Journal of Sustainable Forestry

ISSN: 1054-9811 (Print) 1540-756X (Online) Journal homepage: http://www.tandfonline.com/loi/wjsf20

Determinants of illegal logging in indonesia: An

empirical analysis for the period 1996-2010

Yaoyao Ji, Ram Ranjan & Chi Truong

To cite this article: Yaoyao Ji, Ram Ranjan & Chi Truong (2017): Determinants of illegal logging
in indonesia: An empirical analysis for the period 1996-2010, Journal of Sustainable Forestry, DOI:
10.1080/10549811.2017.1369436

To link to this article: http://dx.doi.org/10.1080/10549811.2017.1369436

Accepted author version posted online: 28

Aug 2017.

Submit your article to this journal

Article views: 1

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=wjsf20

Download by: [University of Michigan] Date: 01 September 2017, At: 06:56

 Determinants of Illegal Logging in Indonesia: An Empirical
Analysis for the Period 1996-2010
YAOYAO JI, 1* RAM RANJAN, 1 and CHI TRUONG 2
1
Department of Environment Sciences, Faculty of Science and Engineering, Macquarie
University, Australia
2
Department of Applied Finance and Actuarial Studies, Faculty of Business and Economics,
Macquarie University, Sydney Australia

t
Illegal logging is a significant problem in Indonesia, which is one of the few countries

ip
with a large forest area. In this paper, we study factors that affect harvesting and

cr
supply of illegal timber from Indonesia to China and Japan. We also look at factors
Downloaded by [University of Michigan] at 06:56 01 September 2017

that lead to demand of Indonesian illegal timber from China and Japan. A

us
simultaneous-equation econometric model of illegally logged timber demand and
supply is developed and tested using annual data over the period 1996-2010. We
find that corruption and decentralization in Indonesia have significant and positive
an
impact on illegally logged timber supply while excess demand in Japanese
construction and furniture industries as well as Japan’s housing starts are significant
M
factors affecting illegal logging in Indonesia. Law enforcement or policies aimed at
reducing illegal harvesting in Indonesia are found to be more effective than policies
targeting the import of illegal logged timber into Japan and China.
ed

Keywords. Illegal logging in Indonesia; Illegal timber trade; Simultaneous-equation

model; Deforestation; Illegal timber demand
pt
ce
Ac

1
 INTRODUCTION

Indonesia is the third largest country in terms of tropical forest area and also a country with the

second highest deforestation rate (FAO, 2010). Each year, a large area1 of forest in Indonesia is

lost, the majority of which is due to illegal logging. Illegal logging has detrimental effects on the

environment, the economy and the society in Indonesia (Obidzinski et al., 2006). The aim of this

t
paper is to study factors that affect illegal harvesting of timber in Indonesia and to evaluate the

ip
effectiveness of policies implemented inside as well as outside Indonesia to reduce illegal

cr
logging.
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
The term ‘illegal logging’ is often used to refer to the cutting of trees in protected forest areas

(Callister, 1999; FFPRI, 2005). However, to study the impacts of different policies that aim at
an
reducing illegal logging, it is necessary to extend the notion of illegal logging beyond harvesting

activity. For example, a policy that restricts the demand of illegally logged timber may help to
M
reduce the incentive to cut down trees illegally. For this reason, we follow the literature and use

the term ‘illegal logging’ to refer to the entire supply and demand process of illegally logged

timber2 (Brack et al., 2002; IGES, 2007; Luttrell et al., 2011). This latter (more general) definition
ed

is especially helpful in studying the determinants and magnitudes of illegal logging in the practical

context that Indonesia trades with multiple partners.

pt

Recognizing the serious impacts of illegal logging, Indonesia has implemented a number of
ce

policies to control forestry related crimes (Obidzinski et al., 2006). A ban on log exports was

implemented for the first time in 1985 by the Indonesian government to promote sustainable
Ac

harvesting and support the development of the domestic wood processing industry (Gellert, 2003;

1
About 498 thousand hectares, according to FAO,2010.
2
For example, in Brack et al., (2002), “Illegal logging takes place when timber is harvested, transported, bought or sold in

violation of national laws. The harvesting procedure itself may be illegal, including corrupt means to gain access to forests,

extraction without permission or from a protected area, cutting of protected species or extraction of timber in excess of agreed

limits. Illegalities may also occur during transport including illegal processing and export, mis-declaration to customs, and

avoidance of taxes and other charges”.

2
 Luttrell et al., 2011). The ban was lifted in 1997 to alleviate the impact of the Asian financial crisis

and was reintroduced in 2001 when the Indonesian economy had recovered. An anti-money

laundering law and an anti-corruption law were introduced in 2003 as indirect measures to

control forest crimes (Luttrell et al., 2011). Two agencies, the Corruption Eradication Commission

and the Financial Intelligence Unit, were set up to monitor the law enforcement by relevant

agencies (Dermawan et al., 2011). These law enforcement agencies have been argued to have

the potential to control illegal logging through a strategy of “follow the money” 3 . In addition,

t
ip
Indonesia and other wood-producing countries have signed multilateral or bilateral agreements

with wood-consuming countries to control illegal logging (Setiono et al., 2005). For example,

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

China and Japan signed Memorandum of Understandings (MoU) with Indonesia to guard against

us
forestry crimes by exchanging information and enhancing cooperation in 2002 and 2003,

respectively (Setiono et al., 2005). However, the exact impact of these bilateral agreements has
an
been challenging to ascertain due to lack of reliable statistics (Luttrell et al., 2011). In 2008, the

US amended the Lacy Act of 1990 to prohibit importing of illegally sourced plant materials and
M
products, which restricted the export of illegal Indonesian timber products to the US (Prestemon,

2015). Further, in 2013, Indonesia developed a timber legality assurance system, known as
ed

SVLK (Sistem Verifikasi Legalitas Kayu), to track the movement of timber from harvesting to

trading stages (Luttrell et al., 2011)4. This system has been considered as an effective way to

curb illegal logging, and also crucial towards Indonesia’s successful signing of a Voluntary
pt
ce
Ac

3
For instance, the President Director of Surya Dumai Group, Marthias, was charged with corruption in the forestry sector, while

the Governor of East Kalimantan, Suwarna Abdul Fattah, was charged with receiving bribes from Marthias (Luttrell et al., 2011)

(Court decision No. 21/PID.B/TPK/2006/PN.JKT.PST on Marthias. Court decision No. 380.K/Pid. Sus/2007 on Suwarna). They

were sentenced to 18 months and 4 years in prison separately (Luttrell et al., 2011). Details of the cases please refer to:

http://www.cifor.org/ilea/_pf/1/_ref/indicators/cases/decision/Martias.htm

http://www.cifor.org/ilea/_ref/indicators/cases/decision/suwarna_AF.htm

4
We would like to thank a referee for pointing out these more recent policies. The impact of these policies is, however, not

captured in our dataset, see the limitation section for further details.

3
 Partnership Agreement (VPA)5 for meeting the requirements of EU Timber Regulation (Luttrell et

al., 2011; Obidzinski et al., 2014).

Factors that affect illegal logging in Indonesia have been analysed in previous studies.

Dieter (2009) estimates illegal timber trade, using an adjusted input-output model and trading

records of multiple countries. International trade is found to increase the supply of illegal timber

while ineffective forestry governance is found to be an important factor that facilitates illegal

t
harvesting activities. Contreras-Hermosilla et al. (2008) suggest that effective controlling of illegal

ip
logging requires both supply and demand related measures. Harwell et al. (2009) provide a

cr
comprehensive report that summarizes illegal logging activities in Indonesia using data collected
Downloaded by [University of Michigan] at 06:56 01 September 2017

from fieldwork as well as information acquired from media, government reports and reports from

us
various NGOs (e.g. FAO, ITTO). It is observed that corruption in the forestry sector is a critical

factor that drives illegal harvesting and the most common corruption activity takes the form of
an
issuing fake documents that allow companies to clear more trees than they are entitled to.

