Toll Road Infrastructure Development in Indonesia: A System Dynamics Perspective
Lukas B. Sihombing
lukas.b.sihombing@gmail.com
Abstract
The development of infrastructure, especially toll roads in Indonesia, is critical because if
compared with toll roads in other countries like China or Malaysia, Indonesia toll road
growth is much slower. This paper will examine the development of toll road infrastructure in
Indonesia from 1995 to 2012, and then analyze it with a System Dynamics methodology. At
the conceptualization stage of SD methodology, there is a significant gap between forecasts
based on the current situation and forecasts based on an econometric data panel method. The
results of this paper, based on a current investment process from the Government of
Indonesia, is the total toll road construction, based on the government’s plan is 2213 km; it
will be finished in 2080. It means that toll road construction in Indonesia is very challenging.
This paper recommends that to accelerate the construction of toll roads in Indonesia, it should
be leveraged with something innovative.
Keywords: gap, econometric, system dynamics, toll road infrastructure
1. Introduction
According to the Masterplan Acceleration and Expansion of Economic Development
(MP3EI, 2011), the infrastructure, including roads, ports, airports, power plants, water
supplies, sewers, telecommunications, and other infrastructures are needed to strengthen the
inter and intra connectivity of future center economic growth. However, the object of this
paper is toll road infrastructure. According to Government Decree No.15, 2005, toll road
organization intended to bring about equitable development and its results as well as a
balance in regional development with regard to justice, which funds are derived from road
users to build a road network. The aim is to improve the efficiency of distribution services
and support economic growth, especially in areas that have high levels of development. A toll
road is not only an alternative road that can be taken from an existing public road, but it is
also an alternative road when there are no public roads in a particular region. Therefore, a toll
road is necessary to develop a particular area.
According to Aschauer (1989), the decline of infrastructure investments in the United States
from 1971 until 1985 reduced the productivity level to 0.8% annually for country-wide
businesses. Related to measuring infrastructure investments, Sutherland, Araujo, Egert, and
Kozluk (2009) said there are two kinds of measuring infrastructure investments. The first is
an estimation of capital stock, which is often used for calculations in a national context.
Meanwhile, the second is to measure physical infrastructure size. The measurement of toll
road infrastructure investments which is used in this study is physical toll road
infrastructures, such as toll road length in kilometers (km).
Tf toll road length is used as a measurement of investment, based on data, it can be shown that
the growth of toll road infrastructure in Indonesia is insignificant. For instance, in 2000 the
toll road length was 585.07 km, whereas in 2001 toll road growth only reached 0.95% or
590.62 km. At that time, there were 5 private investor funded toll roads or equal to 23%
(BPJT, 2007). The participation of the private sector is low due to the impact of the financial
crisis at the global level especially in Asia in mid-1997, and also the collapse of the domestic
commercial bank sector (The World Bank, 1999).
According to Jasa Marga (2012), “Jagorawi (Jakarta Bogor Ciawi) is the first toll road which
was operated by Jasa Marga in 1978. With a total length of 59 km, the toll road connects
Jakarta, Cibubur, Citeureum, Bogor, and Ciawi. The operation of Jagorawi marks a milestone
in the establishment of Jasa Marga (Persero) Tbk, as the company to develop and operate a
toll road in Indonesia.” Based on BPJT (2013), the total length of functional toll roads as of
2013 has reached 784.06 km (BPJT, 2013).
In fact, the Government of Indonesia through the Toll Road Authority Board (BPJT) engaged
in accelerated efforts in Indonesia for toll road investments as was issued by the Toll Road
Acceleration Program 2005-2009 for 1099.08 km length, consisting of 32 sections of Trans
Java toll roads or equal to 763.24 km length; and 19 sections of non-toll roads for Trans Java
or equal to 335.84 km length (BPJT, 2010). Nevertheless, the realization until 2010 was only
78.45 km length; so the average growth per year from 2006 to 2010 was 15.69 km or 2.3%
per year.
On the other side, the acceleration of toll road programs was also carried out by other
countries such as China and Malaysia, but the actual acceleration of toll road construction in
China and Malaysia is better than Indonesia. It can be shown from the Toll Road
Acceleration Program in China that the length of toll roads in 2007 reached 45,000 km
(World Bank, 1999); in 2010 it reached 65,000 km (CIA, 2012) or an average growth of 30%
annually from 1996 until 2010. By 2020, the expected length of toll roads in China is
expected to reach 85,000 km (Peopledaily, 2007). Meanwhile, Malaysia toll road length in
1966 was 20 km (Wikipedia, 2009), and the length of toll roads in 2010 reached 2,671.28 km
(Wikipedia, 2007) or an average growth of 9.24% annually from 1996 until 2010.
From the description above, the following question arises: How is the complexity of toll road
infrastructure development dynamics depicted in Indonesia?
