Analisis Volatilitas Harga Beras di Indonesia: Efek Stok dan Perubahan Kebijakan.

PRICE VOLATILITY ANALYSIS OF RICE IN INDONESIA:
THE IMPACT OF STOCK AND CHANGES IN POLICY

DESSY ANGGRAENI

GRADUATE SCHOOL
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2015

STATEMENT OF THESIS, SOURCE OF INFORMATION
AND COPYRIGHT TRANSFER*
I hereby declare that thesis titled Price Volatility Analysis of Rice in
Indonesia: The Impact of Stock and Changes in Policy, is my work under the
direction of the advisory committee and has not been submitted in any form to any
other universities. Sources of information derived or quoted from works published
and unpublished by other authors have been mentioned in the text and listed in the
References at the end of this master thesis.
I hereby transfer the copyright of my master thesis to the Bogor
Agricultural University.
Bogor, August 2015

Dessy Anggraeni
H451110141

*

Copyright transfer due to the collaborative research work with other parties outside the
Bogor Agricultural University should be based on a related agreement.

SUMMARY
DESSY ANGGRAENI. Price Volatility Analysis of Rice in Indonesia: The
Impact of Stock and Changes in Policy. Supervised by ANDRIYONO KILAT
ADHI, AMZUL RIFIN, and BERNHARD BRÜMMER.
This thesis aims to understand the volatility trend of rice price in
Indonesia, by exploring some possible drivers and test the significant impact of
those drivers to overall volatility measurement. There is overwhelming evidence
in recent literature showing that volatility has generally been lower put the
variability has been high, with the important exception of rice. The work covered
in this thesis is motivated by the complex challenges of rice price stabilization in
Indonesia. As most rural households are both producers and consumers of rice,
unbalanced proportion between rice productions and price can be harmful to

households living in poverty or at the brink of poverty. This research is a
contribution to the studies on the econometric estimation of price volatility and
impact of its policy mitigations. The estimation is based on daily rice price from
January 2006 to December 2013 with total 2086 observations, obtained from
Trade Ministry of Indonesia. Meanwhile, estimated stock data was obtained from
USDA WASDE from the same period. First, basic GARCH model is applied to
estimate rice price volatility. Then, macroeconomic variable of interest is tested
using GARCH-X with the introduction of external regressors as proxy. These
external regressors are tested both on mean model and variance model.
Formally, the research objectives are formulated as follow: 1) to
understand what is the volatility trend of rice price in Indonesia, 2) to explore
possible influential macroeconomic variables to rice price volatility measurement
in Indonesia, and 3) to test the significance impact the above variables to the
development of rice price volatility measurement in Indonesia.
The main conclusions of this thesis are as follow. First, in order to
understand the volatility trend of rice price in Indonesia, we applied empirical
method to historical rice price data. The result shows that the volatility of rice
price was driven more by its own-variance rather than external shocks.
Introduction of relevant external regressors such as stock level estimations or
dummy vector for other macroeconomic variable to GARCH-X were found to be

able to overall volatility estimation. Parameters of these external regressors were
statistically significant to better describe price observations in GARCH-X model.
However, the impacts of these external regressors are very small compare to the
impacts of own-variance or external shocks.
Second, all four microeconomics variable tested are proven to have impact
on overall volatility. This means that the introduction of relevant external
regressors such as stock level estimations, rice harvest seasonality, international
rice price or dummy vector for other macroeconomic variable to GARCH-X can
help improve the overall volatility estimation.
Third, based on the diagnostic test, we found that these variables are
significant to the model. By comparing the result of diagnostic test, it seems that a
combination of dummy matrix for food policy before and after 2009, and
estimation of stock can give better model fit. Nevertheless the other variable also
improves overall measurement.

With the above conclusions, the author can suggest the following
recommendation. In relation to price stabilization, it is important to develop an
effective and efficient rice supply chain that involves all market stakes holder in a
balanced and proportional way. This can be done by removing obstacles in each
chain of rice supply chain by improving transportation means and infrastructure,

simplify administration process, and reliable monitoring mechanisms. Due to
production uncertainties and food price volatility, government needs to be more
flexible especially in relation to food policy in order to make faster response to
any situations. It is important to hold sufficient amount of rice at storage in order
to cope up with ongoing food issues.
Keywords: price volatility, rice price, GARCH-X