Harwell and Blundell (2009) suggest that to reduce illegal logging and corruption in Indonesia,
M
the forestry management system needs to be improved and the anti-money laundering legislation

needs to be augmented with corresponding bank regulations.

ed

In addition to corruption, ill-designed institutions for forest management also contribute to the

illegal logging problem. Casson et al. (2002) analyze the problem of illegal logging in Indonesia
pt

from the political economy perspective. They suggest that the decentralization process, which
ce

assigns substantial forestry management decision making power to local government,

encourages illegal logging in Indonesia. As local governments are allowed to issue timber
Ac

permits, and also to keep district taxes from these permits, they have strong incentives to permit

excessive logging (Casson et al., 2002). Logging practices permitted by local governments are

observed to be in contradiction with guidelines of sustainable forests development in many cases,

and induce the same detrimental impacts as those by ‘illegal’ logging (Casson et al., 2002).

Mismanagement of logging permits at the local government level is found to reduce the impacts

5
VPA is an essential component of the EU bilateral trade agreements between the EU and its timber exporting partner

countries (Obidzinski et al., 2014).

4
 of the forest law enforcement by the central government. Burgess et al. (2012) analyze the effect

of decentralization on deforestation (including ‘illegal’ logging) in Indonesia. They find that by

increasing the number of districts, the decentralization process, in fact, has accelerated the

deforestation rate. Smith et al. (2003) conduct an analysis by distinguishing between collusive

and non-collusive corruption. Collusive corruption is argued to be the main driver of illegal

logging in the period that follows the fall of President Suharto.

t
Determinants of illegal logging in other countries have also been examined in the literature.

ip
Alemagi et al. (2010) discuss factors that affect illegal logging in Cameroon, including systemic

cr
corruption, poverty, conflicts, licensing society, usurpation of property rights and inadequate
Downloaded by [University of Michigan] at 06:56 01 September 2017

institutional support. Hansen et al. (2008) estimate the amount of illegally logged timber in Ghana

us
from 1996 to 2005 using timber export statistics, domestic timber consumption estimates and

official harvesting records. Around 70 percent of timber harvested in Ghana is found to be illicit,
an
due largely to policy failure and the lack of effective economic incentives for farmers and local

communities to protect forestry (Hansen et al. (2008). In contrast, Nanang (2010) suggest that
M
the main cause of illegal logging in Ghana is the high timber export demand, and to reduce

deforestation, policies are implemented to reduce timber export, which then restricts both legal
ed

and illegal harvesting. Richards et al. (2003) analyze the social, economic, environmental and

governance impacts caused by illegal logging in Honduras and Nicaragua. Besides adversely
pt

impacting on the environment, illegal logging is found to hurt the economy as firms that harvest

and process legal timber find it difficult to compete with illegal operations. Richards et al. (2003)
ce

suggest that to reduce illegal logging, measures curbing corruption and incentive schemes for

improving forest management are required. McElwee (2004) examines the problem of illegal
Ac

logging in Vietnam and finds that creating personal incentives for forest protection and

management is an effective method for controlling illegal logging.

Despite the large number of studies on illegal logging, no study has analysed factors driving

illegal logging within a multiple market framework using econometric models. The impacts of

domestic and bilateral policies have not been formally estimated either. In this study, we develop

a simultaneous-equation model to estimate factors that affect illegal logging. The model

5
 incorporates both demand and supply characteristics of logging. We select Japan and China,

which are Indonesia’s main timber trading partners and also the suspected main illegal timber

importers, to represent the demand side of Indonesia’s illegal logging. Different from the standard

simultaneous-equation model, we analyse the illegal timber demands from Japan and China in

two separate sets of equations to allow the incorporation of country specific factors.

In addition, we test the impacts of key policies that were introduced by the Indonesian

t
government to reduce illegal logging activities, including the ban of log exports as well as the

ip
anti-money laundering and anti-corruption laws. The effectiveness of forest related policies

cr
adopted in importing countries is also examined. We also test the effectiveness of bilateral
Downloaded by [University of Michigan] at 06:56 01 September 2017

agreements, which forbid the trade of illegally logged timber, in both the supply and demand

us
equations of our model.
an
The remainder of the paper is organized as follows. A discussion on the data used for this

study is provided in next section, following which, we outline the modelling framework and the
M
testing procedures. Estimation results and policy implications are discussed next. In particular,

we first present the results and analysis of factors causing illegal logging, then we discuss the

impact of political factors in Indonesia. The impact of the green purchase laws on curbing illegal
ed

timber import from Japan are analysed next. We also discuss some key limitations of our

approach. The conclusion summarizes the main findings of the paper.

pt
ce

ILLEGAL LOGGING DATA FOR INDONESIA

Ac

The complex underground operation of timber smugglers makes it almost impossible to obtain

accurate information for quantifying illegal logging and illegal trade. Official statistics may reflect

a small portion of illegal logging activities that are discovered, and are therefore not reliable,

while most quantitative reports on illegal logging provide estimates based upon ‘speculation and

anecdotal information’ (Burgess et al., 2012). Among frequently cited sources, Palmer (2001)

uses a materials balance model to estimate illegal logging by comparing the supply and demand

of the round wood in Indonesia. The volumes of illegal logging in Indonesia in 1997 and 1998

6
 were estimated to be 49.176 million m3 and 64.612 million m3 of round wood equivalent,

respectively. The Indonesian Ministry of Forestry provides estimates of illegal timber supply over

the period 1980 to 2005 based on the differences between the official round wood production

and consumption in the processed wood industry (Manurung et al., 2007). The highest amount of

illegal round wood was estimated at 42.2 million m3 in 2002 and fell to 20.3 million m3 in 2005

(Manurung et al., 2007). Harwell et al. (2009) use the difference between legal timber supply and

the total timber consumption to estimate the level of illegal timber supply in Indonesia. Their

t
ip
estimates indicate that the level of illegal timber produced in the country over period 2003-2006

is about 30 million m3 per year (Harwell et al., 2009).

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

After introduction of the log export ban in the mid-1980s, log exports from Indonesia were

us
reduced significantly while export of other wood products such as plywood, increased rapidly

(Makkarennu, 2013). Indonesia became the largest supplier of plywood in 1997. Fig. 1 depicts
an
Indonesian plywood production and exports between 1996 and 2010. It shows that most of the

plywood produced in Indonesia (on average 87.60 percent) is exported, and a large proportion
M
(around 40 percent) is exported to Japan and China (Fig. 2). These are the main importers of

Indonesian plywood, accounting for around 40 percent of the total plywood exports from
ed

Indonesia during the period 1996-2010.

The ITTO (1997-2012) report provides information on the annual timber trade in the world
pt

based on the data provided by individual countries. However, for the same trading channel, the
ce

quantity reported by the exporting country is often different from the quantity reported by the

importing country. For example, in 2002, the export of plywood to Japan was reported by
Ac

Indonesia to be 1.956 million m3, while the plywood import from Indonesia was reported by

Japan to be 2.637 million m3, a difference of 0.681 million m3. Similar discrepancies were also

observed between China and Indonesia’s trading records. In the report by ITTO (1997-2012),

these discrepancies have been attributed to different measurement approaches and timber

smuggling.