To answer the question, a relevant approach to analyze the complexity of toll road
infrastructure development dynamics in Indonesia is the System Dynamics (SD) Approach,
because according to Sterman (2001), “System dynamics draws on cognitive and social
psychology, organization theory, economics, and other social sciences. To solve important
real world problems, we must learn how to work effectively with groups of busy
policymakers and how to catalyze change in organizations.” So in a system, Indonesia toll
road infrastructure development is very complex and dynamic. In the real world, for example,
there are delays in land acquisition, a shortage of development funds both by the government
and private parties, etc. Because of that, using a system dynamics method is relevant to
analyze the development of the complexity of toll road infrastructure development dynamics
in Indonesia.
The purpose of this paper is to describe the dynamics of Indonesia toll road infrastructure
development. The dynamic depiction of toll road development in Indonesia can be analyzed
by using a System Dynamics Approach. This method can be used to assist us in
comprehending the feedback system of toll road infrastructure development as being dynamic
and complex. According to Sterman (2000), System Dynamics can be used as a method to
improve understanding of complex systems in an interdisciplinary and fundamental way
because SD is based on a nonlinear dynamics theory and feedback control developed through
mathematics, physics, and engineering. SD draws on cognitive and social psychology,
organization theory, economics, and other social sciences, solving problems in the real world,
and as a bridge to a nonlinear system. According to Luan, Chen & Wang (2010), System
Dynamics (SD) can bridge a simulation system through an internal system structure
simulation; it can particularly handle a multi-loop nonlinear system.
The structure of this paper is as follows: Section 2 is the methodology and data sources. The
methodology used in this paper is a System Dynamics Approach, i.e. starting from a
conceptual design until making a learning strategy/infrastructure design. Meanwhile, the
source of data is that which supports the conceptual process of the SD method. Section 3 is a
literature review of toll roads that outlines the literature which explains the correlation
between the toll road lengths with the population and GDP. The literature outlines the
econometric analysis of infrastructure and describes system dynamics. Section 4 describes
SD conceptual methods to get a description of problems in toll road development and
compare them with 21 countries. This section also describes the SD model formula for toll
roads, starting from causal loops until stock and flow diagrams and then SD results. Finally,
Section 5 is the conclusion.
2. Methodology
The methodology used in this paper is a System Dynamics Approach. The stages of a system
dynamics modeling process according to Martinez-Mayano and Richardson (2013) are: 1)
conceptualization — problem identification and definition, system conceptualization; 2)
formulation — model formulation; 3) testing — model testing and evaluation; 4)
implementation — model use, implementation, and dissemination; and 5) design of learning
strategy/infrastructure. An overview of the system dynamics modeling approach can be seen
in Figure 1.
Design of Learning
Strategy/Inffastructure:
Model Use,
Implementation, and.
Dissemination
Understanding of the
Problem and the System,
Understanding of
the Model Problem Identification
and Definition
System
Model Testing and Conceptualization
Evaluation
Model
‘Formulation
Source: Martinez-Mayano and Richardson (2013)
Figure 1: Overview of the System Dynamics Modeling Approach
2.1 Source of Data
To identify the problems associated with physical gaps in toll road lengths in Indonesia,
namely to get the length of a toll road which is appropriate with the current situation,
historical data of Indonesia toll road lengths was used, starting from 1995 until 2012.
Meanwhile, to get desired toll road lengths, an econometric data panel method was used.
Besides toll road length data, data from the population and GDP for 21 countries had to be
collected, starting from 1995 until 2012. The data of the toll road lengths was collected from:
BPJT, 2013; World Road Statistics, 1999-2004 (The IRF World Road Statistics 2006);
2001(The IRF World Road Statistics 2001); 2007 (The IRF World Road Statistics 2007); CIA
Factbook, 2012; population data (data for 1995-2010 from Penn Tables (the Penn World
Table, 2012); for 2011 from UNFPA 2011; for 2012 from UNFPA 2012, GDP data (The
World Bank, 2012).
3. Literature Review of Toll Roads
In road infrastructure planning, Fuller and Morency (2013) used a population approach to
reduce the exposure of motor vehicle traffic volumes for all road users, which may greatly
reduce the total number of transportation fatalities. According to Ament, Clevenger, Yu, and
Hardy (2008), as taken from the TRB/Transportation Research Board (2002), it stated that the
national economy and population were expected to pose major new challenges for
transportation and the environment. Yue, et al. (2003) used data population and road
infrastructure as a scenario of spatial population distribution that could be developed as a
basic forecast of population in China. Moller-Jensen and Knudsen (2008) used population
data to gain a proximity index for locality; i.e. value was measured by network roads between
one locality and other urban localities. Orenstein and Hamburg (2010) used population
growth to measure land use/land cover, including roads in Israel.