RINGKASAN
DESSY ANGGRAENI. Analisis Volatilitas Harga Beras di Indonesia: Efek Stok
dan Perubahan Kebijakan. Supervised by ANDRIYONO KILAT ADHI, AMZUL
RIFIN, and BERNHARD BRÜMMER.
Studi ini bertujuan untuk memahami tren volatilitas harga beras di
Indonesia, dengan cara mempelajari faktor-faktor yang berpengaruh dan menguji
seberapa besar pengaruh dari fakrot-faktor tersebut terhadap keseluruhan estimasi
fluktuasi harga. Berdasarkan studi literatur yang banyak ditemukan akhir-akhir
ini, tingkat fluktuasi harga untuk berbagai komoditas pertanian cenderung rendah,
kecuali untuk beras. Studi ini tergerak untuk mengetahui berbagai tantangan
dalam menstabilkan harga beras di Indonesia. Sebagian besar rumah tangga di
pedesaan berperan baik sebagai produsen dan konsumen beras, dengan demikian
proporsi yang tidak seimbang antara jumlah beras yang diproduksi dengan harga

yang terbentuk di pasar akan sangat merugikan bagi rumah tangga miskin dan
berada diambang batas kemiskinan. Penelitian ini memberikan kontribusi bagi
studi tentang estimasi ekonometrika terhadap volatilitas harga dan pengaruh
kebijakan penanggulangannya. Estimasi pada studi ini dilakukan berdasarkan data
harga beras nasional mulai bulan Januari 2006 hingga Desember 2013 dengan
total 2086 observasi, di mana data tersebut diperoleh dari Departemen
Perdagangan Republik Indonesia. Sementara itu, estimasi jumlah stok beras
menggunakan data dari USA WASDE untuk periode yang sama. Pada tahap awal,
model GARCH digunakan untuk memperkirakan volatilitas harga beras.
Selanjutnya, beberapa variable makroekonomi yang diteliti dimasukkan ke dalam
model GARCH-X sebagai regresor eksternal. Seluruh regresor eksternal tersebut
diuji baik pada model rata-rata dan model varians. Hasil pendugaan menunjukkan
bahwa volatilitas harga beras lebih banyak didorong oleh variansnya sendiri
dibandingkan guncangan eksternal.
Secara formal, tujuan penelitian dapat diformulasikan sebagai berikut: 1)
memahami trend volatilitas harga beras di Indonesia, 2) mengekplorasi beberapa
variabel makroekonomi yang berpengaruh terhadap pengukuran volatilitas harga
beras di Indonesia, 3) menguji dampak signifikan variable-variable tersebut
terhadap pengukuran volatilitas harga beras di Indonesia.
Ada 3 kesimpulan utama yang bisa diambil dari hasil penelitian ini.

Pertama, volatilitas harga beras didorong lebih oleh variasinya sendiri
dibandingkan lonjakan yang berasal dari luar. Dengan mempertimbangkan
regresor eksternal yang relevan, seperti tingkat stok atau vector dummy sebagai
variable makroekonomi pada model GARCH-X, maka secara keseluruhan
estimasi model volatilitas harga beras dapat dijelaskan dengan lebih baik.
Parameter-parameter untuk regresor eksternal tersebut teruji signifikan secara
statistic dan dapat menjelaskan deret harga pada model GARCH-X. Namun,
pengaruh dari regresor eksternal tersebut terbukti sangat kecil bila dibandingkan
dengan pengaruh variasi deret harganya sendiri atau pun pengaruh luar lainnya.
Kedua, seluruh variabel yang diuji terbukti memiliki pengaruh terhadap
volatilitas. Hal in berarti bahwa, memasukkan regresor eksternal yang relevan
seperti estimasi tingkat stok, musim panen padi, harga beras internasional, dan

vector dummy untuk makroekonomi, ke dalam model GARCH-X dapat
membantu memperbaiki keseluruhan estimasi volatilitas harga.
Ketiga, berdasarkan hasil uji diagnostic, kami menemukan bahwa
variabel-variabel yang diuji berpengaruh signifikan terhadap model. Dengan
membandingkan hasil diagnostic test, maka kombinasi antara dummy matrx untuk
kebijakan pangan sebelum dan seudah tahun 2009, dengan estimasi stok dapat
memberikan kesesuaian model yang paling baik. Namun, variabel lainnya juga

terbukti memiliki pengaruh, walau pun lebih rendah, terhadap pemodelan
volatilitas harga beras.
Dengan demikian, penulis dapat merekomendasikan beberapa hal berikut.
Dalam rangka menstabilkan harga, penting sekali untuk memastikan kestabilan
rantai pasok beras, dengan demikian setiap agen dalam distribusi beras dapat
berperan secara seimbang dan sesuai proporsinya. Salah satu jalan yang dapat
ditempuh untuk mencapai hal ini adalah dengan menhilangkan segala bentuk
rintangan yang mungkin menghambat kelancaran rantai pasok, contohnya dengan
memperbaiki sarana transportasi dan infrastruktur, menyederhanakan proses
administrasi pergudangan, dan mengaplikasikan mekanisme monitoring yang
handal. Untuk mengatasi ketidakpastian dan volatilitas harga, pemerintah perlu
lebih fleksibel terutama dalam kaitannya dengan kebijakan pangan, sehingga
dapat memberikan respon yang cepat pada berbagai situasi. Pemerintah juga perlu
untuk memiliki cadangan beras yang cukup untuk menghadapi berbagai kendala
pada penyediaan bahan pangan utama.
Keywords: volatilitas harga, harga beras, GARCH-X

© All Rights Reserved by Bogor Agricultural University, 2015
Copyright Reserved
It is prohibited to quote part or all of this paper without including or citing the

source. Quotations are only for purposes of education, research, scientific
writing, preparation of reports, critics, or review an issue; and those are not
detrimental to the interest of the Bogor Agricultural University.
It is prohibited to announce and reproduce part or all of this paper in any form
without the permission of the Bogor Agricultural University.