The gap between Japan’s timber imports from Indonesia and Indonesia’s timber exports to

Japan is presented in Fig.3. An account of similar discrepancies in timber trade between China

7
 and Indonesia6 is presented in Fig.4. The unexplained gaps increase after 1998 in Fig.4 and after

1999 in Fig.3, peak in 2001 and decline thereafter in both Fig. 3 and 4. If these gaps are due

purely to different measurement approaches used by different countries, the gaps between

trading records should be constant over time. In terms of the fluctuations in gaps observed in

Fig.3 and Fig.4, a possible explanation could be the ongoing trade in illegal timber. It is apparent

from Fig.3 and Fig.4 that most of the recorded timber trade in the importing countries, Japan and

China, are higher than the recorded exports from Indonesia. It seems that Indonesia under-

t
ip
reports the actual trade records, which consistently exceed the annual forest harvesting

concessions issued by the government. In addition, the annual discrepancies observed between

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Japan and Indonesia, and China and Indonesia have a similar trend (they both increase from

us
1998 and decline after peaking in year 2001) (see Fig.5). This is also consistent with illegal

logging estimates generated by previous studies, such as Manurung et al. (2007), Harwell et al.
an
(2009) and Lawson et al. (2010). Given that our focus in this study is on two of Indonesia’s major

timber trading partners, Japan and China, using illegal values based upon the annual
M
discrepancies (in export and import records) offers a novel way to estimate the illegal timber

volumes, as compared to relying on overall illegal logging amount that includes all of Indonesia’s
ed

trading partners. Therefore, in this study, we use the discrepancies in the trading records

between Japan and Indonesia, and China and Indonesia as proxies for the export of illegally

logged timber from Indonesia (see Fig.5)7.

pt
ce
Ac

6
The timber trading records include trading of log, swan timer, veneer and plywood. These wood products were converted to

the log equivalent by index of logs 1:1, swan timber 1:1.43, veneer 1:1.9, plywood 1:2 (Hou et al., 2011).

7
Other main timber importing countries of Indonesia are UK and US. Around twenty percent of Indonesian plywood was

exported to these two countries (ITTO 1997-2012). However, they were not selected as the main illegal timber importers as

their estimated amount of illegal timber imports are not significant. Appendix B presents discrepancies in the trading records

between US and Indonesia, and UK and Indonesia between 1996 and 2010.

8
 MODELLING FRAMEWORK

The Simultaneous-Equation System Model

In a single-equation model, the dependent variable is expressed as a linear function of one or

more explanatory variables (Gujarati, 2003). However, when a two-way flow of influence among

t
variables is present (for instance, Y is determined by the X, and X is, in turn, determined by Y), a

ip
simultaneous-equation system comprising two or more equations is required (Gujarati, 2003). In

cr
this study, the price and the quantity of illegal logging are determined by the intersection of the
Downloaded by [University of Michigan] at 06:56 01 September 2017

demand and supply curves. Considering this simultaneous relationship, a simultaneous-equation

us
model including a supply function, two demand functions and an equilibrium condition function

was required: the demand for Indonesian illegally logged timber (proxied by the total demand of
an
Indonesian illegal timber by Japan and China) depends on legal timber price and economic

conditions of the importing countries, while the supply of Indonesian illegally logged timber
M

depends on the timber price and the economic and policy environments in Indonesia.
ed

Chinese demand for Indonesian illegal timber ( QC ) is modelled to depend on Indonesian

timber price (P), the number of new housing starts ( HSC ) and the log production in China
pt

( PLOGC ):
ce

l o g ( Q C ) = β 1 0 + β 1 1 lo g ( P ) + β 1 2 l o g ( H S C ) + β 1 3 lo g ( P L O G C ) + η C (1)
Ac

In several studies, illegal timber import is argued to be an important source of input for the

booming China economy (EIA&Telapak, 2005; Robbins et al., 2012). Variables such as the

number of new housing starts ( HSC ) is expected to positively influence the illegal timber

demand from China, while the domestic production of round log ( PLOGC ) is expected to have a

negative coefficient.

9
 Japanese demand for Indonesian illegal timber ( QJ ) is similarly modelled to depend on

Indonesian timber export price (P), the domestic production of round log ( PLOGJ ) and the

number of new housing starts ( HSJ ) in Japan:

lo g ( Q J ) = β 2 0 + β 2 1 lo g ( P ) + β 2 2 l o g ( H S J ) + β 2 3 lo g ( P L O G J ) + η J (2)

t
Japanese domestic production of timber products has decreased since the early 1990s due

ip
to recession and a tariff reduction after trade liberalization (Araya et al., 2008). Environmental

cr
concerns also led to a decline in the domestic wood production. As a result of the reduction in
Downloaded by [University of Michigan] at 06:56 01 September 2017

domestic supply, demand for imported timber, including illegal timber from Indonesia, increased.

us
Therefore, the domestic production of round log in Japan ( PLOGJ ) is expected to have a
an
negative impact on the illegal timber demand. The number of new housing starts variable ( HSJ )

is expected to positively influence the illegal timber demand from Indonesia.

M

Indonesian illegal timber supply ( QI ) is modelled to depend on Indonesian timber export

ed

price (P), the increase in the number of districts in Indonesia ( INDISI ), Indonesian GDP ( GDPI ),

Indonesian corruption index ( CI I ) and the percentage of paved road out of the total road in
pt

Indonesia ( PROADI ):
ce

lo g ( Q I ) = β 3 0 + β 3 1 lo g ( P ) + β 3 2 lo g ( I N D I S I ) + β 3 3 lo g ( G D PI ) + β 3 4 lo g ( C I I ) (3)
+ β 3 5 lo g ( P R O A D I ) + η I
Ac

The choice of explanatory variables in Equation (3) is based on the analysis of previous

studies. As discussed in Casson et al. (2002), Indonesia embarked on the path of decentralized

governance since 1998, leading to the addition of several new districts. Since the heads of these

districts are allowed to issue small-scale forest clearing concessions to companies, increasing

the number of districts may lead to an increase in forest clearance. The increasing number of

districts has been argued by Burgess et al. (2012) to accelerate the level of unsustainable

10
 harvesting in Indonesia. In addition, corruption has been seen as a culture at the top government

levels during the last thirty two years of governance by the ex-president Suharto (Palmer, 2001).

This culture has gradually spread over to the sub-national governments. Changes in corruption

perception index of Indonesia may therefore exhibit potential correlations with illegal logging

activities. Since the road network and road building were considered as key factors relating to the

transportation costs for forest harvesting and timber supply (Brown et al., 1994; Merry et al.,

2009), the percentage of paved road out of the total road in Indonesia was included in the supply

t
ip
equation. This variable may have positive effects on illegal timber supply as road paving has

been argued to exacerbate deforestation through logging and fire in Amazon forests (Kirby et al.,

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

2006; Nepstad et al., 2001).

us
The model is completed with a fourth equation that states the market equilibrium condition

(Indonesian illegal timber supply ( QI ) equals the sum of Chinese demand for Indonesian illegal
an
timber ( QC ) and Japanese demand for Indonesian illegal timber ( QJ )):
M

QI = QC + QJ (4)
ed

To investigate the impacts of different policies, we use dummy variables that take the value

of 1 when a policy is in place and 0 when it is not. The dummy variable, REGU, represents the
pt

ban on log exports, which is aimed at stopping round log exports and promoting sustainable

harvesting in Indonesia (Gellert, 2003; Luttrell et al., 2011). The dummy variable ANTIM
ce

represents the anti-money laundering law and anti-corruption law. These laws were argued to

contribute significantly towards curbing illegal logging in Indonesia (Luttrell et al., 2011). The
Ac

dummy variable, BIL, represents the bilateral agreements between Indonesia and its main timber

importing partners (including UK, the US, Malaysia, China and Japan) to restrain the trading of

illegal timber beyond Indonesian borders.

The model is estimated using annual data over the period 1996-2010. There are a number

of unique features of the model in terms of the selection of explanatory variables that differentiate

it from previous studies. The variable relating to Indonesian corruption index ( CI I ) has only

11
 been qualitatively analysed in previous studies of Indonesian illegal logging (such as those by

Smith et al. (2003), Harwell et al. (2009), Dermawan et al. (2011) and Downs et al. (2012)). None

of the previous studies has included the Indonesian corruption index ( CI I ) in an econometric

model to test its impact on illegal logging supply in Indonesia. Addtionaly, the total number of

districts in Indonesia has been increasing over time, which has led to an increase in harveting

quotas. While number of districts has been used previsouly for testing its impact on

t
deforestation in Indonesia in a study by Burgess et al. (2012), no econometric analysis exists

ip
thus far to directly test for its effects on illegal logging supply. Similarly, some other key variables

cr
such as the percentage of paved road in Indonesia, the number of new housing starts and the
Downloaded by [University of Michigan] at 06:56 01 September 2017

domestic production of round log in Japan and China have been used to analyse their impact on

us
deforestation, however, their role in explaining illegal logging in Indonesia has not been explicitly

explored. Further descriptions and data sources for these variables are presented in Appendix A.
an
Estimation of the Two Stage Least Square Model
M

With a simultaneous-equation model, applying the Ordinary Least Square (OLS) method to
ed

each demand and supply equation of the model will result in inconsistent estimates. This is due

to the correlation between price (P) and the error terms in demand and supply equations (Greene,
pt

2003). The model can be consistently estimated using the Two Stage Least Square (TSLS)

method. In the first stage, an instrumental variable called predicted P, which is highly related to P
ce

but unrelated to the error terms is estimated. In the second stage, the endogenous variable, P, is

replaced by predicted P obtained from the first stage. Equations that use predicted P, instead of
Ac

P, can be estimated using the OLS.