Meanwhile, researchers who used GDP data as a factor that affects the increase of toll road
development are as follows: Newel and Peng (2008) by citing OECD (2006) and the World
Bank (2006) mentioned that developed countries allocated an average of 2.2% of the GDP to
infrastructure, while developing countries invested 7% of the GDP to the new infrastructure
(Newell, Peng, 2008, PG. 22). This statement was supported by Oronje, Rambo, and Odundo
(2014) in citing Heggie (1995) and Brushett (2005), who mentioned that road transportation
is very important for socio-economic development, because it is able to provide the
necessities between center production and the market; it makes the mining, industrial, and
agricultural sectors become more valuable. Road transportation also facilitates access to the
workplace, as well as education, health, social and leisure facilities. As mentioned by Heggie
(1994; 1995) and Brushett (2005), the quality of road networks can determine the ability of
cost transportation to be affordable, commodity price basis, and the quality of life, especially
for people who have low incomes. In this case, an efficient transportation system can
minimize transportation costs, so that it can support economic growth by promoting
production and trade (Oronje, Rambo, Odundo, 2014, p.75).
According to Nufez (2008), in the aggregated transportation demand forecast model,
individual mobility and income are represented by traffic and gross domestic product (GDP).
Mobility generates traffic and it is supposed that GDP growth also increases purchasing
power ability. In economics, this relationship is represented by an elasticity of traffic with the
leads on GDP usually greater than one (Nujiez, 2008, p.258).
For example, in India, as described by Lakhmani and Sikroria (2012), economic growth in
India, especially in the previous three years, was robust in some matters. Economic growth
has accelerated and now boasts an average growth of 8% annually. The economic growth is
supported by investments in infrastructure such as road transportation, railways, air and
water, power and telecommunications, and also water supply and irrigation. During the
Eleventh Plan (2007-2012), the GDP grew 4.6% to 7-8%, with nearly US $320 billion
expenditures spent in the period of the Plan (Lakhmani, Sikroria, 2012).
Another example is in China. According to Wang (2011), it has most of the toll roads in the
world. One of which is the longest in China, namely from 140,000 km in the world, 100,000
km is in China. He stated that toll roads greatly influence the economic development in
China. Based on data, China's spending on logistic services amounted to 3.3 trillion RMB
($570 billion) in the first half of 2011 (Wang, 2011, p.1303).
To get a description of the ideal length of toll roads in Indonesia, namely compared with 21
countries in the times series and cross-country, a panel data econometric method is used.
According to Gujarati (2004), econometrics means “economic measurement”. Baltagi (2001)
said that the term “panel data” refers to the pooling of observations on a cross-section of
countries, households, firms, etc. over several time periods. This can be achieved by
surveying a number of households or individuals and following them over time.
Various researchers have conducted studies related to the use of the econometric model in
infrastructure; they are: Khanam (1996), who used the Cobb-Douglas and translog functional
forms; Zhu and Tan (2000) who examined the causal relationship between the intensity of
FDI inflow and the increase of technical efficiency; Globerman and Shapiro (2003) who
analyzed the statistical importance of governance infrastructure; Tomljanovich (2004) who
examined one possible source of growth and per capita output levels; Chien and Hsiau (2005)
who researched an empirical analysis of factors that have influenced economic growth;
Aristovnik (2006) who investigated the empirical link between the current account and the
fiscal balance; Lall (2006) who examined the contribution of publicly supplied infrastructure;
Henderson and Kumbhakar (2006) who estimated the returns to capital, employment, private
and public capital in gross state product; Mohapatra and Giri (2009) who explored the
relationship between economic development; Amdal, Bardsen, Johansen, and Welde (2007)
who studied the ease of planning new toll projects by a panel data; Marquez, Ramajo, and
Hewings (2010) who investigated databases where the number of cross-sectional units is
small for a typical panel of data; Gompert and Buerkle (2011) who developed a hierarchical
Bayesian model to quantify the genome-wide population structure and identify candidate
generic regions affected by selection; Sin-Yu (2011) who examined the causal relationship
between financial development and poverty reduction; and Hamilainen and Malinen (2011)
who estimated the elasticity of private production with respect to public capital in a regional
framework.
4. System Dynamics Modeling for Toll Road Development
As described in the methodology above, the first step to construct System Dynamics
modeling is to identify and define problems and then create a System Conceptualization.
4.1 System C onceptualization
According to Lyneis and Ford (2007), a System Dynamics methodology is used as the form
of a model conceptual structure and the approximate chronological order of development.
The model structure was followed by some typical project behavior. Then by citing Roberts
(1964, 1974), an understanding of management in the project conditions at the conceptual
phase is the perception gaps — differences between perceived progress and real progress, and
between perceived productivity and real productivity.
Thus, the perception gap is physical gaps in Indonesia toll road length, which explains toll
road length forecasts by looking at possible toll road lengths and comparisons to 21 countries.
According to Vaishnavi and Kuechler (2008), a physical gap is a comparative description
between ideal situations or desirable ones, so a “vision” needs to be created.