PRICE VOLATILITY ANALYSIS OF RICE IN INDONESIA:
THE IMPACT OF STOCK AND CHANGES IN POLICY

DESSY ANGGRAENI

Master thesis
as one of the requirements to obtain a degree of
Magister Sains
in
Master of Science in Agribusiness Study Program

GRADUATE SCHOOL
BOGOR AGRICULTURAL UNIVERSITY
BOGOR

2015

Examiner Commission on Affairs of Master Thesis Examination:
Dr. Ir. Nunung Kusnadi, MS

Thesis Title

:

Name
NRP

:
:

Price Volatility Analysis of Rice in Indonesia:
The Impact of Stock and Changes in Policy
Dessy Anggraeni
H451110141

Approved by
Advisory Committee

Dr Ir Andriyono Kilat Adhi
Chairman

Dr Amzul Rifin, SP, MA
Member
Agreed by
Head of Agribusiness Study Program

Prof Dr Ir Rita Nurmalina, MS
Examination Date: July 3rd, 2015

Prof Dr Bernhard Brümmer
Member
Dean of Graduate School

Dr Ir Dahrul Syah, MScAgr
Submission Date:

ACKNOWLEDGMENT
I would like to express the deepest appreciation to my advisory committee
chair Dr. Ir. Andriyono Kilat Adhi, who has shown the attitude and the substance
of a great mentor: he continually and persuasively conveyed a spirit of adventure
in regard to research and scholarship, and an excitement in regard to teaching.
Without his supervision and constant help this dissertation would not have been
possible.
I would like to thank my committee members, Dr. Ir. Amzul Rifin, MA.,
Prof. Dr. Bernhard Brümmer and Dr. Tinoush Jamali Jaghdani, whose work
demonstrated to me that concern for farm economics theory in the context of
global and open market society, should always transcend academia and provide a
quest for our times.
In addition, a thank you to Prof. Dr. Ir. Rita Nurmalina, MS., who always
encourage me to pursue higher, and whose passion for the agribusiness had lasting
effect. I thank the Bogor Agricultural University and University of Göttingen for
providing supportive academia environment. I also thank Beasiswa Unggulan by
Planning and Cooperation of Foreign Affairs (BPKLN) Ministry of Education and
Culture of Indonesia for their financial support granted through Joint-Degree
Master Program scholarship. I also want to thank to my fellow classmates of
Magister Sains Agribisnis (MSA) in their ceaseless supports during my thesis
writing.
Bogor, August 2015
Dessy Anggraeni
H451110141

CONTENTS
CONTENT S

viii

LIST OF TABLES

ix

LIST OF FIGURES

ix

LIST OF APPENDICES

ix

LIST OF ABBREVIATIONS

x

1

INTRODUCTION
Background
Problem Statement
Research Objectives
Benefit of the Study
Scope and Limitation of the Study

1
1
3
6
7
8

2

LITERATURE REVIEW
Commodity Price Dynamics
Sources of Price Shocks
Mitigation Strategy
Storage Arbitrage
Definition, measurement, and drivers of price volatility

8
8
14
19
21
22

3

FRAMEWORK
Types and Sources of Data
Theoretical Framework
Operational Framework

27
27
27
29

4

RESEARCH METHODOLOGY
GARCH
GARCH-X

31
31
32

5

RESULT
Rice Price Volatility in Different Period
Rice Price Development 2006 – 2013
Policy Implications

34
34
35
43

6

CONCLUSION AND RECOMMENDATIONS
Conclusions
Recommendations

44
44
44

REFERENCES

46

APPENDICES

50

BIOGRAPHY

59

ix

LIST OF TABLES
1
2
3
4
5
6

Comparison of standard deviation and coefficient of variations
Rice sowing and harvest period in Indonesia
Government purchasing price 2002–2012
ADF test for unit root of rice price in levels and returns
Summary statistics of rice price in level and returns
GARCH and GARCH-X estimates

6
12
20
34
34
39

LIST OF FIGURES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Consumer Price Index in Indonesia 2006 - 2013
Domestic rice price Indonesia compare to international price 2005 - 2013
(in USD per Tonnes)
Variable supply and fix demand curve
Paddy harvesting pattern 2012 - 2013
Rice harvested area (Ha) in Indonesia in 2000–2013
Supply chain of paddy grain and rice in Karawang Region
BULOG stock mechanism to replenish and distribute rice stocks
Operational framework
Flow chart estimation process for GARCH and GARCH-X
AIC and BIC criteria for AR lag from 1-30
Domestic rice price in Indonesia 2006 - 2013
Daily return of rice price in Indonesia 2006-2013
Daily squared return of rice price in Indonesia 2006-2013
Estimation of beginning stock level of government stock 2006-2013
Comparison of actual rice price return and simulated return
Conditional variance of rice price return