To estimate the instrumental variable in the first stage, log ( P ) is regressed on all

exogenous variables in the model as in equation (5).

log ( P ) = θ 0 + θ 1 log( HS C ) + θ log( PLOGC ) + θ log( HS J ) + θ log( PLOG J ) + θ log( INDIS I )

2 3 4 5

+θ 5 log( CI I ) + θ log(GDPI ) + θ lo g( PROAD I ) + π (5)

6 7

12
 The predicted value of log ( P ) based on the OLS estimates in (5) is then used as the

instrumental variable (predicted P).

In the second stage, the endogenous variable, log ( P ) , is replaced by predicted P in

Equations (1), (2) and (3). Each of these equations is then estimated using an OLS regression.

As explained above, the regression of TSLS can be done step by step manually. However,

several software packages offer an easier way to perform it automatically. We estimate the

t
model using Gretl software 8(Cottrell, 2001).

ip
cr
Tests for Appropriateness of the Instrumental Variable
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
Tests are conducted to determine if TSLS is required and whether the instrumental variable

helps overcome the inconsistency problem when TSLS is used. TSLS may help overcome the
an
inconsistency problem but increases the variances of estimators at the same time. This is

because the instrumental variable is just an estimate of the endogenous variable ( log ( P ) ) and is
M

not perfectly correlated with the endogenous variable. To determine whether TSLS is required,

the Hausman test is conducted when the endogenous variable ( log ( P ) ) (rather than the
ed

instrumental variable) is used in the demand and supply equations to test the correlation

between the endogenous variable and disturbances. The null hypothesis of the Hausman test is
pt

that no correlation exists (Greene, 2003). The null hypothesis cannot be accepted when the P
ce

value of the test results is or lower than 0.1. TSLS is then required for the model.

To test the validity of the instrumental variable in the TSLS model, we use Sargan test to
Ac

determine whether the instrumental variables are correlated with disturbances and the weak

instrument test to determine whether the instrumental variables contain sufficient information

from the original endogenous variable to be relevant. The null hypothesis of Sargan test is that all

instrumental variables are valid (Gujarati, 2003). If the P values are larger than 0.1, the null

hypothesis can be accepted. The weak instrument problem can be tested using the F-statistics of

8
Details related to Gretl Software can be found at the website: http://gretl.sourceforge.net/

13
 the first stage of the TSLS (Bound et al., 1995; Stock et al., 2002). A value larger than ten implies

a good fit without the weak instruments problem (Cottrell et al., 2012). The F-statistics of the first

stage of the TSLS are presented in the third to last row of Table 1, and are all higher than ten.

Test results, as shown in Table 1, demonstrate that TSLS method performs well in this

model except for the Hausman test and Sargan test in the equations estimating China’s demand

and Japan’s demand. The P values of the Hausman test indicate that the OLS is consistent in

t
the Chinese and Japanese demand function. The Sargan test of the functions suggests that the

ip
instruments are invalid. Therefore, the China and Japan’ demand functions were regressed

cr
separately using OLS. The results are also presented in the Table 1. However, the differences
Downloaded by [University of Michigan] at 06:56 01 September 2017

between results under TSLS and under OLS for the Japan and China’s demands turn out to be

us
insignificant.
an
ESTIMATION OF RESULTS AND DISCUSSION
M

Analysis of Factors Causing Illegal Logging

ed

The estimation results are provided in Table 1. Since all variables are log transformed, the

estimated model is a double log model. As such, the coefficient of an explanatory variable is the
pt

elasticity of the dependent variable with respect to a change in that explanatory variable. The
ce

estimation results indicate that the increase in the number of districts (INDISI) has a negative and

significant effect on the illegal timber supply from Indonesia. An increase in the number of INDISI,
Ac

by design, leads to more logging rights, and increases the area of legally harvested forest. With a

reduction in total area of forest available for harvesting and everything else staying the same,

illegal logging would decline. However, some logging practices permitted by local governments

might be considered as ‘illegal’ by the central government (Casson et al., 2002). That is because

the district governments have issued permits for forest clearing activities without considering the

sustainability of the forest. Consequently, even if the logging activities permitted by district

14
 governments are legal, the amount of forests harvested increases despite a reduction in illegal

logging.

The coefficient of CII is negative and highly significant, suggesting that higher level of

corruption is positively correlated with illegal logging. Note that CII is a corruption perception

index that takes a lower value when corruption is high. This finding is consistent with the findings

by Palmer (2001) and Downs et al. (2012), that corruption is a key driver of accelerating

t
harvesting and trading in illegal timber. Most illegal and corrupt activities related to timber trade,

ip
such as harvesting, delivery and making forged documents, happen in Indonesia rather than in

cr
the importing countries. In the model presented here, introducing a variable relating to the
Downloaded by [University of Michigan] at 06:56 01 September 2017

corruption index in Indonesia is much more meaningful than including the corruption perception

us
index of importing countries (the corruption perception indices for Japan and China were tested

in the model, but they emerged as insignificant).

an
As expected, the percentage of paved roads in the total road length in Indonesia (PROADI)
M
shows a positive and significant coefficient. A good road network reduces the transportation

costs for delivering round-log from the harvest locations to the milling centre and to the ports.

The increasing percentage of paved roads in Indonesia plays a positive effect in promoting illegal
ed

timber export by reducing transportation costs.

pt

The third and fourth columns in Table 1 present the estimates of illegal timber demand in

Japan. The number of new houses (HSJ) has a positive and significant coefficient. The use of
ce

wood in buildings is considered as a part of Japanese culture and tradition. Wood and wood

products used in the construction industry accounted for a large proportion (40 percent) of the
Ac

total wood demand in Japan (MAFF, 2010). In a traditional house, about 0.2 m3 of wood is

required for each squared meter of non-structural area (MAFF, 2010). However, most of the

wood (80 percent) used in construction is imported (MAFF, 2010). Therefore, the number of new

housing starts plays an important role in influencing the import of wood and that of the illegal

timber.

15
 As expected, domestic wood production (PLOGJ) in Japan has a negative and significant

impact on Japanese demand for illegally logged timber import from Indonesia. As shown in Fig.6,

the domestic timber self-sufficiency rate in Japan fell in the beginning of the 1990s and rose back

after 2003. The rate in 2007 was 23 percent which is similar to the rate (25 percent) in 1991.