The physical gaps in toll road infrastructure in Indonesia are compared between toll road
lengths based on current situations and ideal (desired) toll road lengths. These gaps are
analyzed using data time series starting from 1995 until 2012 and then forecasted until the
year 2040. To determine the potential gap (output gap), according to Chaudhry (2010), an
econometric data panel is used with required data time series and cross-country. According to
Gujarati (2003), the panel data regression model is a combination of features between time
series data and cross-sections, because it provides “more informative data, more variability,
less collinearity among the variables, more degrees of freedom, and more efficiency”. The
use of statistical and econometric data panel tools can be used for the length and distribution
of a delay. The other reason is that the econometric technique is designed to spell out the time
in most business and economic data being reported regularly, separating intervals such as
quarterly, bi-annually, or annually. System Dynamics models are always developed for
sustainable times. The second reason mentions that a regression equation for a lag has fixed
lag weights, implying a fixed delay time. The structure of information and material delay
used in system dynamics responds to changes in delay times (Sterman 2000, p.439).
Econometric panel data can be used in a system dynamics model. Then to get the desired
results, it requires the determination of cross-country data that is taken based on countries
that have the same income per capita or appropriate with Indonesia’s position based on World
Bank standards for lower income and upper income. As a result, there are 21 countries such
as: Algeria, Morocco, Nigeria, South Africa, Tunisia, Argentina, Bolivia, Chile, El Salvador,
Guatemala, Jamaica, Mexico, Panama, China, Indonesia, Malaysia, Pakistan, the Philippines,
Thailand, Iran, and Syria. In addition to toll road length data, data related to population and
GDP should be collected in each country starting from 1995 to 2012. According to Sharma
and Vohra (2009), toll road length, population, and GDP data among those countries can be
used as a comparison of data, where GDP data is taken to describe the economy of a country,
and the population data is taken to represent the number of people in a country. Furthermore,
the toll road length time series data for 21 countries can be seen in Table 2.
To find the physical gaps above, an analysis of the current situation is carried out by using
data about the length of a single Indonesia toll road. Meanwhile, to get an ideal or desired toll
road using the panel data method, one should take into account 21countries as times series
and cross-country. This method will result in a linear regression model and data panel model.
So, this model will show that the development of toll roads in Indonesia will be significant.
To determine the regression model with single data, time series data of toll road length in
Indonesia can be used starting from 1995 until 2012 and then forecasted using ARIMA
(Autorefressive Integrated Moving Average) model. Then the process to determine the best
data panel is as follows: 1) collect data time series for 21countries as cross-country; 2) make
a common structure model of each country considered the same and without any effect; 3)
make a random structure model, which is indicated by each country as different individuals;
and 4) make a random model, which is indicated by each country as a unique individual. It is
further analyzed to obtain the ideal structure regression model. Then the hypothesis is tested
to finally get the best model. The steps are shown in Figure 2.
YEAR
109
Algeria Congo, De Gabon Morocco Nigeria South Affi Tunisia Argentina El Salado GuatamelsJamaica
_{Toll ARTOLL cfTout Gx STOLL MATOUL NATOLL: STOLL TUTOLE ra sero. arin CHTOUL FITOUL GUPOLL rar M rae Pen mr
os 30 1142, 567 6368,
as eee sat ” e407
4030 0 219s 57 as 77 ons
4030 0 0 7 1194 200 ™ 0 638 oF 6289
4030 30 30a t94 2082 ™ 1 638 4 6289
4030 3030399 4 629
4030 30304529 4 6129
4030 30306? 638 4 oT
4030 0 467 118M OT 638 4 0987
os 30 3 SO) SIS 9S 638 m4 3B 6lad
os 30 303063) 638 4 4463
os 30 Ee ee en ee) 638 m4 44 Ist
os 30 0 0 7 119892 638 i re
4s 0 RSLS OT 638 4 4 6I3L
os 0 WO 86MM OT 38 m4 4 BL
ois 0 3030 ROHN 39 DTS 638 4 4 oI3t
os 0 3030 R66 SIH 638 4 4 6st
India
sis
53265
20074192
‘Table 2: Toll Road Length Time Series Data for 21 Countries from 1995-2011
Common structure
model
Data
Random structure
model
Fixed structure
model
‘The best
model?
ja Mabiysin Pakistan Phiippines Thailand Iran, Ishur Lebano
¥TOLL 1 wv "NTOUL BjPOLL PATOL PHTOLL TOLL TOLL arm 7
S424 06.85 ®
Yes->|
Testing model
—}
Regression
structure model
Figure 2: Steps of Panel Data.
4.1.1 Single Regression Model
To obtain single data regression results, an ARIMA (Auto Regressive Integrated Moving
Average) linear method is used. ARIMA models are, in theory, the most general class of
models for forecasting a time series which can be stationary through differencing and
transformations such as logging. In fact, the easiest way to think of ARIMA models is as
fine-tuned versions of random-walk and random-trend models: the fine-tuning consists of
adding of the lags to be differenced series and/or lags of the forecast errors to the prediction
equation, as needed to remove any last traces of autocorrelation from the forecast errors
(Introduction to ARIMA, 2014).
