and the economic
allocation of regional budget allocation can be found.
REFERENCES
Ades, A. and R. Di Tella. 1994. “Competition and Corruption. Institute of
Economics and Statistics.” Discussion Papers 169, University of Oxford.
Baraldi, A., 2008. Effects of electoral rules, political competition and corruption
on the size and composition of government consumption spending: An
Italian regional analysis. The BE Journal of Economic Analysis & Policy, 8:
1–37.
Barro, R. 1990. Government spending in a simple model of endogenous
growth. Journal of Political Economy, 98 (5): S103–S125.
Becker, G., 1983. A theory of competition among pressure groups for political
INTRODUCTION
It has been argued that the root cause of environmental degradation in
Indonesia is a combination of weak institutions and poor governance.
Dearth of transparency, lack of accountability, and limited participation
of stakeholders in formulating and implementing forest land management
policies have contributed to the continued deforestation (Bisson et al. 2003;
Cisneros et al. 2020; Smith et al. 2003). As a result, over the last two decades,
Indonesia has lost 9.75 million hectares of its humid primary forest, making
up 36% of its total tree cover loss in the same period (Global Forest Watch
2021).
Decentralisation has been blamed as one of the major institutional factors
driving local capture that trigger deforestation in Indonesia (Adrison 2014).
Decentralisation in Indonesia gave local government discretionary power
over the supply of logging permits, which has led to excessive logging and
bribe-taking (Smith et al. 2003). Furthermore, the direct mayoral elections
did not yield an increase in political accountability and resulted in even
more deforestation (Burgess et al. 2012), possibly by changing the structure
of political rent extraction, with fewer candidates running for office and
larger governing coalitions exerting their influence.
However, theoretically, Indonesia can still benefit from the
decentralisation system through the check and balances mechanism between
the executive and legislative branches at the local level. This ideal version
of decentralisation requires a strong local parliament—characterised by
65
high competition among political parties—that is expected to hold the local
leaders accountable for their policy choice, leading to better governance and
ultimately reducing deforestation.
A body of research has found that political competition plays a key
role in influencing socio-economic performance. A high degree of political
competition encourages efficiency in public goods provision and dissuades
rent-seeking behaviours. Political competition and the checks and balances
imposed in a well-functioning democracy restrict the ability of governments
to engage in rent-seeking behaviour (Olson 2000; North 1990; North and
Weingast 1989). Political competition has built-in mechanisms for increasing
the probability of detection and punishment due to the competition among
parliament members of the ruling party and opposition parties. Studies
dealing with environmental policies also showed that political competition
raises the stringency of environmental policies (Fredriksson et al. 2005). That
being said, it is important to investigate to what extent does the competition
among political parties at the local level influences the environmental
outcomes, specifically around forest management.
In Indonesia, a higher degree of political competition appears to
improve the economic outcome (Rezki 2020) and local services delivery
(Skoufias et al. 2014) through pro-business policies and better expenditure
choices, respectively. While evidence highlights the link between local
political incentives and deforestation in Indonesia (see Cisneros et al. 2021),
there is no empirical evidence to date linking local political competition and
deforestation.
Our research, therefore, attempts to explore the association between
political competition and the rate of deforestation. We hypothesise that a
higher degree of political competition within the parliament, hence more
accountable government, is associated with a decreasing rate of deforestation.
This chapter offers a novel measurement of political competition by
simultaneously employing fractionalisation and polarisation index. The
Fractionalisation index measures the degree of political party diversity in a
parliament. While this measurement has been widely used in the previous
literature to illustrate the degree of political party fragmentation (Wang
2014), we argue that such an index is insufficient to measure the likelihood
of conflicting ideas in the policymaking process. We, therefore, complement
the measurement with the polarisation index.
Using subnational-level data of Indonesia from 2009 to 2018, we
EMPIRICAL STRATEGY
The focus of our research is closely linked with the empirical literature
exploring the connection between institutional quality and environmental
outcomes. Previous studies usually use binary variable contrasting
democracy and autocracy or employ a sort of index to represent the degree
of democracy, civil liberty, and political freedom (Farzin & Bond 2006;
Bhattarai & Hamming 2001; Torras & Boyce 1998). Other studies use a range
of proxies to measure the quality of the institution, such as the strength of
the environmental NGOs (Binder & Neumayer 2005) or electoral system
(Fredrikkson & Millimet 2004). As for the environmental outcomes, past
studies mainly use deforestation, air quality, or carbon emissions as the
representative indicators.
A more recent literature that have the most similarity to ours are
Morjaria (2011) and Su, et al. (2020). Morjaria (2011) exploited the timing
of introduction of multiparty system as the instrument to measure the
impact of political competition on deforestation in Kenya. Su, et al. (2020)
leveraged the party fragmentation index and margin of victory between the
first and second place party to test the relationship between political party’s
competition and air quality in Mexican municipalities. A set of control
variables were employed in those studies, including population, GDP per
capita, density of economic activity, and district-specific fixed-effects.
In this study, fixed effects regression panel data are employed to
capture the relationship between political competition and deforestation at
the district level. This approach is used to capture the time-invariant that
occurs in each district. As GFW data continue to improve methods and
technology to measure the deforestation every year, year fixed-effects are
therefore also employed to measure the effect of methodological difference.
This methodological difference is due to adjustments made by GFW, such
as algorithm adjustments, updated satellite data, and variations in the
availability of satellite imagery1. Our analysis uses the following specification:
1 https://www.globalforestwatch.org/blog/id/data-and-research/menilai-tren-hilangnya-
tutupan-pohon-dalam-kurun-waktu-data-20-tahun-2/
DATA
The analysis was conducted using an unbalanced panel dataset. The number
of districts/cities in our sample is 241 out of a total of 514 districts/cities. We
dropped districts/cities with no primary forest. We addressed the ongoing
process of district proliferation by refining our district framework to the
end of the observation period, building a panel of geographic entities as
observed in 2009 to 2018. This study covers ten years ranging from 2009 to
2018, covering two election cycles. Table 1 provides the summary statistics
of the data used in this study.
Table 1. Descriptive Statistics
Variables Obs Mean Std. Dev. Min Max
Deforestation 1321 2463.623 5755.156 .076 59816.148
Political party 1321 .87 .058 .53125 .9376
fractionalization
Political party polarization 1321 .93 .045 .9259259 .9981689
Political party ideology 1321 .447 .112 0 .6479012
fractionalization
Political party ideology 1321 .954 .121 0 1
polarization
Audit BPK 1321 4.215 1.004 1 5
Price oil exposure 1321 376904.41 458797.5 2 2152481
Environmental 1321 1.260e+10 3.066e+10 0 5.734e+11
expenditure
Natural Resource Revenue 1321 9.828e+10 2.385e+11 2.815e+08 3.622e+12
Sharing
Poverty Rate 1321 10.996 5.906 2.009 45.709
Regional GDP 1321 .183 .544 .009 9.494
Temperature 1321 25.307 1.868 13.17 28.98
Wind Speed 1321 1.138 1.013 .04 4.61
Cloud Amount 1321 81.112 7.32 50.56 91.08
Political Competition
This chapter uses two measures to represent the level of political competition
in the local parliament by adopting the index proposed by Montalvo and
Reynal-Querol (2005) with the value ranging between 0 and 1, namely
fractionalisation and polarisation index:
where FI is the fractionalisation index. PI is the polarisation index. Si,j is
political party’s seat share in the parliament for district i and year j.