During the time period from 1992 to 2007, the domestic timber self-sufficiency rate was lower

than 25 percent. This meant that Japan’s domestic wood supply was insufficient, and this, in turn,

led to a higher demand for illegal timber from Indonesia (the impact of Japan’s GDP on illegal

t
ip
timber demand was also tested in the model, however, it emerged as insignificant).

cr
The last two columns in Table 1 present the estimates for illegal timber demand in China.
Downloaded by [University of Michigan] at 06:56 01 September 2017

Domestic wood production (PLOGC) in China has a negative and significant impact on illegal

us
timber import. Fig.7 illustrates round log production figures in China from 1996 to 2010. As

shown in Fig.7, the production of round log declined since 1998. China’s Natural Forest
an
Protection Programme was also implemented in 1998. This Programme was aimed at controlling

the resource crisis in the national forest areas and preventing environmental degradation
M
(Witness, 2009). Timber harvesting for commercial purposes was halted completely in the upper

Yangtze River and also in the middle and upper reaches of the Yellow River. Commercial logging
ed

was also sharply reduced in the Northeast and Inner Mongolia. This Programme contributed a

great deal towards the protection of China’s forests. The natural forest cover increased by 393.05
pt

million m3 within the last ten years, as indicated by comparisons between the 6th and 7th national

forest resources inventory (Cao et al., 2011). In the meanwhile, the number of reforestation and
ce

afforestation projects have significantly increased in China. In 2000, the reforested area was

around 46,000 hectares, including 4,000 hectares of fast-growing forests (Liu et al., 2000). After
Ac

the log production hit a low in 2002, production has increased as a result of the reforestation

projects leading to additional harvesting options. The fluctuations in domestic wood supply have

had significant effects on the timber imports demands. With a decrease in the domestic wood

supply, domestic consumption has increasingly relied on imports. This has also added to the

demand for illegal timber, especially for the tropical wood from Indonesia. However, whenever

the domestic wood supply has increased, the illegal timber demand has declined.

16
 One crucial aspect that we have not included in our analysis is related to the impact of palm

oil prices on illegal logging. One would expect that as palm oil prices increased, the tendency to

clear forest would increase as well. However, between 1990 and 2005, nearly half of oil palm

expansion occurred through conversion of cropland area (Koh et al., 2008) and oil palm was

often planted in the area previously degraded by fire and logging (Fitzherbert et al., 2008). The

effect of palm oil expansion on deforestation has been discussed in a number of studies in the

past (Abood et al., 2015; Fitzherbert et al., 2008; Sandker et al., 2007). However, it has not been

t
ip
possible to find a quantitative linkage between palm oil expansion and illegal logging in Indonesia.

In this paper, we included the variables concerning palm oil price9, palm kernel oil price10 and

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Indonesian palm oil production 11 to test for their impact on illegal logging. However, these

us
variables had to be dropped as they turned out to be insignificant and of wrong sign. Moreover,

they were also found to be correlated with a key variable in our model, namely the Indonesian
an
corruption index.
M
Impact of Policies to Mitigate Illegal Logging
ed

The impacts of various policies are examined by including a dummy variable that represents the

introduction of a particular policy in the illegal logging supply equation. The studied policies

include a ban on log exports (REGU), the anti-money laundering and anti-corruption law of 2005
pt

(ANTIM) and the bilateral agreements with timber-importing countries enacted in 2002 (BIL).
ce

Since the correlation coefficients between CII and the regulation related dummy variables, as

well as between INDISI and the dummy variables, are higher than 0.5, the variables CII, INDISI
Ac

were dropped from these regressions to avoid potential multi-collinearity issues. The estimated

results are described in Table 2.

9
Source: World Bank Commodities Price Data
10
Source: World Bank Commodities Price Data
11
Source: United States Department of Agriculture https://pecad.fas.usda.gov/highlights/2015/12/Indonesia/Index.htm

17
 The results of the regression including REGU as a variable are presented in the third column

of Table 2. The coefficient for REGU is significant and negative, which means that the total illegal

supply has dropped since the implementation of the log export ban. This could also be confirmed

through the Indonesian log trading records from ITTO (1997-2012). The trading records of

Indonesian log exports to China and Japan, and Chinese and Japanese log imports from

Indonesia show a significant drop when the log exports ban was reintroduced in 2001 (see Fig.8

and 9). Note that log is not the only product included in the definition of timber export recorded by

t
ip
Indonesia or timber import recorded by China and Japan. Other products include veneer,

plywood and sawn timber. Burgess et al. (2012) have suggested that after the ban of log export,

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

illegal as well as legal logs are transformed into processed products such as plywood before

us
being exported, and the effectiveness of the log export ban in reducing illegal logging is doubtful.

This view may be supported by the ITTO (1997-2012) report as well, which suggests that the
an
amount of log exported by Indonesia has declined significantly, while at the same time, Indonesia

has been the largest exporting country for plywood since 1997. However, since the definition of
M
illegal logging includes log as well as other products, our results suggest that the ban of log

export is actually effective in reducing illegal logging, even after the effect of product
ed

transformation is taken into account.

The coefficient of ANTIM (anti-money laundering and anti-corruption laws) is negative and
pt

significant (in the fourth column of Table 2). Since corruption is a key driver of illegal harvesting,

the laws and related policies are effective in curbing illegal logging by catching the grey money
ce

and preventing corruption. Controlling corruption will also provide a climate for designing and

implementing other policy instruments to reduce illegal logging (Amacher, 2006).

Ac

The regression results incorporating the bilateral agreements (BIL) are presented in the last

column of Table 2. Note that the coefficient for BIL is also negative and significant. Illegal logging

is an activity that involves several parties and countries. Curbing illegal logging needs the

cooperation of all countries involved. The estimates of their effects on timber-importing countries

will be discussed in the next section.

18
 The Green Purchase Law in Japan

The Green Purchasing Law was introduced in Japan to promote the consumption of eco-friendly

goods and services by the states and other entities. The item ‘legal wood’ was added in the

Green Purchasing Law by the Japanese government after the issue of illegal logging was

discussed at G8 Summit in 2005 (MOE, 2007). Based on this law, the wood and wood products

need to be verified for their legality and sustainability before being imported to Japan (MOE,

t
ip
2007). Meanwhile, methods, such as tracing timber source, using satellite data and cooperating

with related countries, were developed by the Japanese government to deal with illegal logging

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

(MOE, 2007). Table 3 presents the regression results using the Green Purchasing Law

us
(GREENJ) as a dummy variable. The variable of GREENJ has a negative and significant

coefficient, which means the law has had a positive effect on reducing the import of illegal timber.
an
In addition, the tests for bilateral agreements (BIL) in Japan’s and China’s demand models

are presented in Table 3. Such agreements, especially involving several parties, however, have
M
had limited effects. The limited effect in the case of China have been attributed to the fact that

the MoU was signed with the Chinese State Forest Administration, which is only responsible for
ed

domestic forestry issues (Jurgens, 2006). This agreement could become more effective if the

China Customs Agency, which is responsible for the screening of imported goods, can get
pt

involved in the future (Jurgens, 2006).

ce

Finally, it is also worth testing whether the global financial crisis (in 2007-2008) had any

impact on illegal timber demand or supply. A dummy variable, (which took the value of 1 for the
Ac

period 2007-2008 and 0 in others years) reflecting the potential effects of the global financial

crisis of 2007 and 2008, was added to the model. However, the variable turned out to be

insignificant.

19
 LIMITATIONS OF THE STUDY

Our study has a number of limitations in terms of data and methodology, which are worth

mentioning. First, as no official statistics of illegal logging could be found, we estimated

Indonesian illegal logging on the basis of the trading discrepancies between Indonesia and its

main importing partner countries. Although the accuracy of the data could not be guaranteed, the

t
annual discrepancies between Japan and Indonesia, and China and Indonesia have a similar

ip
trend (they both increase since 1998 and decline after peaking in year 2001), which is also

cr
consistent with illegal logging tendencies estimated by previous studies, such as Manurung et al.
Downloaded by [University of Michigan] at 06:56 01 September 2017

(2007), Harwell et al. (2009) and Lawson et al. (2010). Second, except for Japan and China,

us
other Indonesian timber importing countries and associated policies were not considered in this

paper. The UK and the US are important timber trading partners of Indonesia and are also
an
suspected of importing Indonesian illegal logging. However, they were not selected in this study

due to lack of credible data (that is, their timber trading discrepancies were not found to be
M

significant). Third, some key policies for curbing illegal logging, such as the EU FLEGT VPA

process, the SVLK and the US Lacey Act Timber Amendments in 2008, were not considered in
ed

this paper as they are beyond the time period of the study (1996 to 2010) and also because they

do not have a direct impact on curbing illegal logging trading between Indonesia and Japan, or
pt

between Indonesia and China.