Forecasting with an ARIMA model example with the ARIMA model (0, 1, 1) (0, 1, 1) is
outlined as follows:
(1-B)(1-B 12)X, = (1 ~ 0;B) (1-O,B"”) e, o)
But to use it in forecasting, it requires a translation of the regression equation to make it more
general. The above model is in the following form:
X, =X 1 + X12—Xi-1 + er — Ope, 1— Ol ey-12 + OO 11-13 (2)
Using the Crystal Ball ® tool and data input of toll road length in 1995-2011, the best
ARIMA equation (1, 1, 2) is obtained as follows:
TOLL_INA_As_ Usual = 0.9767* TOLL_INA,;_; + 1.39* YR,1-0.765 (3)
where:
TOLL_INA_As Usual is toll road lengths in Indonesia, TOLL_INA,.; is toll road lengths
from the previous year, while YR,.; is a dummy variable from the previous year.
4.1.2 Panel Data Model for Toll Roads
By combining the time series of a cross-section of toll road lengths, GDP, and population in
21 countries (Algeria, Morocco, Nigeria, South Africa, Tunisia, Argentina, Bolivia, Chile, El
Salvador, Jamaica, Mexico, Guatemala, Panama, China, Indonesia, Malaysia, Pakistan, the
Philippines, Thailand, Iran, and Syria), we can use panel data. A simple dynamic panel data
model with heterogeneous coefficients is presented as Gujarati (2004) has outlined:
Vir = Bi + BoXoi + P3Xait + wir (4)
This depicts where i stands for the ith cross-sectional units and ¢ for the th time period.
For an empirical illustration, Baltagi and Levin (1992) estimate a dynamic model of the
demand for cigarettes based on panel data from 33 American states over the period of 1963-
87. The estimated equation is (Baltagi, 2001):
InCi= a. + BilnC;, 21 + BolnPirt B3ln Viet BslnPict uit (5)
This depicts where the subscript i denotes the ith state (i = 1, ..., 46), and the subscript ¢
denotes the fth year (¢ = 1, ..., 26). Cjis real per capita sales of cigarettes by persons of
smoking age (14 years of age and older).
So, in the same way as the equation model (2), we can write the structure model for toll road
length as follows:
Intolli: = & + Brit Ingdpi + Bi2lnpopit + eit (6)
where:
Injoll: toll road lengths (km), In,gdp: GDP (millions of dollars), Inop: population (millions
of people), a; = intercept, £; , 6 until,= slope, e:= stochastic disturbance term and i =the
number of individuals. Notice that a and B vary individually for all. This model indicates that
the parameters for a and B are various.
So the data panel of the toll road length equation in Indonesia is:
LNTOLL_INA =-4.493-0.223 * LNYR + 0.246 * LNPOP_INA + 0.750 * (7)
LNGDP_INA
or
TOLL INA Not As usual = EXP (-4.493)*YR_IND®?3* POP_ INA°*4**GDP_ (8)
INA*™
where:
TOLL_INA_Not As usual is the toll road length in Indonesia by comparing it with 21
countries, POP_INA is the total population in Indonesia, GDP_INA is Indonesia's GDP, and
YR_IND is a dummy variable of that year.
By combining Eq. (3) and (8) above, then a physical gap of toll road lengths in Indonesia
until 2040 is forecasted as in Figure 3 below.
Physical Gap: As Usual vs Not As
Usual
—TOLL_INA_AS_USUAL
——TOLL_INA NOT AS
USUAL
Toll Road Length (km)
a
r=)
8
8
a Amana
a aNqnm
a Sesooog
a aAaqaqq
2011
2015
2039
on
So
os
aq
1995
Figure 3: Physical G ap of Toll Roads between As Usual and Not as Usual
From Figure 3 above, it describes that forecasting as usual is based on Eq. (3) and not as
usual based on Eq. (8). In Figure 3, it states that with not as usual, after calculating for the
year 2012, Indonesia toll road length in 2040 should be 9,549 km, but with as usual Indonesia
toll road length will achieve 1,269 km, so the physical gap in the year 2040 is 10,357 km.
From the results of the physical gap above, it can be explained that toll road development in
Indonesia is required due to the needs of population and GDP. But in fact, the need for the
toll road length in Indonesia as reported by the BPJT (2013) is 2,213.83 km length, where the
toll roads operated are 784.06 km, establishment of toll roads is 874.79 km length, tender
process is 159.77 km, and tender preparation is 395.21 km length. So, if compared with the
results of Eq. (8), then a toll road length of 2,213.83 km should have been achieved in 2007.
Unfortunately, this plan was never realized. So, the main problem of this study is that there is
a gap between the current situation and the desired toll road length in Indonesia.
4.2 Formulation of System Dynamics Model for Toll Road Development
After knowing the problems of the development of toll roads in Indonesia with a physical gap
analysis, a System Dynamics model is formulated.
4.2.1 Causal Loop Diagram
According to Sterman (2000, p.179), there is a need for new road construction because of
congestion and delays. The construction of toll roads in Indonesia aims to streamline traffic
in areas that have been developing, improve the distribution of goods and services to support
economic growth, promote equitable development and justice outcomes, as well as ease the
burden of government funding through the participation of road users (BPJT, 2013). So as the
mental model of the outline of toll road development in Indonesia, it can be described in
Figure 4.
Toll Road
——_ cea
A
2c ‘ + t
Accessibility Toll Revenue
.