𝐹𝐹𝐹𝐹𝑖𝑖,𝑗𝑗 = 1 − ∑𝑁𝑁
𝑖𝑖=1
2
𝑠𝑠𝑖𝑖𝑖𝑖 (1)
0.5−𝑠𝑠𝑖𝑖𝑖𝑖 2
𝑃𝑃𝑃𝑃𝑖𝑖,𝑗𝑗 = 1 − ∑𝑁𝑁
𝑖𝑖=1 ( ) 𝑠𝑠𝑖𝑖𝑖𝑖 (2)
0.5
Control Variables
To minimise the issue of omitted variable bias, we use several control variables
2 Global Forest Watch is the database that provides data and tools for monitoring forests. This data
is retrieved from https://www.globalforestwatch.org/dashboards/
RESULTS
Empirical Results
The main result from the empirical analysis is provided in Table 2, where
Columns (1-6) show the relationship between political competition
represented by political party fractionalisation and political party
polarisation with deforestation rates. The results of this regression strengthen
the descriptive correlation found in Appendix 4. We found a different
relationship between political party fractionalisation and political party
polarisation with deforestation. A significant and negative relationship is
seen in the political party fractionalisation. An increase in the fractionalisation
index is associated with a slower rate of deforestation. On the other hand, a
positive relationship is observed in political party polarisation, meaning that
the deforestation rate increases along with a higher degree of political party
3 We recognised that government of Indonesia has been imposing moratorium of forest conversion
for oil palm plantation since 2011. However, there are still limitations to its enforcement. This was
revealed in the findings of Gaveau et al. (2021) which showed that based on satellite monitoring,
forest conversion for agricultural land still occurred even after the moratorium despite a declining
trend in recent years. Therefore, price exposure remains a valid control in this case.
REFERENCES
Adrison, V. 2013. “Deforestation in Decentralised and Democratic Indonesia.”
Working Paper No. 01. Institute for Economic and Social Research,
Faculty of Economics Universitas Indonesia (LPEM-FEUI). <https://
www.lpem.org/deforestation-in-decentralized-and-democratic-
indonesia/>
Arnold, L. L. 2008. Deforestation in decentralised Indonesia: What’s law got
to do with it. Law Env’t & Dev. Journal, 4: 75.
Aytimur, R. E. 2011. “Coalition Governments and Political Rents.” Working
INTRODUCTION
It has been argued that voters lean toward picking candidates who have
similar identities with they are than those who have different identities.
For example, in the UK, Heath (2013) found that working class voters are
suspected to be more likely to vote for working class candidates. Nida and
Fahmi (2017), from US survey data in 2012, identified that employees tend to
choose candidates from the same employment level and those voters tend to
choose the same race. From their research in India in 2012, Heath, Verniers, &
Kumar (2015) found that Moslem voters tend to choose Moslem candidates.
On candidate‘s religion, Tilley (2014) explained three things that matter
why religion can influence a person’s choice. First, dominant explanation of
voting behaviour is correlated with religion. Second, religion is correlated
with a particular set of values or at least like policy offerings; and third,
past religious conflicts (history) could shape voting behaviour. Candidates’
religion could influence voter’s decisions directly—i.e., voters choose
candidates based on their beliefs—or indirectly—i.e., voters can conclude
the attribute and belief from candidates’ religion. (Campbell, Green and
Layman, 2011). Only a few research, however, is available describing the
socioeconomic characteristics for such behaviour.
89
As in the case of Indonesia, candidates’ identity is considered as an
important factor in gaining the election. Particularly important are their
religion and gender. Based on a survey data conducted by the Centre for
Strategic and International Studies/CSIS (2017), 58.4% of Indonesians could
hardly accept leaders of different religions. In addition, an Asia Foundation
survey in 2013 found that 38.6% of candidate’s religion becomes an important
factor to elect the Governor, Regent, and Mayor (Strachan 2014). Other
studies elsewhere also find similar results. Ramadlan & Masykuri (2018),
for instance, found similar results in which 62% of the sample community
in their study in Boolang Mongondow Regency, North Sulawesi Province,
Indonesia in 2006 believe that having leaders with the same religion (Islam)
is of importance. Therefore, the role of religion in the election of local heads
is still taken as a determinant factor.
It is also important to look further the reasons why voters’ socio-
economic backgrounds determine their political options. First, since 2007,
Indonesia has conducted direct and concurrent local government head
election (Pilkada langsung dan serentak) as an important instrument
to establish the democratic local government (Martinus, 2013; Suyatno,
2016). The most crucial thing of Pilkada langsung dan serentak is voters’
participation because it can be a reflection of democracy (Akbar, 2017).
Nevertheless, according to Putra (2017), the low participation in Pilkada is
one of the common problems. This confirms the finding by Rochmi (2018)
that the participation rate in pilkada 2018 decreased compared with the 2017
elections, which will have an impact on the success of a democratic country.
An argument for this low and declining participation in elections is that
voters could not find suitable candidates with characteristics as they expect,
particularly related to candidates’ religion and gender.
Second, during the election campaign period, religious issues usually are
typically raised with respect to choosing the “right” regional heads. Indonesia
consists of many different individual and community backgrounds, raising
religious issues during regional elections are very sensitive and prone to
conflics. The candidates’ gender that now becomes one important unwritten
requirement for regional heads is another factor emerging in electing the
“proper” leaders, resulting in a larger number of men becoming regional
leaders than women. Marwah (2016), fore example, mentioned that only 15
female candidate pairs were selected from a total of 56 pairs in simultaneous
regent election in Central Java in 2015.
DATA
The data used in this chapter—i.e., the Indonesian Family life Survey (IFLS)—
are based on a longitudinal survey focused on collecting information on
health and socio-economic indicators of selected household and individuals
living in 13 out of 27 provinces in Indonesia. The selected 13 provinces are
North Sumatra, West Sumatra, South Sumatra, Lampung, DKI Jakarta, West
Java, Central Java, Yogyakarta, East Java, Bali, West Nusa Tenggara, South
Kalimantan, and South Sulawesi (Strauss, Witoelar and Sikoki 2016). The
main reason for using IFLS, dataset instead of other available datasets, is
that information on whether voters consider candidates’ religion and gender
METHODS
We modify the model used by Putra (2017) as follows:
(1)
This chapter uses 2 models where the dependent variable (y) is a dummy
variable individual preferences when voting local head elections, i.e.,
Consider Religion (Model 1), and Consider Gender (Model 2). Independent
variable is divided into 2 groups, i.e., individual characteristics (Ikit) and
community characteristics (Cjit). Also, this model includes a fixed effect to
control some macro shocks, i.e, year fixed effect (Dt) and island fixed effect
(Smd). While individual characteristics consist of education, sex, religion,
age, and marital status, education is grouped into 5 level, no schooling
until university. Community characteristics consist of household per capita
expenditure, living in urban, near market, terminal, living near provinces or
district centre, and high belief in the same religion.