4
4
9
13
13
16
26
29
30
35
36
37
37
38
42
43

LIST OF APPENDICES
1
2
3
4
5
6

The history of Indonesian food policy from 1969 to now
Summary food policy in Indonesia after 2009
ACF and PACF result of rice return
GARCH(1,1) Fit Diagnostic Result
R simulation result of GARCH-X with 2 external regressors: estimation
of stock in variance model and rice harvest season in mean model
Complete Rugarch plot

50
51
52
53
54
57

x

LIST OF ABBREVIATIONS
BKP

: Badan Ketahanan Pangan
(Indonesian Food Security Agency)
BLBI
: Bank of Indonesia Liquidity Support
BULOG : Indonesian National Food Logistic Agency. A state-owned
enterprise (BUMN) which was established through Government
Regulation (PP) No. 7/2003.
CBP
: Cadangan Beras Pemerintah
(Government Rice Reserve)
Gapoktan : Gabungan Kelompok Tani
(Farmers Association)
GKP
: Gabah Kering Panen
(dried paddy grain at harvest)
HPP
: Government purchase price
PNS
: Domestic Civil Servant
RASKIN : Rice for the poor program
Satgas
: Procurement unit under BULOG
UPGB
: Unit Pengolahan Gabah dan Beras
(Rice and Grain Processing Units)

1

INTRODUCTION
Background

Price variability in agricultural commodities is useful for both producers
and consumers because it gives rooms to maneuver for both market actors:
producers have the opportunities to reap higher profits by supplying more when
prices are high and withhold supply when prices are low; meanwhile consumers
can also be better off by taking the opposite position. The wisdom to buy cheap
and sell dear are well-known among traders for centuries. Excess supplies are then
to be stored for later use or held temporary until price rise to a desirable level in
the future. Indeed storage arbitrage could smoothen price series, but this
mechanism also generates additional positive first order autocorrelations to the
series itself. This condition makes future prices even more unpredictable, thus
more difficult to handle. In times of high price volatility, public and policy makers
tends to discredit traders and speculators for inducing larger spikes. The truth is,
the logic that drives these actions are also the one that account for storage
mechanism, only the latter is aimed to improve welfare for both producers and
consumers.
The call to set the price right, especially on important staple food such as
rice, is compelling as food insecurity still an ongoing issue to most part of
Indonesian regions. Despites significant progress in reducing overall poverty and
the number of people living bellow national poverty line, which reach over 43
percent of total populations, are still living with less than 2 USD per day.
Meanwhile food expenditures account for more than half of total per capita
household expenses. Within the poor household income group, almost 20 per cent
of this expenditure is spent on rice (Saifullah and Dawe, 2010). Sudden jump in
rice price would then translate into putting millions of vulnerable households who
live in the brink of poverty not being able meeting their daily nutritional needs.
Price gives signal to both consumers and producers either to increase
supply and/or decrease demand. As for rice, this signal carries information about
quality, such as variety, place of origin, degree of processing, etc. Every day, price
signals carry forward, via series of bargains and negotiations between buyers and
sellers, throughout the distribution channels: from farmers to grocers; until they
reach end customers. Government can intervene this process by setting a price
reference or Government Purchasing Price (GPP) in order to protect farmers from
loss when market price falls. On the other hand, this policy also affects farmers’
expectation that they should sell their paddy at higher price, which in return would
trigger traders to sell dry unhusked paddy and also rice at even a higher price. In
the end, overall market price for rice increases.
In practice, GPP faced many challenges especially due to the presence of
oligopolistic rice supply chain, where many rice distribution channels, especially
millers and wholesalers, are dominated by only several actors. Private institutions
controls around 80 percent of total rice stock domestically, meanwhile public only
20 percent. This raise challenge for BULOG, semi government institutions that