.
Saving Cost
Operation Car 4 Productivity Toll Road Investor
"
a
Economic Growth
Figure 4: Conceptual Mental Causal Effect Diagram of Toll Road Investment
Development in Indonesia
From Figure 4 above, there are two reinforcing loops: loop R1 and R2. Loop R1 explains that
with toll road construction in Indonesia it is expected to facilitate accessibility and traffic in
developing areas, increasing the effectiveness and distribution of goods and services, so that
the toll road users will benefit from vehicle operating costs. This will increase productivity
and finally support economic growth. Loop R2 explains that with the construction of toll
roads in Indonesia, it will get toll revenues, so that with toll revenues the corporations will get
investment returns, depending on the certainty of traffic volume and toll rates, which will
finally support Indonesia's economic growth.
4.2.2 Stock and Flow Diagram
According to Martinez-Moyano and Richardson (2013, p.115) to make a model formulation,
it uses two approaches. It starts small, adds complexity as necessary, and uses realistic
operational thinking. Thus, in this paper the process is focused on the realities of the toll road
investment process in Indonesia and the process of toll road construction, according to BPJT
(2013). The toll road investment process can be illustrated as shown in Figure 5.
PREQUALIFICATION 5 mN
Ub :
Sd
WINNER DECISION
ae
\7
ESTABLISHEMENT OF TOLL ROAD COMPANY
wy
TOLL ROAD CONCESSION AGREEMENT
WZ
BANK FINANCING AGREEMENT
BIDDING PROCESS iaMontHs
Source: BPJT, 2013
Figure 5: Indonesia Toll Road Investment Procedures
Based on Figure 5 above and the process of toll road construction in Indonesia, it can be
described in the Stock and Flow Diagram as illustrated in Figure 6.
oD Tver Toon Paces
88 Land a
caus aytaing fraction
‘ave Land Acquistion
Time per km
Establehemens
Bing process
Tae pe km
Figure 6: Stock and Flow Diagram of Indonesia Toll Roads
From Figure 6 above, the process of the wait establishment of the toll road company requires
pre-qualifications, a bidding process, and a winner decision; then there is a toll road
concession agreement between the Indonesia Toll Road Authority and the company. Then,
the equation of the prequalification rate is as follows:
MIN (Wait _ Establishment_of_Toll_Road_Company * Correction _ (9)
Performance_Effect/Time_Standard, Early Finish _ Time * Correction _
Performance_Effect)
Meanwhile, the Correction Performance Effect is an effect that is obtained based on the
Correction Performance that affects the Prequalification Rate and Re-Prequalification Rate
and Bidding Process Rate, as seen at Figure 7.
Graphical Function
1 Conecton
To0 Too
100 0030
Conection 0.200 1060
Petfonanc 0.300, 0105
eect 0.400, 0130
0500 460
0.600 0745
0700 0965
00 0320
0.300 0.985
(0000 ects 4.000 4.000
— _——
<= fico
a Years DataPoints: [11 ]
eaoupae
Figure 7: Correction Performance E ffect
Next, to run the toll road concessioner, the company must show their 'bankable' financials to
BPJT, and then the company performs land acquisitions and toll road construction.
4.3 System Dynamics Model Result
The system dynamics model results are as described in Figure 6 above with the stock. It is
started with the toll road length in 2013 at 784.06 km (BPJT, 2013), Ave Land Acquisition
Time per km= 5/30 year and Ave Construction Time per km = 2/25 years, with the Toll Road
Building Fraction = 0.0456. Then the toll road length in the year 2040 is 1,488.04 km.
Meanwhile, land acquisitions will be executed starting in 2014 with 2213.83-784.06 =
957.843 km, so that in 2014 the remainder is 948.41 km.
WP 1: Land sequstion Process 2: Tod Road Bung
F
La
4 e254... ae
-" ——
4 a I
f 13.10 200 2047 80 2085.20 2082.80 2100;
mt jn
2 Unetied
Figure 8: Indonesia Toll Road Infrastructure System Dynamics Model Results
If followed, land acquisition will be completed in the year 2067, but the toll road plan which
is planned by the government for 2213.83 km will be reached in the year 2080.
Meanwhile, as can be seen in Figure 9, with the same input data above but using a leverage
option, for example if the leverage is 2 times, land acquisition will be completed in the year
2032, but the toll road plan which is planned by the government for 2213.83 km will be
reached in the year 2054. If leverage is 3 times, land acquisition will be completed in the year
2024, and the toll road plan which is planned by the government for 2213.83 km will be
reached in the year 2043.
® TonRced sutseg 2 Lond Acton Process
5: TollRonterevarge Land eg. te kovege (QE Tod Rend Bung 2: Land Adan Process 3 Tok Rea. Tet levenge © Lone hcg. evenege
———T
a es
»§ SEL » BEE
? ?