We estimate the model using pooled-logistic-regression to measure
the influence of socioeconomic characteristics on the probability of a person
having voting preferences by considering religion and gender. Our estimate
is expected to show whether these socioeconomic characteristics associated
with their voting behaviour. The analysis uses various specifications to check
whether the results are robust. As our unit of analysis is at the individual
level, we use restricted-unrestricted mechanism or plug-and-play some
of the community characteristics from the models and see the stability
of the estimated marginal effect of the individual-level characteristics
variables (Appendix 1 and Appendix 2). In addition, we also provide some
heterogeneity analysis to observe if there are differences in the estimates due
RESULTS
Results of marginal effect from Table 2 shows that the level of voter’s
education affects the probability of voter’s religion consideration in choosing
a candidate for local head elections. The pooled logistic results show that
voters who have the highest education in Junior, Senior, and University
tend to have less probability each of 5.15, 7.61, and 7.55 p.p. (percentage
point) in considering candidate’s religion for local head than voters who did
not attend school. This negative effect means that individuals do not have
tendency to choose a leader based on the same religion status as theirs (Nida
and Fahmi 2017).
HETEROGENEITY ANALYSIS
This chapter also present several heterogeneities analyses that are divided
into 3 sections: (1) By sex, between male and female voters, covering samples
of 17,465 female voters and 20,041 male voters; (2) by Island, which compares
voters from Java Island and Off Java Island, the Java Island including DKI
Jakarta, West Java, Central Java, Special Region of Yogyakarta, and East
Java. Outer Java Island are Bali, West Nusa Tenggara, South Kalimantan,
South Sulawesi, North Sumatra, West Sumatra, South Sumatra, and
Lampung; and, (3) by income group, which classifies voters as Poor, when
Per Capita Expenditure is less than poverty line and Not Poor when Per
Capita Expenditure is more than poverty line.
Table 3 shows more results of male and female voters having the
probability of using religion and gender consideration in choosing a
candidate for local head elections. In general, both male and female voters
have a similar preference in terms of religion consideration when choosing
CONCLUSION
Voters mainly consider two factors—i.e., religion and gender—in choosing
local head candidates. As stated earlier, this chapter aims to investigate and
measure the influence of socioeconomic characteristics on the probability of
a person having preferences based on religion and gender considerations
in voting for a candidates in local head elections in Indonesia. This chapter
employs IFLS data of 2007 and 2014 and pooled-logistic model to measure
the influence of individual/community characteristics in considering
religion and gender.
The results found that socioeconomic factors influence voters’ preference
in considering candidates’ religion and gender in local head elections,
especially in 13 selected provinces. The finding also shows that voters’
level of education and those who live in urban have negative significance
to considering candidates’ religion and gender. Moreover, Moslem-female
voters who live in district and province centre, and community who have
“high level of belief” in the same religion tend to consider candidates’
religion and gender in local head elections. It can be concluded that a voter
who has the same social characteristics will have a higher probability of
choosing candidates with the same social characteristics as theirs.
REFERENCES
Akbar, I. 2017. Pilkada Serentak Dan Geliat Dinamika Politik Dan
Pemerintahan Lokal Indonesia. CosmoGov, 2 (1): 95. doi: 10.24198/
cosmogov.v2i1.11852.
Bhue, G. S. et al. 2014. Does Education Affect Voting Preferences? Evidence
from Voting Records in Indian Elections. SSRN Electronic Journal. doi:
10.2139/ssrn.2396784.
Botterman, S. and M. Hooghe. 2012. Religion and voting behaviour in
Belgium: An analysis of the relation between religious beliefs and
Christian Democratic voting. Acta Politica. Nature Publishing Group,
47(1), pp. 1–17. doi: 10.1057/ap.2011.11.
Campbell, D. E., Green, J. C. and Layman, G. C. 2011. The party faithful:
Partisan images, candidate religion, and the electoral impact of party
INTRODUCTION
Cognitive ability is one of the components to increase individual capacity
and essentially plays a role in child development and their future life (Zahra
and Wisana 2019). According to Glewwe et al. (2017) and Cawley et al.
(2001), cognitive ability is also an indicator of individual achievement in
adult life, such as educational attainment, career, and wage outcomes in the
labour market. In their study, Acosta et al. (2020) reported that individuals
with higher cognitive abilities tend to have better jobs and higher incomes.
This finding implies that childhood cognitive ability could be as predictor
of individual for success in their later life, especially in terms of education
achievement and labour outcome. Therefore, investment on cognitive ability
should be prioritised to improve the development outcome.
Hewi and Saleh (2020) stated that the Programme for International
Student Assessment (PISA) measured in reading, mathematics, and science
can capture students’ cognitive abilities. These cognitive abilities can be seen
in the way students solve problems in each question area and the scores. Based
on international cognitive measurements through PISA, children’s cognitive
scores in Indonesia are still relatively low (Figure 1). Indonesian students’
abilities are still below the average international standard. Evidence can be
investigated in the 2018 PISA score report in which Indonesia ranked 73 out
of 79 countries (OECD 2019). Indonesia’s performance among other Asian
111
countries in literacy, mathematics, and science is still far behind compared
to China, Singapore, and Malaysia. In this regard, developing countries such
as Indonesia urgently need to improve cognitive abilities of children in order
to perform better.
Figure 1. PISA’s Scores in Some Asian Countries in 2018
800
549 569 551 555 591 590
600 487 489 489
393 419 426 440 438
371 379 393 415
400
200
0
The Average Singapura Thailand China Indonesia Malaysia
OECD Score
𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 𝑜𝑜𝑜𝑜 𝑠𝑠𝑠𝑠ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑔𝑔𝑖𝑖𝑖𝑖 = 𝛽𝛽0 + 𝛽𝛽1 𝑎𝑎𝑖𝑖𝑖𝑖 + 𝛽𝛽2 𝑏𝑏𝑖𝑖𝑖𝑖 + 𝑒𝑒𝑖𝑖𝑖𝑖 and (2a)
ESTIMATION RESULTS
Table 5 shows the estimation results explaining the relationship of
socioeconomic status, especially parents’ education, with their children’s
cognitive abilities, as shown in Model 1. Model 1, which was developed
based on Erola et al. (2016) and Nieland (2021) study, claims that parental
socioeconomic status and educational attainment correlate with their
children’s educational achievement. Hence, the question is: Are intelligent
sons associated with intelligent mothers, and conversely, are fathers with
higher education associated with daughters intelligent? On the other hand,
Model 2 explains the relationship of the child’s cognitive abilities with
educational and work outcomes in adulthood. The estimation results related
to the hypothesis are presented in Table 5.
Table 5 shows that parents, especially mothers, passing her intelligence
genes to their children. Mother’s education has a positive relationship
with children’s cognitive scores. These results are supported by Wang’s et
al. (2020) study which states that parental education, especially mothers’
education, significantly impacts their children’s academic achievement.
However, it turns out that fathers tend not to pass their intelligence genes on
to daughters or sons, as evidenced by the estimation results above. Mothers
tend to pass their intelligence genes on to boys and girls, as evidenced by
the estimation results, which show that the mother’s education variable is
significantly positively correlated with the cognitive scores of girls with a
significance level of 1% and boys at a level of 5%. Mothers’ education, as
measured by the length of schooling, can also improve the cognitive scores of
their sons and daughters, but the increase is relatively small. If the mother’s
length of education is increased by one year, it can increase 1.75 cognitive
scores for girls and 1,022 for boys’ cognitive scores.
The natural logarithm of education expenditures also has a positive
and significant relationship with children’s cognitive scores. One percent
increase in spending on education associated with a 3.4 percent increase
of children’s cognitive scores. These results also support research by Kang
INTRODUCTION
Education has been emphasised as one of the keys to encourage female
labour force participation, not only by providing better job opportunities
for women but also by increasing their income (Mammen and Paxson
2000; Cameron, Dowling, and Worswick 2001; Jaumotte 2003; Besamusca,
Tijdens, Keune, and Steinmetz 2015). In addition, previous studies found
that the relationship between educational attainment and female labour
force participation is not linear across all levels of education, but forms a
U-shape relationship (Cameron, Dowling, and Worswick 2001; Klasen and
Pieters 2015; Schaner and Das 2016; Mehrotra and Parida,2017; Klasen et al.