ce
Ac

CONCLUDING REMARKS

In this paper, we develop a simultaneous-equations model to estimate the impact of various

factors that affect illegal harvesting and trading of timber in Indonesia. The model is estimated

using annual data for the period 1996-2010. We estimate the determinants of illegally logged

timber demand from Japan and China as well as the supply of illegal timber from Indonesia. For

20
 modelling illegally logged timber supply, the key explanatory variables used in our study include

the number of new districts created in Indonesia during the period of analysis, corruption index

as well as related policies, such as Indonesian log exports ban, Indonesian anti-money

laundering and anti-corruption laws and the various bilateral agreements between timber trading

countries. For estimating the demand for illegally logged timber, we include the excess demand

for timber as well as the number of new housing constructions in the importing countries.

t
The main findings of this study can be summarized as follows. On the supply side,

ip
corruption in Indonesia has a significant impact on the supply of illegally logged timber. In

cr
addition, the increase in forest extraction permissions resulting from an increase in the number of
Downloaded by [University of Michigan] at 06:56 01 September 2017

districts may merely disguise ‘illegal’ logging as a legal practice. Although the results suggest

us
that ‘illegal’ logging has reduced as a result, the total rate of deforestation has in fact increased.
an
On the demand side, the timber demand in China did shift from domestic to import based,

including that for illegal timber, due to the implementation of natural resource protection policies
M
in the recent past. Whereas, domestic supply has increased due to a rapid implementation of

afforestation and reforestation based projects. Therefore, the demand for imported illegal timber

from Indonesia has declined as a consequence.

ed

The number of new housing starts is found to be a significant driver of imports demand from
pt

Japan for illegal Indonesian timber. The Japanese building construction industry has become the

largest consumer of wood products from illegal sources (Araya et al., 2008). Our analysis also
ce

suggests that the level of domestic timber supply in Japan has an important impact on the import

of Indonesian illegally logged timber into the country.

Ac

We further find that law enforcement related activities in Indonesia, or policies with a law

foundation, are significant factors that help reduce illegal logging. Since corruption is a key driver

of illegal harvesting, the associated laws and related policies are effective in curbing illegal

logging by catching grey money and preventing corruption. Finally, it is also worth noting the

limitations of this study as pointed out earlier.

21
 REFERENCES

Abood, Sinan A, Lee, Janice Ser Huay, Burivalova, Zuzana, Garcia-Ulloa, John, & Koh, Lian Pin. (2015).

Relative contributions of the logging, fiber, oil palm, and mining industries to forest loss in Indonesia.

Conservation letters, 8(1), 58-67.

t
Alemagi, Dieudonne, & Kozak, Robert A. (2010). Illegal logging in Cameroon: Causes and the path

ip
forward. Forest Policy and Economics, 12(8), 554-561.

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Amacher, Gregory S. (2006). Corruption: A challenge for economists interested in forest policy

us
design. Journal of forest economics, 12(2), 85-89.

Araya, Akihiko, & Katsuhisa, Hikojiro. (2008). Japanese Wood Market and Use of Tropical Wood.
an
Bound, John, Jaeger, David A, & Baker, Regina M. (1995). Problems with instrumental variables

estimation when the correlation between the instruments and the endogenous explanatory variable
M
is weak. Journal of the American Statistical Association, 90(430), 443-450.

Brack, Duncan, Gray, Kevin, & Hayman, Gavin. (2002). Controlling the international trade in illegally
ed

logged timber and wood products. Sustainable Development Programme, Royal Institute of

International Affairs. London.

pt

Brown, Sandra, Schreier, Hans, Thompson, William A, & Vertinsky, Ilan. (1994). Linking multiple
ce

accounts with GIS as decision support system to resolve forestry/wildlife conflicts. Journal of

Environmental Management, 42(4), 349-364.

Ac

Burgess, Robin, Hansen, Matthew, Olken, Benjamin A, Potapov, Peter, & Sieber, Stefanie. (2012).

The Political Economy of Deforestation in the Tropics. The Quarterly Journal of Economics, 127(4),

1707-1754.

Callister, Debra. (1999). Corrupt and illegal activities in the forestry sector: Current understandings,

and implications for world bank forest policy: Draft for discussion. World Bank, Washington.

22
 Cao, Yu-kun, Lv, Tian, & Chen, Ning-jing. (2011). Analysis of the impact on natural forest protection

program to forestry policies in China. Issues of Forestry Economics, 5, 377-391.

Casson, Anne, & Obidzinski, Krystof. (2002). From new order to regional autonomy: Shifting

dynamics of “illegal” logging in Kalimantan, Indonesia. World Development, 30(12), 2133-2151.

Contreras-Hermosilla, Arnoldo, Doornbosch, Richard, & Lodge, Michael. (2008). The economics of

illegal logging and associated trade.

t
ip
Cottrell, Allin. (2001). Gretl Manual.

Cottrell, Allin, & Lucchetti, Riccardo. (2012). Gretl user’s guide. Distributed with the Gretl library.

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Dermawan, Ahmad, Petkova, Elena, Sinaga, Anna, Muhajir, Mumu, & Indriatmoko, Yayan. (2011).

us
Preventing the risk of corruption in REDD+ in Indonesia. Bogor Barat: Center for International

Forestry Research (CIFOR).

an
Dieter, Matthias. (2009). Analysis of trade in illegally harvested timber: Accounting for trade via third

party countries. Forest Policy and Economics, 11(8), 600-607.

M

Downs, Fiona, & Tacconi, Luca. (2012). Corruption, deforestation and environmental injustice: the

case of Indonesia. Paper presented at the Paper presented at the 11th Global Conference on
ed

Environmental Justice and Global Citizenship.

EIA&Telapak. (2005). The Last Frontier: Illegal Logging in Papua and China's Massive Timver Theft.
pt

FAO. (2010). Global Forest Resources Assessment 2010 Main Report: FAO.
ce

FFPRI. (2005). Current Activities to Combat Illegal Logging and Associated Trade in Illegally Sourced

Wood Products in Japan.

Ac

Fitzherbert, Emily B, Struebig, Matthew J, Morel, Alexandra, Danielsen, Finn, Brühl, Carsten A,

Donald, Paul F, & Phalan, Ben. (2008). How will oil palm expansion affect biodiversity? Trends in

ecology & evolution, 23(10), 538-545.

Gellert, Paul K. (2003). Renegotiating a timber commodity chain: lessons from Indonesia on the

political construction of global commodity chains. Paper presented at the Sociological Forum.

Greene, W.H. (2003). Econometric Analysis. New Jersey: Prentice Hall.

23
 Gujarati, Damodar N. (2003). Basic econometrics (4th ed.). Boston: McGraw Hill.

Hansen, Christian Pilegaard, & Treue, Thorsten. (2008). Assessing illegal logging in Ghana.

International Forestry Review, 10(4), 573-590.

Harwell, Emily, & Blundell, Arthur. (2009). " Wild Money": The Human Rights Consequences of Illegal

Logging and Corruption in Indonesia's Forestry Sector. New York: Human Rights Watch.

Hou, Fangmiao, & Song, Weiming. (2011). An analysis on the supply and demand of timber in China.

t
ip
Paper presented at the E-Business and E-Government (ICEE), 2011 International Conference.

IGES. (2007). Japan's Public Procurement Policy of Legal and Sustainable Timber: Progress,

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Challenges and Ways Forward.

us
ITTO. (1997-2012). Annual Review and Assessment of the World Timber Situation.

Jurgens, Emile. (2006). Learning lessons to promote certification and combat illegal logging in
an
Indonesia: September 2003 to June 2006: CIFOR.

Kirby, Kathryn R, Laurance, William F, Albernaz, Ana K, Schroth, Götz, Fearnside, Philip M, Bergen,
M

Scott, . . . Da Costa, Carlos. (2006). The future of deforestation in the Brazilian Amazon. Futures,

38(4), 432-453.
ed

Koh, Lian Pin, & Wilcove, David S. (2008). Is oil palm agriculture really destroying tropical biodiversity?