Figure 9: Indonesia Toll Road Infrastructure System Dynamics Model Results after
Leverage
5. Conclusion
The problems of toll road conditions in Indonesia by using a physical gap analysis between
the existing toll road length and estimated econometric panel data foretell a large difference
in 2040, as long as 10,357.580 km. It proves that the construction of toll roads in Indonesia
still needs to be accelerated by up to 8 times, because toll road development is based on
population and GDP by comparing it with 21 countries.
If simulated by using real toll road construction in Indonesia, with Ave Land Acquisition
Time per km = 5/30 year and Ave Construction Time per km = 2/25 years, with the Toll
Road Building Fraction = 0.0456, then the toll road length in 2040 is 1,488.04 km.
Meanwhile, the land acquisitions that will be executed starting in 2014 are 2213.83-784.06 =
957.843 km; so in 2014 the remainder is 948.41 km. Therefore, to accelerate toll road
development in Indonesia, there must be leverage.
References
Amdal, E, Bardsen, G, Johansen, K, Welde, M. 2007. Operating Costs in Norwegian Toll Companies:
A Panel Data Analysis, Transportation Vol. 34:681-695.
Ament, R, Clevenger, AP., Yu, O. and Hardy, A. 2008. An Assessment of Road Impact on Wildlife
Populations in U.S. National Parks, Environmental Management, Vol. 42: 480-496, DOI
10.1007/s00267-008-9112-8.
Aristovnik, A. 2006. The Influence of Fiscal Policy and Private Ii on External Imbalances in
Transition Economies, Economic and Business Review for Central and South — Eastern Europe, Apr.
Aschauer, DA.1989. Is Public Expenditure Productive? Journal of Monetary Economics 23 (1989)
177-200, North-Holland.
Baltagi, BH. 2001. Econometric Analysis of Panel Data, John Willey & Sons, Ltd.
Board of Toll Road Authority (BPJT). 2007. www.bpjt.net. Accessed on October 15, 2007.
Chaudhry, AA. 2010. A Panel Data Analysis of Electricity Demand in Pakistan, The Lahore Journal
of Economics, Vol. 15.
Chien-Hsun C, Hsiau-Ling Wu. 2005. Determinants of Regional Growth Disparity in China's
Transitional Economy, Journal of Economic Studies, 32, 5/6.
CIA Factbook, 2012.
Fuller, D. and Morency, P. 2013. A Population Approach to Transportation Planning: Reducing
Exposure to Motor-Vehicles, Journal of Environmental and Public Health, Vol. 2013, Article ID
916460.
Gompert, Zachariah and Buerkle, Alex, C. 2011. A Hierarchical Bayesian Model for Next-Generation
Population Genomics, Genetics, March 187:903-917.
Government of Ind ia. 2011. Mastery Acceleration and Expansion Economic De
2011-2025 (MP3El).
Globerman, S. and Shapiro, D. 2003. Governance Infrastructure and US Foreign Direct Investment,
Journal of International Business Studies, Jan: 34, 1. p.19.
Gujarati, DN. 2004. Basic Econometrics, Fourth Edition, McGraw-Hill.
Hamiilainen, P and Malinen, T. 2011. The Relationship between Regional Value-Added and Public
Capital in Finland: What do the New Panel Econometric Techniques Tell Us, Empirical Economics
40:237-252.
Henderson, DJ. and K, SC. 2006. Public and Private Capital Productivity Puzzle: A Nonparametric
Approach, Southern Economic Journal, 73(1), 219-232.
Hua, W. 2011.High Logistics Cost, Toll Road and Institutional Factors Countermeasure in China,
Journal of Modern Accounting and Auditing, ISSN 1548-6583, Nov. Vo. 7, No. 11, 1301-1306.
International Road Federation, 2001, The IRF World Road Statistics 2001, IRF Geneva
International Road Federation, 2006, The IRF World Road Statistics 2006: Data 1999-2004, IRF
Geneva.
International Road Federation, 2007, The IRF World Road Statistics 2007, IRF Geneva.
Khanam, BR. 1996. Highway Infrastructure Capital and Productivity Growth: Evidence from the
Canadian Goods-Producing Sector, Logistic and Transportation Review, Sep, 32, 3.
Lakhmani, P, Sikoria, R. 2012. Infrastructure Financing Instruments with a Special Emphasis on
Highway and Roads, International Journal of Management Research and Review, Vol. 2, Issue 9,
Article No-20/1668-1677, ISSN:2249-7196.
Lall, SV. 2006. Infrastructure and Regional Growth, Growth Dynamics and Policy Relevance for
India, Development Research Group, The World Bank.
Lyneis, JM. and Ford, DN., 2007. System Dynamics Applied to Project Management: A Survey,
Assessment, and Directions for Future Research, System Dynamics Review, Vol. 23, No. 2/3
(Summer/Fall); 157-189, p.159.
Martinez-Moyano, I J. and Richardson, GP.2013. Best Practice in System Dynamics Modeling,
System Dynamics Review Vol. 29, No. 2 (April-June 2013): 102-123, p.108.
Marquez, MA., Ramajo, J, Hewings, Geoffrey J.D. 2010. A Spatio-Temporal Econometric Model of
Regional Growth in Spain, Springer-Verlag.