2021). The findings show that female labour force participation is high for
women who are not in school, low for women with primary and secondary
education levels, and high for women with tertiary education level.
However, this relationship differs across countries. Cameron, Dowling,
and Worswick (2001) analysed the relationship between educational
attainment and female labour force participation (FLFP) in Korea, Thailand,
Sri Lanka, Indonesia, and the Philippines for the period of 1975-1976 and
only found the U-shape pattern in Sri Lanka. Examining the driven factors
of FLFP in eight developing countries for the period of 2000-2014, Klasen
et al. (2021) found that the relationship was positive and linear for Brazil
131
and South Africa, U or J-shape for Jordan, India, Indonesia, and the mixture
of both for Bolivia, Vietnam, and Tanzania. Consequently, it is important
to have further empirical studies on the relationship between educational
attainment and FLFP, especially in the case of a particular country.
This study aims to examine the effect of educational attainment,
represented by years of schooling, on FLFP in Indonesia. Investigating
this relationship in Indonesia is important because despite the significant
improvement in female educational attainment, FLFP rate in Indonesia has
remained constant for the past 20 years. Female school enrollment rates for
secondary education in Indonesia experienced a large increase from only
53% in 1999 to 90% in 2018. Female enrollment rate for higher education also
shows an increasing trend from 13.4% in 1999 to 39% in 2018 (Figure 1).
Figure 1. Female School Enrollment Rate (%)
120,00 109,70
104,59
100,00 90,04
80,00
60,00 53,10
38,99
40,00
20,00 13,42
0,00
Female Male
LITERATURE REVIEW
A large number of studies on female labour force participation have been
conducted, especially in the developing countries where women face several
barriers to participate in labour market. Klasen and Pieters (2015) and
Mehrotra and Parida (2017) analysed the causes of low and stagnant female
labour force participation in India for periods 1980s to 2011. They found that
the cause of stagnation in India was a combination between labour supply
and demand factors. From the demand side, structural change in economic
activities from agriculture to construction sector has limited women to
participate in labour force. From the supply side, the rising household income
led to a withdrawal of women from the labour market due to income effect.
Klasen and Pieters (2015) also found that the increase in women’s education
did not make a great contribution to FLFP due to the U-shaped relationship
between women’s educational attainment and FLFP.
In the case of Indonesia, Cameron, Dowling, and Worswick (2001)
conducted an empirical study on relationship between educational attainment
and FLFP within the period of 1975-1976 in five Asian countries, including
Indonesia. The result showed no U-shape relationship between educational
attainment and FLFP. However, they found that intermediate, secondary,
and tertiary levels of education have positive impact on Indonesia women’s
participation.
Schaner and Das (2016) and Cameron, Suarez, and Rowell (2019) also
analysed the main drivers of FLFP in Indonesia for periods 1990s to 2010s.
These two studies found that the main drivers of FLFP in Indonesia are
marital status, educational attainment (especially tertiary level), and the
presence of young children under the age of five. According to Schaner and
Das (2016), the cause of FLFP stagnation in Indonesia is due to the U-shape
relationship between FLFP and educational attainment where an increase in
FLFP at the higher education group has been offset by a decrease in FLFP at
lower education group.
Klasen et al. (2021) also investigated the determinant of FLFP for urban
married women in eight developing and emerging countries, including
Indonesia. Their data covers period from 2000 to 2014. Similar to Schaner
and Das (2016), they confirmed U-shaped relationship between educational
attainment and FLFP in Indonesia. Relative to women who did not finish
DATA
The main data in this study are taken from the Indonesian Family Life
Survey (IFLS). IFLS is a longitudinal survey in Indonesia, which has been
conducted six times. The first wave (IFLS 1) was conducted in 1993, the
IFLS 2 in 1997, and the IFLS 2+ in 1998; whereas the IFLS 3, IFLS 4, and
IFLS 5 were conducted in 2000, 2007, and 2014, respectively. Besides IFLS,
there are other data sources which better represent Indonesian population
such as SUSENAS and SAKERNAS. This study uses IFLS since it is publicly
accessible. In addition, IFLS sample is representative of about 83% of the
Indonesian population living in 13 of the nation’s 26 provinces in 1993.
However, IFLS is not a representative for the whole Indonesia, particularly
those in the eastern part of the country.
This study utilises the IFLS 3 (Strauss et al. 2004), the IFLS 4 (Strauss et
al. 2009) and the IFLS 5 (Strauss, Sikoki, and Witoelar 2016) to obtain a pooled
dataset of individuals. The 2000, 2007, and 2014 waves of the Indonesian
Family Life Survey (IFLS) are selected to cover periods of female labour force
participation stagnation.
The samples are limited to married women aged 25 to 54 in each
survey wave. Working women in the age of 25-54 are selected to avoid
the issue of school enrollment and retirement. Married women are chosen
because previous studies have shown that the issue in female labour force
participation is among married women (Cameron, Suarez, and Rowell
2019; Mehrotra and Parida 2017). In addition, previous studies on labour
supply focused on the family context in work decision making (Blundell and
MaCurdy 1999).
The main outcome of interest is female labour force participation. FLFP
is measured as a binary variable which equals one if women participate in
labour force and equals 0 if women did not participate in labour force1. This
1 This variable refers to the definition of labour force participation by International Labour
Organization (2022) which is the number of persons in the labour force as a percentage of the
working-age population. The labour force is the sum of the number of persons employed and
the number of persons unemployed. This study includes people who was, last week, working for
minimum 1 hour, including unpaid family workers, as part of persons employed (Badan Pusat
Statistik 2022).
This model is modified from Klasen and Pieters (2015) and Mehrotra and
Parida (2017). This study will estimate the model for each survey year
separately to determine whether the correlation change over time.
The dependent variable is a dummy variable which equals 1
(one) if women participate in labour force and zero otherwise2. The main
independent variables are that are the female years of schooling and
its square The control variables consist of four vector variables.
The vector u consists individual characteristic variables including age and
its square, religion fixed effects, father’s years of schooling, and mother’s
years of schooling. The vector w contains household characteristic variables,
including household head’s years of schooling, the natural log of per capita
weekly expenditure, dummy of having at least having one male with
salaried employment, household size, number of children aged 0-4 and 5-14,
number of adult females aged 15-65, and number of elderly women and men
aged 65+. The vector z includes regional characteristic which is provincial
unemployment rate in each survey wave. The subscripts, i, t, and r indicate
an individual, time index, and province, respectively. The term indicates
province fixed effect, and εit is error term.
This study first estimates the effect of education on FLFP using probit
regression, which is then followed by a calculation of the derivation from
each estimation result to determine a turning point when women enter
the labour market. Lastly, this study conducts heterogeneity analysis by
disaggregating the samples based on respondent region (Java vs Off-Java)
and the presence of babies.