Conservation letters, 1(2), 60-64.

pt

Lawson, Sam, & MacFaul, Larry. (2010). Illegal logging and related trade: Indicators of the global
ce

response: Chatham House London.

Liu, Ai-min, Li, Fei, & Liao, Jun-guo. (2000). Study on the balance of supply and demand of timber and
Ac

forest resources in China. Resources Science, 22(6), 9-13.

Luttrell, Cecilia, Obidzinski, Krystof, Brockhaus, Maria, Muharrom, Efrian, Petkova, Elena, Wardell,

Andrew, & Halperin, JAMES. (2011). Lessons for REDD+ from measures to control illegal logging in

Indonesia: CIFOR.

MAFF. (2010). Annual Report on Forest and Forestry in Japan

24
 Makkarennu, Nakayasu Akira. (2013). Prospective Indonesian Plywood in the Global Market. Journal

of Life Sciences and Technologies Vol, 1(3), 190-195.

Manurung, EGT, Simangunsong, BCH, Sukadri, D, Widyantoro, B, Justianto, A, Ramadhan, S, . . .

Priyono, BM. (2007). A road map for the revitalization of Indonesia’s forest industry.

McElwee, Pamela. (2004). You say illegal, I say legal: the relationship between ‘illegal’logging and

land tenure, poverty, and forest use rights in Vietnam. Journal of Sustainable Forestry, 19(1-3), 97-

t
ip
135.

Merry, Frank, Soares-Filho, Britaldo, Nepstad, Daniel, Amacher, Gregory, & Rodrigues, Hermann.

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

(2009). Balancing conservation and economic sustainability: the future of the Amazon timber

us
industry. Environmental Management, 44(3), 395-407.

MOE. (2007). Japan's Green Purchasing Policy: Tackling Illegal Logging.

an
Nanang, David M. (2010). Analysis of export demand for Ghana's timber products: A multivariate co-

integration approach. Journal of forest economics, 16(1), 47-61.

M

Nepstad, Daniel, Carvalho, Georgia, Barros, Ana Cristina, Alencar, Ane, Capobianco, João Paulo,

Bishop, Josh, . . . Prins, Elaine. (2001). Road paving, fire regime feedbacks, and the future of Amazon
ed

forests. Forest ecology and management, 154(3), 395-407.

Obidzinski, Krystof, Andrianto, Agus, & Wijaya, Chandra. (2006). Timber Smuggling in Indonesia:
pt

Critical Or Overstated Problem?: Forest Governance Lessons from Kalimantan: CIFOR.

ce

Obidzinski, Krystof, Dermawan, Ahmad, Andrianto, Agus, Komarudin, Heru, & Hernawan, Dody.

(2014). The timber legality verification system and the voluntary partnership agreement (VPA) in
Ac

Indonesia: Challenges for the small-scale forestry sector. Forest Policy and Economics, 48, 24-32.

Palmer, Charles. (2001). The extent and causes of illegal logging: An analysis of a major cause of

tropical deforestation in Indonesia. CSERGE Working Paper.

Prestemon, Jeffrey P. (2015). The impacts of the Lacey Act Amendment of 2008 on US hardwood

lumber and hardwood plywood imports. Forest Policy and Economics, 50, 31-44.

25
 Richards, M, Wells, A, Del Gatto, F, Contreras-Hermosilla, A, & Pommier, D. (2003). Impacts of

illegality and barriers to legality: a diagnostic analysis of illegal logging in Honduras and Nicaragua.

International Forestry Review, 5(3), 282-292.

Robbins, A, & Perez-Garcia, John. (2012). Impacts of Illegal Logging Restrictions on China’s Forest

Products Trade. Paper presented at the Conference paper delivered on March.

Sandker, Marieke, Suwarno, Aritta, & Campbell, Bruce. (2007). Will forests remain in the face of oil

t
ip
palm expansion? Simulating change in Malinau, Indonesia. Ecology and Society, 12(2).

Setiono, Bambang, & Husein, Yunus. (2005). Fighting forest crime and promoting prudent banking

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

for sustainable forest management: The anti money laundering approach.

us
Smith, Joyotee, Obidzinski, Krystof, Subarudi, Subarudi, & Suramenggala, Iman. (2003). Illegal logging,

collusive corruption and fragmented governments in Kalimantan, Indonesia. International Forestry

an
Review, 5(3), 293-302.

Stock, James H, & Yogo, Motohiro. (2002). Testing for weak instruments in linear IV regression:
M

National Bureau of Economic Research Cambridge, Mass., USA.

Witness, Global. (2009). A Disharmonious Trade: China and the Continued Destruction of Burma's
ed

Northern Frontier Forests: a Review by Global Witness, 2006-09: Global Witness.

pt
ce
Ac

26
 Appendix A Descriptions and sources for variables in the model

Variables Description Units Source

QJ Illegal timber imports in Japan m3 ITTO
PLOGJ Round-log production in China 1000m3 (Araya et al., 2008)
HSJ Number of housing starts in Japan Unit Ministry of Land,
Infrastructure,
Transport and
Tourism in Japan
QC Illegal timber imports in China m3 ITTO
HSC Number of housing starts in China Million Urban housing
m2

t
Markets in China
(2009)

ip
PLOGC Round-log production in China 1000m3 National Bureau of
Statistics of China
QI Sum of illegal exports of Japan and China m3 ITTO

cr
P Legal timber export price in Indonesia $/m3 FAOSTAT
Downloaded by [University of Michigan] at 06:56 01 September 2017

CII Corruption Perception Index of Indonesia Index Transparency

International

us
INDISI Increase in number of districts in Indonesia Number (Burgess et al.,
2012)
PROADI Percentage of paved road on the total road in % World Bank
Indonesia
GDPI
an
Indonesian GDP (basic year is 2005) US$ World Bank, CEIC
REGU The ban of log exports first introduced from 1985 Binary (Luttrell et al., 2011)
to 1997 and reintroduced in 2003(1-the policy variable
has been introduced;0-no policy)
ANTIM The anti-money laundering and anti-corruption Binary (Luttrell et al., 2011)
M
law introduced in 2005(1-the policy has been variable
introduced;0-no policy)
BIL The bilateral agreements with timber-importing Binary (Luttrell et al., 2011)
countries introduced in 2002 (1-the policy has variable
ed

been introduced;0-no policy)

GREENJ The item of ‘legal wood’ was added in the Green Binary (MOE, 2007)
Purchasing Law by the Japanese government in variable
2006 (1-the policy has been introduced;0-no
policy)
pt
ce
Ac

27
 Appendix B Timber imports from Indonesia and Indonesian records
of timber exports to US and UK from 1996 to 2010

2500
US's Records
Timber Imports/Exports (1000m3)

2000 Indonesian Records

1500

t
ip
1000

cr
500
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
0

an
Fig. B.1. Discrepancy between US records of timber imports from Indonesia and Indonesian records
of timber exports to US from 1996 to 2010

Source: ITTO (1997-2012)

M
ed
pt
ce
Ac

28
 600
UK's Records

Timber Imports/Exports (1000m3)

500
Indonesian Records

400

300

200

t
100

ip
0

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
Fig. B.2. Discrepancy between UK records of timber imports from Indonesia and Indonesian records
of timber exports to UK from 1996 to 2010