2

hold mandate to control domestic stocks, to increase its importance in stabilizing
price effectively.
Since many farmers sell their harvest in the form of wet unhusked paddy,
thus they could not enjoy the benefit whenever there is an increase in the price of
dry unhusked paddy, dry milled paddy, and rice. To farmers, a higher reference
price for dry unhusked paddy could be a boomerang for them because they are
also rice consumers. The wider the gap between paddy and rice, the smaller the
profit farmers would receive. This inefficiency may have caused due to lack of
technology owned by farmers, especially for drying paddy which still mainly
relies on the hot weather.
A stable price for rice can solve part of this problem by giving assurance
to business actors along rice supply chain to keep on investing to rice farming. As
part of price stabilization policy, government also maintain a certain level of rice
stocks in order to absorb excess supply of rice in the market and then distribute
them when supply is low. A combination of reference price at farm level and
stock management is expected to be able to stabilize rice price in the long run.
The goal to stabilize price can only happen when the policies are correctly
implemented, especially in relation to timing and bureaucracy. For example, there
are cases where rice price at market is high, but at the same time rice stock level is
high in storage. This is due to BULOG could not dispatch its stock fast because it
needs recommendations from regional state government and the Ministry of Trade
in order to conduct market operation (Lensaindonesia, 2015).
Theoretically there is a relationship between price and stock level.
Assuming that the source of price shock is exogenous or predetermined variables
on the market, thus it is possible to alter the course of price in the short to medium
run by using endogenous responses of storage mechanism. Deaton (1992) and
Wright and Williams (1991) studies provide brilliant insight about the economics
of storage as intertemporal arbitrage. Following this study, many researchers
(Cafiero et. al., 2011; Wright, 2011) also studied the relationship between
commodity price behaviors and stock level. Although it is difficult to test the
mechanism empirically, storage model implies that there are strong negative
correlations between prices and stock level. In the periods of low stocks, the
availability of stock appears to be able to mitigate the condition of higher
volatility in price series (Serra and Gil, 2012). In return, insertion of a stock
variable may open the possibility of transferring the shock in one grain market to
price of another in a later period, thus increase complexity in making inference
about what are the underlying drivers of price volatility (Bobenrieth et al., 2012).
Yet the biggest challenge lies ahead unanswered: how to bring theory into
practice.
In order to manage or anticipate adversity, experts as well as policy
makers would be interested to answer these questions: Is this the right time to
apply floor/ceiling policy? What is the optimum level? What policy instrument
should be use: market operations or import quota? What are the impacts to
producers and consumers welfare? etc. Insights about how severe the impact of
price volatility is can help policy maker to adopt better hedging strategy or even to
increased government intervention in the allocation of investment resources
(Apergis and Rezitis, 2010). However, there has not been much study for
Indonesia that focus on rice price volatility in order to explain price shocks events

3

with the existence of storage arbitrage. This thesis attempts to untangle some
elements of rice price volatility conundrum and provide reliable measurement as
reference to formulate the best price strategy that can favor both rice farmers and
consumers.
This thesis will begin by reviewing recent development commodity price
dynamics. Typical agricultural commodity prices are often characterized by an
autocorrelation series with seasonal and cyclical components. Decomposition of
these series became more challenging due to the biological nature of agricultural
productions that caused systematic behavior of price movements. The next chapter
will discuss about sources of price shocks. In the long run, price series tend to be
mean reverting. Price series may jump to a new level and stay in that level for a
relatively longer period, creating spikes (Peterson and Tomek, 2005). This
movement seems to follow a typical pattern, such that price increases in
accelerating pace in one year followed by a sharp drop in the following year
(Bobenrieth et. al., 2012; Brümmer et. al., 2013; Gouel, 2011; Huchet-bourdon,
2011). The last two chapters will discuss about some alternatives of mitigation
strategies and the logic behind storage arbitrage.
Domestic rice price volatility in Indonesia between 2006 and 2013
provides good insight about volatility structures and drivers behind it. This sample
allows us to observe how rice price volatility changes during “commodity boom”
in 2007/2008 where prices of some agricultural products jump more than double
only within a few months, and then swiftly decline to lower level. This particular
event challenged the effectiveness of existing price stabilizing, such as buffer
stock mechanism as well as other agricultural price control policies.
The methodology follows GARCH (Bollerslev, 1986) technique to model
volatility. We apply external regressors in the model to gauge the impact of some
possible drivers to volatility measurement. This allows conclusions to be drawn
about whether one or more drivers, for example stock forecasts published by
public institution or other policy, may have a big impact to influence the
magnitude of price fluctuations. By the end to understand the volatility trend of
rice price in Indonesia, explore some possible drivers and test the significance
impact those drivers to overall volatility measurement.

Problem Statement
Over the past years there has been an interesting development in food and
other agricultural commodity prices. Price movement seems to follow typical
pattern (Bobenrieth et al., 2012; Brümmer et al., 2013; Huchet-bourdon, 2011;
Prakash, 2010) such that price increasing with accelerating pace in one year
followed by sharp drop in the following year. A very distinct example of this
pattern was during “commodity boom” in 2007/2008 where prices of some
agricultural products jump more than double only within a few months, and then
swiftly decline to lower level. Grain price series also reveals such instances: short
period high volatility interspaced with longer period low volatility. Although in
the recent decades, volatility has generally been lower but the variability has been
high, with the important exception of rice (Gilbert and Morgan, 2010). This
implies higher risks for both consumers and producers. To commodity that hold

4

240.00
220.00
200.00
180.00
160.00
140.00
120.00
100.00
Jan-06

Jan-07

Jan-08

Jan-09

Jan-10

Jan-11

Jan-12

Jan-13

Year
CPI

Foods

Prepared Food

Housing

Clothing

Health

Education

Transportation

Figure 1

Consumer Price Index in Indonesia 2006 - 2013

Source: Ministry of Finance, Government of Republic Indonesia (2014)