RESULT
Estimation Results
Probit estimation results are shown in Table 2. The table presents the average
marginal effect (AME) for each independent variables, which shows the
change in probability of female participation in labour force due to one-unit
2 Ibid
HETEROGENEITY ANALYSIS
Java vs Off-Java
This study conducts heterogeneity analysis by separating the result between
Java and Off-Java. The result is shown in Table 3. This study confirms a
U-shape relationship between educational attainment and female labour
force participation in Java and Off-Java, except for the 2000 period. The
turning point for educational attainment in Java is increasing from 6 years
of schooling in 2007 to 7 years of schooling in 2014. However, for Off-Java,
the turning point is decreasing from 8 years of schooling in 2007 to 7 years
of schooling in 2014. This result is consistent with the earlier findings by
Cameron, Suarez, and Rowell (2019) in which a strong correlation between
educational attainment (especially upper secondary and tertiary school) and
FLFP for Java-Bali and other islands.
CONCLUSION
Using data from the Indonesian Family Life Survey (IFLS) of 2000, 2007,
and 2014 periods, this study finds that educational attainment significantly
affects female labour force participation in Indonesia. This study also
confirms the U-shaped relationship between educational attainment and
female labour force participation. The average marginal effect from probit
estimation shows that educational attainment will have a negative effect on
the probability of female labour force participation until reaching the turning
point of 7 years of schooling, equivalent to first-grade junior high school. The
turning point is also similar to the mean of female schooling in this sample.
After achieving 7 years of schooling, the additional years of schooling will
increase the probability of female labour force participation. By estimating
the probit model for each year separately, this study finds that the turning
point did not change much from 2007 to 2014. The result is also consistent
for both Java and Off-Java. Lastly, this study provides evidence that women
who have babies enter the labour market later than their childless peers.
Our findings provide important implications for social welfare policy
in Indonesia. First, since higher education has a positive significant effect on
female labour force participation, policies that focus on increasing women
access to upper secondary and tertiary school would improve women’s
participation in labour force. The government can increase the effectiveness
of the program by providing better information about the importance of
college and the availability of tuition assistance for the under-informed
and underprivileged communities. Several studies have found that simple
interventions by providing more complete information about the return of
education and tuition assistance can improve educational attainment (Jensen
2012; Castleman and Page 2015). The provision of this information can be in
the form of a campaign through text messages or peer mentoring which does
not cost a lot of resources. The government should take this programme into
consideration, since the enrollment rate of higher education, especially in
rural areas, is still low (Badan Pusat Statistik 2020).
REFERENCES
Badan Pusat Statistik. 2020. ‘Angka Partisipasi Kasar (APK) Perguruan
Tinggi (PT) Menurut Daerah Tempat Tinggal, 2011-2019.’
(“Badan Pusat Statistik”) 20 July. https://www.bps.go.id/
dynamictable/2018/07/11/1524/angka-partisipasi-kasar-apk-
perguruan-tinggi-pt-menurut-daerah-tempat-tinggal-2011-2019.html.
________. 2022. ‘Tenaga Kerja: Konsep/Penjelasan Teknis.’ 07 February.
https://www.bps.go.id/subject/6/tenaga-kerja.html .
Baker, Michael, Jonathan Gruber, and Kevin Milligan. 2008. Universal
Child Care, Maternal Labour Supply, and Family Well-Being. Journal of
Political Economy 116 (4): 709–45.
Berlinski, Samuel and Sebastian Galiani. 2007. The Effect of a Large Expansion
of Pre-Primary School Facilities on Preschool Attendance and Maternal
INTRODUCTION
After experiencing a period of submerging during the New Order period,
Indonesia’s agrarian reform policy was finally strengthened again through
the National Agrarian Reform Programme (Programme Pembaruan Agraria
Nasional or PPAN) launched by Susilo Bambang Yudhoyono (SBY)
government in 2007 (Shohibuddin and Salim 2012). Then, the Jokowi
government continued the programmeme through its inclusion in 2015-2019
National Medium-Term Development Plan (Rencana Pembangunan Jangka
Menengah Nasional/RPJMN) and the 2020-2024 RPJMN as one of the national
goals priority programmes.
The acceleration of the agrarian reform promotes forest lands as the fifth
priority programme of the 2016-2019 National Strategy, simultaneously bring
about the management of forest resources on the hand of the community
people (Kompas.com 2014; BAPPENAS 2014). In 2017, a legal foothold on
the Settlement of Land Tenure in Forest Areas (Penyelesaian Penguasaan
Tanah Dalam Kawasan Hutan or PPTKH) was published through Presidential
Regulation No. 88/2017. In the following year, Presidential Regulation No.
86/2018 on Agrarian Reform entailing forest areas as one of the objects
of agrarian reform was also published. These two regulations are the
government’s instruments to establish the Land Objects of Agrarian Reform
151
(TORA) programme in forest areas which is expected to solve problems
related to land settlement in forest areas.
The implementation of these policies, however, received many criticisms,
especially from civil society organisations. These organisations then
actively proposed TORA based on people participatory mapping processes
(Berdikarionline.com 2017; Wirawan 2018; Ekawati, Salim, and Utami 2019;
Celebespos.com 2020; Susetyo 2021; Tempo.co 2021; Panahpapua.or.id 2021).
They argue that people participation is the pre-request for the programme to
be successful (Carranza 2015; Kantor Staf Presiden Republik Indonesia 2017;
Djauhari, Lubis, and Moenir 2018; Suharti, 2019; Mitchell, Safik and Dodd
2020; Winarti, Joko, and Budiman 2021).
In Sigi District, Central Sulawesi Province, there has been a collaboration
between the local government of Sigi District, GTRA, and civil society
organisations (i.e., YMP, JKPP and KPA)1 together with residents did
participatory work to propose a TORA programme to the central and
provincial governments (JKPP 2017; YMP 2017; Jemali 2017; YMP 2019;
Saputra 2020). In 2017, they proposed 168,000 hectares of forest areas to be
part of their TORA programme. The Inver PTKH Team (Tim Inventarisasi dan
Verifikasi Penguasaan Tanah dalam Kawasan Hutan), however, approved only
3,000 hectares to be processed as the TORA programme of forest areas in Sigi
District. This case shows how significant it can be the gap in size between the
area proposed to be in a TORA programme by local communities and the
final results by the PTKH Inver Team. Even in the case of high community
participation, the final result is mainly determined by the PTKH Inver Team
and this can be far from communities’ expectations.
This TORA scheme is finally implemented under the initiative of the Sigi
District Government of Sulawesi Tengah Province (Savitri 2020), which in
turn has made Sigi District a pilot district for the national TORA programme.
Ironically, the implementation of this forestry agrarian reform in the Sigi
District encounters many obstacles. For example, one village in the Sigi
District proposed a TORA programme for forest areas is the Balumpewa
Village. Residents in this village have participated in various endeavours,
ranging from participatory mapping, allocating budget for agrarian reform
from village funds, to taking direct actions. Unfortunately, the TORA
programme in the Balumpewa Village is hitherto stagnant.
1 YMP stands for Yayasan Merah Putih, JKPP stands for Participatory Mapping Network, and KPA
stands for Consortium for Agrarian Reform
METHODOLOGY
The first part of this research, conducted from July until October 2020, was
a comprehensive desk study to unpack and review agrarian reform policies
and regulations in Indonesia, from Soekarno’s Era to the era of Jokowi. The
second part of this researcher was conducting a fieldwork on November-
December 2020 in Balumpewa Village, Sigi District, Central Sulawesi
Province. During this fieldwork, direct observations among farmers and
two rounds of Focus Group Discussions (FGDs) with various stakeholders
related to the theme of this research followed by in-depth interviews were
organized.