Mohapatra, G and Giri, A.K. 2009. Ec ic De and Envii ! Quality: An
Econometric Study in India, Management of Environmental Quality: An International Journal, Vol.
20, No. 2.
Moller-Jensen, L and Knudsen, MH. 2008. Pattern of Population Change in Ghana (1984-
2000)Urbanization and Frontier Development, Geography Journal, Vol. 73: 307-320, DOI
10.1007/s10708-008-9209-x.
Newel, G, Peng, HW. 2008. The Role of U.S. Infrastructure in Investment Portfolios, Journal of Real
Estate Portfolio Management; Jan-Mar; 14, 1; ABI/INFORM, p.28.
Nujiez, A. 2008. Estimating the Functional Form of Road Traffic Maturity, Network Spatial
Economic, 8: 257-271, DOI 10.1007/s11067-007-9049-0.
Oronje, DO., Rambo, CM., Odundo, PA. 2014. Agency Level A of Road Mai Levy
Fund: Evidence from Kenya, Global Journal of Business Research, Vol. 8, No. 1.
Orenstein, DE. and Hamburg, SP. 2010. Population and Pavement: Population Growth and Land
Development in Israel, Population Environmental, Vol. 31: 223-254, DOI 10.1007/s11111-010-0102-
4.
Sharma, A.K. and Vohra, E. 2009. Critical Evaluation of Road Infrastructure in India: A Cross-
Country View, Engineering, Construction, and Architectural Management, Vol. 16, No. 1.
Sin-Yu Ho. 2011. Finance and Poverty Reduction in China: An Empirical Investigation, The
International Business & Economics Research Journal, Aug: 10, 8.
Sterman, JD. 2000. Business Dynamics: System Thinking and Modeling for a Complex World, Irwin,
McGraw-Hill.
Sterman. JD. 2001, System Dynamics Modeling: Tools for Learning in a Complex World, California
Management Review, Vol. 43. No. 4, Summer.
Sutherland, D., Araujo, S., Balazs Egert, and Tomasz, K, 2009, Infrastructure Investment: Link to
Growth and the Role of Public Policies, Economic Department Working Paper No. 686, Organization
for Economic Cooperation and Development (OECD).
Tomljanovich, M. 2004. The Role of State Fiscal Policy in State Economic Growth, Contemporary
Economic Policy, Jul, 22, 3.
The University of Pennsylvania 2008. The Center for International Comparisons at the University of
Pennsylvania. Penn World Table. Retrieved June 12, 2008.
The World Bank, 1999, Impact of the Asian Financial Crisis on Toll Road Development in Selected
Asian Countries.
The World Bank, Ministry of Construction Japan. 1999. Asian Toll Road Development Program:
Review of Recent Toll Road Experience in Selected Countries and Preliminary Tool Kit for Toll Road
Development.
Vaishinavidan, VK., Kuechler Jr. W. 2008. Design Science Research Methods and Patterns:
Information and Ce i ry, Auerbach Publicati New York.
Weixin, LUAN, Hang, CHEN, Yuewei, WANG. 2010. Simulating Mechanism of I ion between
Ports and Cities Based on System Dynamics: A Case of Dalian, China, China Geography Science,
Vol. 20. No. 5.
Yue, TX; Wang, YA; Chen, SP; Liu, JY; et al. (2003) Numerical Simulation of Population
Distribution in China, Population and Environment; Nov. 2003; 25, 2; Proquest.
Zhu, G and Tan, KY. 2000. Foreign Direct Investment and Labor Productivity: New Evidence from
China as the Host, Thunderbird International Business Review, Sep.
Introduction to ARIMA: Non-Seas 1 Models, http: le.duke.edu/~rnau/41 larim.htm_cited
January 30, 2014.
JasaMarga.Jagorawi. Retrieved Aug 7, 2012, from http://www.jasamarga.com/en_/layanan-jalan-
tol/jagorawi.html.
http://www. bpjt.net/main.php?s id=jartol&parentid=2. id=13&strlang=id, accessed on
December 26, 2012.
http://bpjt.net:8802/websit in.php? id=jartol&parentid=4. id: strlang=id, accessed on
October 29, 2010.
https://www.cia.gov/library/publications/the-world-factbook/geos/ch.html accessed on_ December 15,
2012.
http://englist ledaily.com.cn/200612/30/eng20061230_337130.html, accessed on October 20,
2007.
http://en.wikipedia.org/wiki/List_of expressways and_high s_in Malaysia, accessed on October
29, 2007.
http://bpjt.pu.go.id/konten/progress/t i_accessed on March 4, 2014.
http://pwt.econ.upenn.edu/php_site/pwt62/pwt62_form.php.
http://www.unfpa.org/public/.
http://databank. worldbank.org/ddp/home.do?Step=2&id=4&hActiveDimensionId=WDI Series.
http://bpjt.pu.go.id/konten/progress/t i_accessed on March 4, 2014.
http://bpjt.pu.go.id/konten/jalan-tol/tujuan-dan-manfaat accessed on March 4, 2014.