Price

Consumer Price Index

strategic value such as rice in Indonesia, unpredictable shocks in price is highly
undesirable.
Food price index is the highest growing index compare to other sectors.
Figure 1 shows the development of Consumer Price Index in Indonesia from 2006
– 2013. In 2013 food price index grew 11.4 percent higher than previous year;
meanwhile the growth of total consumer price index is at 8.4 percent. Increasing
food price index can be traced back to New Order Era (1969-1998), when
protectionism policy started losing its ground to market liberalization. At the same
time, raising nationalism gives more favor to domestic farmers at the cost paid by
domestic consumers. As a result, domestic rice price sets to deviate from
international and stays persistently at higher level (Figure 2).
In case of rice in Indonesia, this broad price movement within a short
period is highly undesirable since it is a staple food and it is also a source of
income for the majority of people, especially those living in the rural areas. Due to
its strategic significance, rice price stabilization plays a key role in the long term
policy making process of Indonesian agriculture.

1000.00
800.00
600.00
400.00
200.00
0.00

Year
Rice, Indo med

Rice, Thai 5%

Rice, Thai A.1

Figure 2
Domestic rice price Indonesia compare to
international price 2005 - 2013 (in USD per Tonnes)
Source: World Bank Commodity Price Data and Indonesian Ministry of Trade (2014)

5

Indonesian domestic price is set to be higher than international price in
order to support domestic farmer. Interestingly, during 2007/2008 commodity
boom, Indonesian domestic rice price did not experience significant shocks. In
fact, it managed to keep price at stable level and even lower compare to
international price. Generally it is accepted that the high price spike in 2007/2008
was a result of accumulation of restrictive trade policy actions by main exporting
countries (Timmer, 2010). If there were no such actions and countries were
opening their market, rice price might still increase but not as high as the level
reached in 2007/2008 (Gouel, 2013).
Figure 2 shows a comparison of Indonesian rice price to Thailand rice
price as international benchmark. Indonesia plays an important part in the world
market for rice mainly due to its substantial rice imports for the past half century,
except for several years in the mid-1980s when self-sufficiency was temporarily
achieved. Rice import had been exceptionally important from 1975 to 1996 when
BULOG was actively committed in stabilizing domestic rice prices by keeping
domestic rice prices on the long-run trend of world prices. A study by Timmer
(2003) concludes that there were two reasons for this policy. First, the world price
represents the opportunity cost of rice to the Indonesian economy and economic
efficiency requires that domestic and world prices track each other over extended
periods of time. Second, Indonesia comprise of many small islands and located
close to several major rice exporting ports, thus it is nearly impossible for
Indonesia’s domestic rice price to be kept substantially above or below prices in
those ports for extended periods of time.
The most common way to analyze price movement is by measuring
standard deviation, which is how far away price deviate from the mean and
coefficient of variation, that is how standard deviation divided by mean. These
measurements are very useful to compare variation across time. Table 1 shows the
comparison of standard deviations and coefficient of variations between Indonesia
and Thailand. The choice of period is very critical in measuring both standard
deviation and coefficient of variations. In this study it is useful to apply cut-off
period before and after 2009. Before 2009, Indonesian food policies were giving
much attention on rice production. After commodity boom in 2007-2008, policies
that promote food consumption diversity and local produce were enacted thru
Presidential Decree No. 22/2009 (Suryana, 2007) which then become the main
program for national food security program in 2010-2014 (Food Security Agency
of Indonesia, 2011). This period happen to be aligned with reelection of
Indonesian President in 2009 who supported to reduce the impact of food crisis in
his campaign.
On overall, in the period of 2006-08 Indonesian rice price variations are
lower compare to Thailand’s, meanwhile it is the other way around in the next
period 2009-13. Although it seems like price variation become much higher for
Indonesian rice in the second period, actually the application of standard deviation
and coefficient of variation cannot accommodate curvature which represent steady
price increase in 2009-2011 then followed by steady price decrease from 2011
onwards. This is always the challenge in measuring price volatility that is to know
what variability band that still consider reasonable. Prices that falls outside this
band can then be considered as excessive variability and if the distance is really
far from the mean, then such movement can be categorized as extreme variability.

6

Table 1

Comparison of standard deviation and coefficient of variations

Type of rice
Rice, Indonesia medium

Period 2006-08
SD

Period 2009-13

CV

SD

CV

50,83

0,10

119,78

0,16

Rice, Thai A.1

184,33

0,43

48,92

0,09

Rice, Thai 5%

145,19

0,45

80,73

0,19

Source: Rice, Indonesia medium: Ministry of Trade Government of Indonesia (2014); Rice, Thai
A.1 Rice, Thai 5%: Thompson Reuters Datastream (2014), calculated