The first round of FGD is to examine the history of Balumpewa Village,
both from an administrative perspective and based on folklore. We were
2 A part of this dissertation discusses (1) How did Plantation Land become Excluded from the
Land Redistribution Programme and (2) How did Forestry Become Separate from the Agrarian,
and How did Forestland become Excluded from the Land Redistribution Programme? It shows
the sectoralism of regulation and institution in Indonesia. Currently, forests are included in the
agrarian reform programme but are too complicated to implement, so this is the main topic of this
chapter.
3 Highlands have a minimum elevation of 500 meters above sea level, while mountains are usually
above 1,000 meters above sea level and have peaked.
4 Terms in Kaili Inde Language
6 Shohibuddin (2020) termed this kind of inequality as “allocation inequality,” namely the condition
of land tenure inequality caused by unfair state policy in allocating land between the interests of
the people and for the benefit of sectoral control on a large scale, such as for forest areas, plantation
HGU, and so on. This is distinguished from the “inequality of distribution,” which more closely
reflects the unequal land tenure conditions among the farmers themselves.
7 The data is processed from the survey results of the research team.
8 The Jokowi administration divides the AR policy into two terms, namely ‘Asset Management’
and ‘Access Management’. This asset arrangement is further divided into two, namely land
redistribution and asset legalization. The forestry AR policy is in the category of land redistribution
(release of forest areas). Meanwhile, social forestry is part of the arrangement of access.
9 Referring to the Presidential Regulation of the Republic of Indonesia Number 88 of 2017 concerning
Settlement of Land Tenure in Forest Areas.
10 ‘adverse formalisation’ where the state claims sovereign control of extensive ‘public lands’ and
embarks on formalisation processes that include local populations into new land-based production
systems on adverse terms, (McCarthy et al. 2022).
CONCLUSION
This chapter explores the obstacles to implementing agrarian reform in forest
areas. We explore Balumpewa Village’s case to show how the dynamics of
the agrarian reform programme are beyond institutional and sectoral issues.
The majority of land in this village is forest. Along the time, the Balumpewa’s
residents have been cultivating the forest. However, since the government
of the Sigi district adopted the agrarian reform programme by president
Jokowi, people in this village are struggling to enforce it. The participation of
the Balumpewa community in implementing agrarian reform programmes
is very high. They take much collective actions, which is supported by
local institutions such as customary institutions, religious institutions or
churches, and the village government. This indicates the enthusiasm and
determination of all components in Balumpewa Village to carry out the
agrarian reform programme.
Nevertheless, the community participation has not been seriously
considered and their efforts were unable to influence the Inver PTKH Inver
Team’s decision-making to change or release the status of the forest area
to be a TORA as requested by their proposals either. The results of the
verification issued by the PTKH Inver Team become a definitive decision.
This ‘filtration process’ becomes the right term used to show how the
regulation (1. Presidential Regulation No. 88/2017 concerning Settlement of
Land Tenure in Forest Areas; and 2. Regulation of the Coordinating Minister
for Economic Affairs No. 33/2018 concerning Guidelines for the Inventory
and Verification of Land Tenure Teams in Forest Areas) is a tool to maintain
the status quo of forest areas.
Implementing forestry agrarian reform has become problematic because
of conventional regulatory and rigid bureaucracy. The community does not
have the opportunity to sue or reject the verification results. So, various
participatory works ended up deadlocked because of rigorous regulations
and generalised the whole situation at the site level. Various participatory
efforts of citizens to propose forest areas as objects of agrarian reform have
stalled. Too many conditions to be met while the offer of solutions from the
state is restricted. The current regulations have not responded to the diversity
of socio-ecological situations in the villages that apply for agrarian reform.
Acknowledgments
This article is part of the Sajogyo Institute research supported by the
Knowledge Sector Initiative (KSI). Our thanks to both of these institutions.
Furthermore, we would like to thank all the local community and agrarian
activists we met during our activities in Sulawesi Tengah, especially in Sigi
District, who have assisted us during the research process.
REFERENCES
Antaranews.com. 2019. Warga 103 desa di Sigi ikut konsolidasi reforma agraria.
[online] https://www.antaranews.com. https://www.antaranews.
com/berita/1052714/warga-103-desa-di-sigi-ikut-konsolidasi-
reforma-agraria [Accessed 20 Mar. 2022].
Ariyo Wicaksono, R. 2020. Perhutanan Sosial Mestinya untuk Pemulihan
Kedaulatan Masyarakat. [online] https://betahita.id/. https://betahita.
id/news/lipsus/5745/perhutanan-sosial-mestinya-untuk-pemulihan-
kedaulatan-masyarakat.html?v=1606983148 [Accessed 20 Mar. 2022].
Bachriadi D. 2017. “Reforma Agraria untuk Indonesia: Kritik atas Reforma
ala SBY.” Working Paper series in kebijakan agraria dan pembangunan
paska Orde Baru, Vol.1 No.07/WP-KAPPOB/I/2017. ARC, Bandung.
BAPPENAS. 2014. Rencana Pembangunan Jangka Menengah Nasional
(RPJMN) 2015-2019. [online] Jakarta: BAPPENAS. https://djsn.go.id/
storage/app/media/RPJM/BUKU%20I%20RPJMN%202015-2019.pdf
[Accessed 16 Mar. 2022].
BPS. 2020. Kecamatan Dolo Barat dalam Angka 2020. Badan Pusat Statistik
Kabupaten Sigi.
Berdikarionline.com. 2017. Ini 7 Alasan Mengapa Programme TORA Jauh
INTRODUCTION
Indonesia, as an archipelagic country, faces challenges in realising an
efficient maritime logistics system. According to the World Bank report
(2018), Indonesia’s logistics performance has several weaknesses. Namely,
the distance to export and import is far; the integration of industrial zones
with port infrastructure is still weak; the load time, the number of documents,
and the clearance time are still relatively high. A study conducted by Cho
(2014) mentioned that high logistics costs incurred at container ports are
associated with completing the procedures to export or import the goods.
These costs include document fees, administrative fees for customs clearance
and technical control, custom brokers, terminal handling charges, and inland
connection. Uncoordinated logistics chain is another factor proved causing
logistic costs to increase. The number of container loads returning empty
and low coordination systems could result in expensive administrative costs
(Mohamed-Chérif and Ducruet 2016).
The logistic issues taking environmental and institutional uncertainty
into account play a significant role in creating transaction costs (Huo et al.,
2018). In maritime logistics, port institutions should adapt to the changing
demands and dynamics of the global market. Ports are needed to make
adjustments to meet international standards both in physical infrastructure
177
and institutional infrastructure. Stough (2005) suggests that institutional
reform in port management is fundamental in dealing with the crisis and
rapid globalisation dynamics. Port authorities and maritime stakeholders
in the regions should be sensitive to environmental changes and market
uncertainty fluctuations that are difficult to predict and vulnerable to global
economic volatility.
Transportation costs, inadequate transportation systems, and timeliness
in the supply chain are components of the logistics system that can impact
maritime logistics costs. Higher port efficiencies are associated with lower
maritime transport costs (Lei and Bachmann 2020). In particular, for island
areas, irregular boat schedules, high sea transportation costs, and minimal
transportation modes are obstacles in minimizing logistics costs (Briguglio
1995). Similarly, an organized transportation system and unsatisfactory
timeliness in the supply chain created an inefficient logistics service system
in archipelago regions (Kerr 2005).