Source: ITTO (1997-2012)

an
M
ed
pt
ce
Ac

29
 FIGURES AND TABLES

Table 1 Results of the simultaneous-equation model

Indonesia’s Japan’s Japan’s China’s China’s

supply demand demand demand demand
QI QJ QJ QC QC
TSLS TSLS OLS TSLS OLS
Methods
constant 29.9526 -29.8480* -26.9903 35.8619*** 34.1234**
(31.0909) (16.2783) (30.1302) (10.2140) (14.4987)
P - - 0.6408 - 0.2071

t
(1.7674) (1.1648)

ip
Predicted P -3.1778 0.800 - 0.0414 -
(2.5229) (0.8453) (0.8952)
INDISI -0.4347* - - - -

cr
(0.2447)
Downloaded by [University of Michigan] at 06:56 01 September 2017

CII -13.1539*** - - - -
(3.2322)

us
PROADI 10.9773**
(4.4364)
GDPI -0.4920 - - - -
(0.8325)
HSJ -
an
5.5056*** 5.4644* - -
(1.0202) (1.6375)
PLOGJ - -4.4033*** -4.3777*** - -
(1.5053) (2.2124)
HSC - - - -0.2987 -0.2933
M
(0.2300) (0.3014)
PLOGC - - - -1.8929*** -1.8974**
(0.4172) (0.7930)
Hausman Chi- Chi- R-squared Chi- R-squared
ed

test square(1)= square(1)= =0.5 square(1)= =0.5

3.1056 0.6343 0.7187
P value= P value= P value=
0.078 0.4258 0.3966
pt

Weak F-statistic= F-statistic= F-statistic=

instruments 136.115 223.529 215.23
test
ce

Sargan test LM=5.6958 LM=12.3755 LM=9.7882

P value= P value= P value=
0.1274 0.03 0.0815
Note: the asterisk*, **, *** denote significance at the 10%, 5% and 1% level.
Ac

Number in parentheses are standard errors.

30
 Table 2 Results of Regulatory Influence on Illegal Trading in Indonesia

QI
constant 250.473*** 174.829*** 229.551***
(91.7428) (49.8416) (47.8729)
predicted P -19.2447** -11.4108*** -15.5242***
(7.9533) (3.8996) (3.8342)
GDPI -2.1299 -3.4888*** -4.9109***
(1.5256) (1.173) (1.0303)
REGU -5.6186** - -
(2.3530)
ANTIM - -3.8387*** -
(0.8750)

t
BIL - - -4.6162***

ip
(0.8750)
Hausman test Chi-square(1)= Chi-square(1)= Chi-square(1)=
5.5332 8.6273 7.1192

cr
P value= P value= P value=
0.0187 0.0033 0.0076
Downloaded by [University of Michigan] at 06:56 01 September 2017

Weak F-statistic= F-statistic= F-statistic=

instruments test 13.7354 19.6667 42.4932

us
Sargan test LM=6.7727 LM=2.7244 LM=2.8579
P value= P value= P value=
0.2381 0.6050 0.5819
Note: the asterisks*, **, *** denote significance at the 10%, 5% and 1% level.
Number in parentheses are standard errors.
an
M
ed
pt
ce
Ac

31
 Table 3 Results of Regulatory influence on Illegal Demand in Japan and China

QJ QJ QJ QJ QC QC
Methods TSLS OLS TSLS OLS TSLS OLS
constant 0.2491 2.4257 61.0896** 65.5900 38.3482*** 36.3421*
(23.5207) (30.2439) (30.4033) (38.878) (12.4875) (17.0971)
predicted P 0.5029 - -1.528 - - -
(0.8117) (1.1632)
P - 0.3315 - -1.7926 -0.1234 0.0558
(1.5731) (1.5969) (1.0504) (1.3306)
HSJ 2.8617* 2.8192 3.5425*** 3.4605** - -
(1.7124) (1.9334) (0.8665) (1.4384)
PLOGJ -3.2523** -3.2283 -8.1483*** -8.2285*** - -

t
(1.5102) (2.0375) (1.3894) (2.1495)

ip
HSC -0.0540 -0.0900
(0.4097) (0.8004)
PLOGC - - - - -2.0778*** -2.0510***

cr
(0.4750) (0.9952)
GREENJ -1.3095* -1.316* - - - -
Downloaded by [University of Michigan] at 06:56 01 September 2017

(0.6927) (0.6358)
BIL -1.9366*** -1.9877** -0.3165 -0.2636

us
(0.5562) (0.6666) (0.4393) (0.9546)
Hausman Chi- R-squared= Chi- R-squared= Chi- R-squared=
test square(1)= 0.4938 square(1)= 0.6111 square(1)= 0.4
0.4695 0.6290 0.5810
P value=
an
P value= P value=
0.4932 0.4277 0.4459
Weak F-statistic= F-statistic= F-statistic=
instruments 160.335 160.26 228.603
M
test
Sargan test LM=11.828 LM=11.801 LM=10.834
3 5 8
P value= P value= P value=
0.0372 0.0376 0.0548
ed

Note: the asterisks*, **, *** denote significance at the 10%, 5% and 1% level.
Number in parentheses are standard errors.
pt
ce
Ac

32
 10000
9000 Production
Indonesian Plywood (1000 m3) 8000 Exports
7000
6000
5000
4000
3000
2000

t
1000

ip
0

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Year

us
Fig.1. Indonesian plywood production and plywood exports between 1996 and 2010

Source: ITTO (1997-2012)

an
M
ed
pt
ce
Ac

33
 9000
8000 Total Exports
Indonesian Plywood (1000m3)
7000 Chinese and Japanese Imports
6000
5000
4000
3000

t
2000

ip
1000
0

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

Year

us
Fig.2. China and Japan plywood imports from Indonesia and Indonesian plywood exports from 1996
to 2010
an
Source: ITTO (1997-2012)
M
ed
pt
ce
Ac

34
 8000
Japanese Records
Timber Imports/Exports (1000m3)

7000
Indonesian Records
6000

5000

4000

t
3000

ip
2000

1000

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
an Year

Fig.3. Discrepancy between Japanese records of timber imports from Indonesia and Indonesian
records of timber exports to Japan from 1996 to 2010
M
Source: ITTO (1997-2012)
ed
pt
ce
Ac

35
 4000
Timber Imports/Exports (1000m3) Chinese Records
3500

3000 Indonesian Records

2500

2000

1500

t
1000

ip
500

cr
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
Year

Fig.4. Discrepancy between Chinese records of timber imports from Indonesia and Indonesian
an
records of timber exports to China from 1996 to 2010

Source: ITTO (1997-2012)

M
ed
pt
ce
Ac

36
 Downloaded by [University of Michigan] at 06:56 01 September 2017

Estimated Illegal Timber Exports from

Indoneisa (1000m3)
Ac

0
1000
2000
3000
4000
5000
6000

ce
pt
ed

37
Year
M
an
Japan
China

us
cr
ip
Fig.5. The estimated illegal timber exports from Indonesia to Japan and China from 1996 to 2010

t
 Downloaded by [University of Michigan] at 06:56 01 September 2017

Domestic Timber Self-Sufficiency in

Japan (%)
Ac

0
5
10
15
20
25
30

ce
pt
ed

38
Year
M
Source: (Araya et al. (2008))
an
us
Fig.6. Domestic timber self-sufficiency percentages in Japan from 1991 to 2007

cr
ip
t
 90000

Production of Round Log in China

80000
70000

(1000m3) 60000
50000
40000
30000
20000
10000

t
0

ip
cr
Year
Downloaded by [University of Michigan] at 06:56 01 September 2017

us
Fig.7. Production of round log in China from 1996 to 2010

Source: National Bureau of Statistics of China

an
M
ed
pt
ce
Ac

39
 1200
Chinese Records

Log Imports/Exports (1000m3)

1000
Indonesian Records
800

600

400

200

t
ip
0

cr
Year
Downloaded by [University of Michigan] at 06:56 01 September 2017

Fig.8. Chinese records of log imports from Indonesia and Indonesian records of log exports to China

us
from 1996 to 2010

Source: ITTO (1997-2012)

an
M
ed
pt
ce
Ac

40
 180

Log Imports/Exports (1000m3)

160 Japanese Records
140 Indonesian Records
120
100
80
60
40

t
20

ip
0

cr
Year
Downloaded by [University of Michigan] at 06:56 01 September 2017

Fig.9. Japanese records of log imports from Indonesia and Indonesian records of log exports to

us
Japan from 1996 to 2010

Source: ITTO (1997-2012)

an
M
ed
pt
ce
Ac

41