From food security point of view, it is important to understand how much
price variation that is still safe for the vulnerable to afford sufficient foods and
nutrients. Indonesia has more than 32 million populations who were still living on
less than US$2 per day. In this setting, even a small price hike in food price may
have big impact to households especially when their budget for food constitutes
up to 48 percent of their total budget spending (Central Bureau of Statistics of
Indonesia, 2013). Raising food price issues seems to be inevitable throughout the
years, especially a month before and after Ramadhan1. In the past 3 years, prices
were higher not only in level but also in volatility despites increasing efforts from
the government to improve distribution, speed up import process for importers
who already hold authorization to proceed, and also allocate additional import
volume2. To handle this issue, Indonesian government reserves a certain amount
of rice as public reserve. Some of the goals are to support Rice for the Poor
Program (Raskin) and anticipate rice price volatility.
From market perspective, maintaining storage to store excess supply has
been known to have effect in stabilizing price (Cafiero et. al., 2011; Serra and Gil,
2012; Wright, 2011). The stock also plays important role in price stabilization
policy in Indonesia. The level of stock is a good reference to decide what kind of
policy needed, when it should be implemented, and how intense it is going to be
implemented at national as well as regional level, especially when it is related to
import. However, this ideal condition is hard to achieved due to accurate data on
stock level is rarely available (Resnia et. al., 2012). Thus the ideal stock buffering
mechanism is very difficult to be achieved in the case of Indonesia. Considering
government policies related to food price, other factors seem to play role in
stabilizing rice price volatility in Indonesia.

Research Objectives
Formally, the research objectives are formulated as follow:
1. To analyze rice price volatility in Indonesia between 2006 to 2013.
2. To identify factors which affect rice price volatility in Indonesia.
1

Ramadhan is one of the most celebrated a Muslim holiday in Indonesia that marks the end of fasting month
As stated by the Head of Indonesian Food Security Council (BadanKetahananPangan), Prof. Dr. Ir.
AchmadSuryana, M.S. in Ministry of Agriculture official website: http://bkp.deptan.go.id/berita-192-hargadan-cadangan-pangan.html
2

7

3. To find suitable alternative forecasting model in order to examine rice
price volatility phenomenon in Indonesia.
The fundamental thesis of this study is derived from macroeconomic
approach to deal with price volatility, which are either by directly stabilize the
market price using price controls mechanism, stock mechanism, or a combination
of both. The methodology follows GARCH (Bollerslev, 1986) and GARCH-X
(Apergis and Rezitis, 2010) model to measure volatility. This thesis applied 4
external regressors in the model to gauge the importance of these variables to rice
price volatility development, which are: (1) the estimations of stock level, (2) rice
harvest seasonality, (3) the impact of Thailand rice as a benchmark for
international rice price, and (4) other important macroeconomic variables related
to new direction on food policy. This allows conclusions to be drawn about
whether certain macroeconomic variables have influential predictive capacity to
the conditional variance of rice prices.

Benefit of the Study
There are 4 main problems of food in Indonesia, as presented at the Jakarta
Food Summit 2014, which encompass some limitations in the area of production
and processing, distributions and stocks, trade and marketing, volatile food price
faced by end consumers. Although the summit were held recently, but the
problem discussed has been around for decades. Interestingly, these are also the
same problems faced by many developing countries, yet not many are able to
solve it thoroughly. Indonesia is among one of a few countries who successfully
manage steady economic growth during the past downturn, but food price index
are still the biggest contributors to the country national inflations.
This study will help us to understand the development of rice price in
Indonesia from the point of view of stochastic analysis, such as trends, variability
around the trends, and extreme movements. These analyses are aimed to help
policy makers to make better judgment about what are the real sources of rice
price volatility and the best way mitigate it. Developing countries often suffer
from ineffective and long supply chain that resulted in the lack the ability to take
fast action in the event of sudden jump in prices. Several main actions such as
removing obstacles in each chain of rice supply chain by improving transportation
means and infrastructure, simplify administration process, and reliable monitoring
mechanisms; are need to be done which signify the need of reliable measurement
of price trend and volatility development.
Another issue is related to production uncertainties and food price
volatility, which come as part of the characteristics of agricultural prices. This has
been an ongoing problem faced by governments and all market stakes holders.
Government need to be more flexible especially in relation to food policy in order
to make faster response to any situations. Such responses may involve holding
sufficient amount of rice at storage in order to cope up with ongoing food issues.
This research will give benefit to such issue by exploring possible influential
macroeconomic variables to rice price volatility measurement in Indonesia and
testing the significance impact the above variables to the development of rice
price volatility measurement in Indonesia. By having a clear measurements and

8

standardized method, this research also contributes to clear up some confusions
and debates around rice price movements in Indonesia.