The research in this chapter took a case in North Maluku, one of the
island provinces in the eastern part of Indonesia. The efficiency of maritime
logistics costs, therefore, is crucial for the island’s economic development.
This chapter attempts to analyse the factors that affect maritime logistics costs
in these archipelagic area using a transaction cost economic approach. The
islanders of this area live with the availability of minimal port infrastructure
facilities, poor land connectivity, inefficient governance in setting tariffs, and
high transportation costs. Hence, it is assumed that there are high transaction
costs in the maritime logistics system for transporting goods between islands.
The transaction costs in the maritime logistics system include operational
costs during the production process, land transportation rental costs, loading,
and unloading costs at island ports. Other costs are administrative costs, sea
transportation rental fees to the destination island, loading and unloading
costs at the destination island port, land transportation costs to warehouses
or storage areas, packing costs, warehouse costs, and filling costs (Zeng and
Rossetti 2003; Panayides 2006).
This chapter is organised as follows: the next section outlines the
literature review related to transaction cost economics in the maritime
logistics system. The later part discusses data and methodology. Section of
result and discussion describes the interrelationships between transaction
cost variables. Conclusions and policy implications are presented at the end
of this chapter.
DATA COLLECTION
Data collection was carried out through the focus group discussion (FGD)
method to assist information sharing and transfers between participants
(Asnawi et al. 2020). This method is used to identify and determine key
variables in measuring the causes of transaction costs in the maritime
logistics system. The study adopts the approach used by Cho (2014) that
divided the components of transaction costs into two categories: first,
expenses caused by the asset specificity, including port infrastructure,
availability of port services, and port accessibility, and second, costs that
arise due to environmental uncertainty include maritime uncertainty and
institutional uncertainty. Maritime uncertainty is a condition that causes
search costs to appear in trade by sea due to market uncertainty (Hallikas,
Virolainen and Tuominen 2002). Meanwhile, institutional uncertainty is a
condition of uncertainty regarding institutional regulations, institutional
professionalism, and the actors’ habits or behaviour that can cause transaction
costs in operationalisation and governance (Grewal and Dhawadkar 2002).
RESULTS
The MICMAC methods grouped the variables into four quadrants according
to the categories of its dependence and influence (Figure 2). Quadrant I is
the influencing variables that display elements that are very dominant with
less dependence. This element is an essential variable in the system because
it can become a critical factor. Quadrant II contains relay variables that are
influential and highly dependent. This variable is often classified as elements
that depict the system instability. Quadrant III includes the dependent
variables or output variables, represented by high dependence and small
influence. This variable is relatively sensitive to the reversal in influence and
relay variables. Excluded variables in Quadrant IV are called autonomous
variables (Saxena, Sushil and Vrat 1990). According to the FGD results, the
maritime logistics transaction costs are determined by 11 main elements
DISCUSSION
The relationship between variables that affect maritime logistics transactions
costs shows port management’s complexity in North Maluku Province. Based
on the prospective MICMAC analysis results of the direct influence between
variables, there are six central elements that trigger high maritime logistics
transactions costs: weather, infrastructure, cargo uncertainty, behaviour,
economic scale of island, and regulation.
Weather is a variable that has a substantial burden on transaction
costs in the maritime logistics system. As the area consists of small islands,
environmental uncertainty due to extreme weather can hamper the
distribution of goods between islands. The loading and unloading goods
at the port with minimal loading infrastructure can be disrupted and run
slowly if the weather is terrible. The berth of the ship may usually take
longer time, which causes the anchor costs to increase. Weather uncertainties
also impact the supply side of goods, considering the vessel used is quite
vulnerable to being affected by sea waves to disrupt logistics activities. The
amount of goods available for sale to consumers on the island is limited,
impacting the selling price accordingly. For example, the commodities most
vulnerable to fluctuations due to the weather are fish and vegetables, as they
are perishable. If the weather is extreme, areas such as Ternate island, are the
most susceptible. Most of their food needs, such as vegetables, are transported
by public boat or ferry from Bitung, North Sulawesi, and then distributed
to the surrounding islands. If the sea waves are high, the transportation of
goods to be delivered to the destination island will be hampered . Delay in
delivering goods to end customers will increase logistics costs if the goods
are damaged, especially perishable goods.
CONCLUSION
The efficiency of maritime logistics costs is crucial for the islands’ economic
development in North Maluku, one of the island provinces in the eastern part
of Indonesia. The islanders of this area have been living with the available
minimal port infrastructure facilities, poor land connectivity, inefficient
REFERENCES
Amin, C. et al. 2021. ‘Impact of maritime logistics on archipelagic economic
development.’ The Asian Journal of Shipping and Logistics, (xxxx). https://
doi: 10.1016/j.ajsl.2021.01.004.
Arcade, J. et al. 2003.’ Structural Analysis with The MICMAC Method &
Actors’ Strategy with MACTOR Method.’ The Millenium Project, 3.0,
0–48.
INTRODUCTION
The fisheries sector has long been developing and has become an important
sector in economic activity in many countries, where there are millions of
people in the world who depend on this sector for their livelihoods. Data
from the Food and Agriculture Organization of the United Nation (FAO
2018) states that fish or fishery products are the most traded type of food in
the world today.
The fishery sector has become one of the important economic sectors
globally. There are 35% of world fish production entered international trade
in various forms for human consumption or non-edible purposes. 60 million
tonnes (live weight equivalent) of total fish and fish products exported in
2016 represented a 245% increase since 1976. During the same period, world
trade in fish and fish products also grew significantly in terms of value, with
exports increasing from USD 8 billion in 1976 to USD 143 billion in 2016.
In the last 40 years, this figure has grown much faster than exports from
developed countries (FAO 2018).
Béné, Devereux, and Roelen (2015), define fisheries in terms of the
people involved, species or types of fish, territorial waters or seabed, fishing
methods, class of boat, the purpose of the activity, or a combination of these.
As an economic activity, fisheries are characterised by their operational scale
201
ranging from small to large scale. There are approximately 40 million people
involved in the fishing sector, and almost 90% operate small-scale fishing
activities, and most live in low- and middle-income countries. This includes
self-employed fishers (who own the boat and fishing gear), crew members
(who are contracted by the boat owner to operate the vessel), as well as
households involved in fishing).
Nonetheless, some other literature describes fishing communities as
‘the poorest of the poor, since their income depends on catches which vary
widely, is uneven and unpredictable (Bailey and Jentoft 1990; Béné and Friend
2011; Jentoft and Eide 2011). Yet, in many places, small-scale fishermen are
considered poorer but happier than those with other occupations, although
fishing is considered a job with a lower social status in society (Pitchford et
al. 2007; Fromentin et al. 2014; Anna et al. 2019).
The relationship between fisheries and poverty in developing countries
has often been debated by experts. Béné, Devereux, and Roelen (2015),
explained that the debate focused on the economic aspect, where the
lives of fishermen are low-income, and the biological aspect, namely the
overexploitation of the sea. The debate about poverty (income) of small-scale
fishermen has been going on for a long time, since FAO revealed that 20% of
those working in the global fisheries sector were estimated to earn less than
USD 1 per day, in 2005.
The relationship between fishing and income poverty is complex. In
many cases fishing communities may be prosperous, or even richer (in terms
of income) than other agricultural communities (Béné and Friend 2011;
Jentoft and Eide 2011). However, fishermen’s incomes depend on the size
of their catch which is usually highly variable, uneven, and unpredictable.
Their level of poverty is often more closely related to the severe degree of
geographic and/or political isolation that characterizes many full-time or
mixed fishing communities.
Fishery resources are mobile, and easily moving from one place to
another is a challenge faced by fishermen. Acheson (1981); Stanford et al.,
(2013) explained that poor fishermen depend entirely on the sea, so they
need modern fishing facilities and other supporting facilities to survive
long on the water. Islam et al., (2014) explain that fishermen who use boats
with larger engine capacities earn large incomes. This reflects an unequal
distribution of benefits, as only those who can invest in fishing facilities and
who spend more on fishing operations can derive maximum benefits.
MULTIDIMENSIONAL POVERTY
Measurement of poverty has evolved since its discovery. In general, the
method of calculating the poverty rate uses a monetary approach in which it
is estimated by the value of money (income, expenditure, or consumption).
The advantage of this method is that the measurement is clear, quantitative
in nature so that it can measure the depth of poverty. However, according to
Hulme, Moore, and Shepherd (2001), such an approach will play a significant
role in future research but must be complemented by a multidimensional
approach to poverty that is qualitative in nature and allows subjective
assessment by the poor.
From the outset economists have recognised some basic limitations of
the income method, namely patterns of consumer behaviour may not be
uniform. Therefore, reaching a poverty line level of income does not guarantee
that a person will meet his or her minimum needs. Price differences that can
reduce the accuracy of the poverty line, and the ability to convert a certain
amount of income into certain functions vary according to age, gender,
health, location, climate, quality of affordable services, such as water, health
and education which are often not provided through the market (Akire and
Santos 2014).
According to Anshory Yusuf (2018), using income or expenditure as a
basis for measuring welfare is indeed quite common. However, measuring
welfare using the income basis is still incomplete in capturing the various
dimensions of welfare itself. If welfare is calculated based on income alone,
it has not fully captured the non-income dimensions of real welfare. People
who may not be categorised as poor because their income exceeds the
poverty line may have exceptionally low education, do not have access to
proper water and sanitation, live in houses of limited quality, and do not
have permanent jobs. This question is to be answered with a concept called
(1/4) 1/20 - Do not have a motorized vehicle and only have one
of the following items: Bicycle, Motorcycle, Radio,
Refrigerator, Telephone, or Television
1/8 - The job status that generates the main income is as a
Employment (1/4) labor fisherman
1/8 - Using traditional boats to fish
Source: Alkire et al. (2011), Modified.
The stages of measuring multidimensional poverty based on the Alkire-
Foster method (Alkire and Foster 2011a) are as follows:
–– Choose the unit of analysis (i.e., the household)
–– Choose the dimension of poverty
–– Choose an indicator from each indicator
–– Determine the cut off deprivation of each indicator
–– Determine the weight (wi) of each dimension/indicator
–– Calculating the deprivation score (∑) experienced by the household:
∑ ci = ωiI1 + ωiI2 + ωiI3 +… + ωdId
Where I1 = 1 if the household lacks the indicator i and = 0 if the household
is not deprived at the indicator of i
METHODOLOGY
The Logit model is used to analyse the relationship between institutional
access and poverty of capture fisheries households in Indonesia. Poverty
status (dependent variable) is measured in two monetary approaches
(income) and a multidimensional approach (multidimensional poverty
index). The association model for institutional access and household poverty
of capture fisheries is defined as follows:
Variable Description
Poverty Household Poverty Value 1 if the household is poor, and zero if non-
Status (monetary and poor
Multidimensional)
Households Value 1 if gender of the household head is Male,
Gender
Characteristic and zero if not
Age Age of the household head
Value 1 if the household head did not finish
Education
primary school, and zero if finished primary school
Household size Number of household members
Institutions Value 1 if household receives credit access, and
Credit Access
Access zero if not
Utilization of Value 1 if household receives utilisation of
Cooperative Facilities cooperative facilities, and zero if not
Member of Value 1 if household is a member of Kelompok
Kelompok Usaha Usaha Bersama (KUB), and zero if not
Bersama/KUB
Access to production Value 1 if household has access to production
facilities facilities, and zero if not
Value 1 if household has access to market, and zero
Access to market
if not
Value 1 if household receives boat grant, and zero
Receive boat grant
if not
Receive fishing gear Value 1 if household receives fishing gear grant,
grant and zero if not
Receive financing Value 1 if household receives financing grant, and
grant zero if not
DATA
The dataset utilised for this chapter is the household sample data for capture
fisheries in the 2013 survey of household incomes for agriculture/capture
fisheries in Indonesia by the Central Bureau of Statistics (BPS). This survey
is part of the agricultural census, which is conducted every ten years. The
number of households selected as the sample in this survey is 28,779. The
poverty analysis focused on capture fisheries households nationally and
divided the sample for poverty analysis on households in Java and outside
Java. This helps to demonstrate the consistency and robustness of the
estimation results, as well as to test whether there are significant differences
in the characteristics of household poverty in Java, Outside Java, and
Nationally.
Moneter Multidimensional
19,91 17,26
80,09 82,74
RESULTS
As explained earlier, this chapter focused on capture fisheries households
nationally, and the samples for poverty analysis on households were divided
into those living in Java and outside Java. The logistic regression estimation
results from the institutional access model and household poverty of capture
fisheries are presented in Table 3.
-0.2669 *** -0.5760 *** -0.2257 *** -0.1540 *** -0.2351 *** -0.1856 ***
Member of KUB
0.1048 0.1438 0.0397 0.4040 0.0371 0.0387
Access to -0.2112 *** 0.0756 -0.2558 *** -0.1785 *** -0.2609 *** -0.2012 ***
production
facilities 0.1097 0.1428 0.0366 0.0373 0.0346 0.0358
Access to -0.8656 0.0289 -0.1045 *** -0.3641 *** -0.1095 *** -0.4007 ***
market 0.1687 0.2131 0.0425 0.0421 0.0410 0.0409
Receive boat -0.4637 -0.8598 0.1396 0.1325 0.1199 0.1437
grant 0.4765 0.7287 0.1120 0.1135 0.1083 0.1111
Receive fishing -1.1891 *** 0.4072 -0.2908 *** 0.0878 -0.4101 *** 0.0783
gear grant 0.0395 0.3091 0.1214 0.1102 0.1146 0.1029
Receive -0.0940 0.1947 -0.1019 -0.0307 -0.1045 -0.0064
financing grant 0.2320 0.2756 0.0948 0.0943 0.0876 0.0888
-2.2371 -0.1947 -1.8746 0.5149 -1.8715 0.5072
Constant
0.3557 0.4334 0.1202 0.1200 0.1131 0.1147
Number of
4728 4728 24051 24051 28779 28779
Observation
LR chi2 (12) 143.26 105.72 1578.09 749.16 1784.91 867.25
Prob > chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Pseudo R2 0.0364 0.0374 0.0639 0.0321 0.0621 0.0328
Source : Processed from survey pendapatan rumah tangga pertanian (Badan Pusat
Statistik) with STATA.
Note : Numbers at the second line are standard errors.
REFERENCE
Acheson, J.M., 1981. Anthropology of fishing. Annual review of anthropology,
10(1):275-316.
Ankrah Twumasi, M. Y. Jiang, B. Addai, Z. Ding, A.A. Chandio, P. Fosu,
D. Asante, A. Siaw, F.O. Danquah, B.A. Korankye, and G. Ntim-Amo.