Scope and Limitation of the Study
This research uses average national rice price data from the Ministry of
Trade Republic of Indonesia from 2006 to 2013. The analysis in this thesis is
limited to national level and cannot captured movements in the provincial level or
lower. When at some periods there are price jump in some provinces and price
sunk in other province, it is possible that the average price data do not expressed
such movements.
The choices of certain macroeconomic variables are based on those which
have significant predictive capacity on the volatility of rice prices, in this case (1)
the estimations of stock level, (2) rice harvest seasonality, (3) the impact of
Thailand rice as a benchmark for international rice price, and (4) other important
macroeconomic variables related to new direction on food policy. Estimation of
stock level is measured using USDA WASDE data, which are not the real stock
level available. It is nearly impossible to get such data, however such
approximations are sufficient enough to use as forecast or trading plan. Rice
harvest seasonality is expressed as dummy variable 0 and 1 following the rice
growing and harvest period published by Indonesian Center for Food and Crops
Research and Development. This is also an approximation and can be improved
with real time climatic data or harvest period. Thailand rice data are taken from
Thompson Reuters Datastream for the same period. Thailand is a major rice
played in the world, however the world rice market may get impacted by other
countries other than Thailand, which do not covered in this thesis. Other
macroeconomic variable is expressed as dummy 0 and 1 to represent food policy
applied in the period before and after 2009.

2

LITERATURE REVIEW
Commodity Price Dynamics

Prices contain information on supply and demand forces that works in a
market. In the case of storable agricultural goods, such as grains, demand for
storage is also considered while calculating total demand. In theory, insertion of
storage into supply and demand model could help stabilized price innovation. The
model thus, assumed demand as relatively fix, meanwhile supply tend to varied
depending on how successful harvest during the year. Figure 3 shows a general
illustration of this model.
In the case of bad harvest where rice available in scarce quantity (S1),
price is expected to rise. When price rises above ceiling price, ideally public
storage should sell some of its stocks, increasing market supply, in order to bring
down price as low as targeted ceiling price. The other way around works for the

9

case of good harvest where rice available in abundant quantity (S2). Through this
mechanism, storage is expected to stabilize price, however in real world situation
things may develop in a more complex ways. For example, sufficient amount of
budget need to be available continuously so that public storage can execute stock
purchase at right market timing. Storage can function as buffer for price only
when there is enough quantity in storage. When there is not enough quantity in
storage, available stocks would go to those who could catch up with rising market
price. As consequence, poorest consumers have to reduce their calorie
consumption, increasing their risk of malnutrition and hunger.

Figure 3

Variable supply and fix demand curve
Source: Economics Online (2014)

The relationship between price and stock can be explained structurally via
storage model. Agricultural commodity prices have systematic behavior due to the
biological nature of production, even if their markets are efficient. Typically,
agricultural price series are auto-correlated with seasonal and perhaps cyclical
components. Due to seasonality and cycles, these prices may be mean reverting to
some long-run average. Occasionally, prices jump abruptly and temporarily to a
high level relative to their long-run average, creating spikes.
Based on the above theory, this thesis considers storage as one of variable
that have impact on rice price development. Moreover, the impact of rice harvest
seasonality, Thailand rice as a benchmark for international rice price, and other
important macroeconomic variables related to new direction on food policy will
also use to measure overall price volatility.
The modern theory of competitive storage is developed by Williams and
Wright (1991). This model assumes supply as the sum of independently and
identically distributed (i.i.d.) harvest and carry-over stocks from previous period;

10

consumption as the difference between available supply and the stocks carried
out, and non-negative stocks. Storage arbitrage works by following either two of
these conditions: (1) when the total price and marginal cost of storage per period
is higher than expected future price, then storer would increase its carry-over
stocks; (2) when the total price and marginal cost of storage per period is equal or
lower than expected future price, then storer would not increase its carry-over
stocks.
Empirical proof of this model is difficult due to absence of satisfactory
time series of aggregate production and stocks for major commodities;
nevertheless the model is useful in estimating and testing of a model of price
formation where supply and demand shocks are mediated through storage. Deaton
and Laroque (1996) applied generalized method of moments (GMM) to identify
and estimate a subset of the parameters of storage model in order to construct
comprehensive comparison between the model and actual commodity price data.
Moreover, their study also pursues the explanation about sources of high
autocorrelation in commodity prices series. They found that if autocorrelation
generated by supply shocks only then the resulting autocorrelation should not be
as high as those observed in actual data. Even with the additional speculative
behavior of storage, i.i.d. prices still could not be filtered into the high level of
observed auto-correlated prices. Although it is unclear how commodity prices
exhibit such high autocorrelation; the study reveals excellently the underlying
mechanism behind profit-maximizing and risk-neutral speculator’s acting on i.i.d.
driven process and how this process in the end modify the original dependency of
price series to supply and demand.
Stock arbitration can only work when there is difference in current price
and expected price in the future. In other word, low price do not guarantee traders
to increase stock, or vice versa. Future price need to be higher than the total of
current price and the marginal cost of storage per period. The decision to buy or
sell is crucial for public storage since purchasing budget is generally limited. The
fact that commodity prices are highly auto-correlated exacerbates the matter as
stocks must be held at sufficient amount as well as making profit (or at least not
making loss) when hold them.
From budget perspective, storage may not be the best solution to stabilize
prices. Nevertheless, storage can be used to reduce costs of varying production
(when marginal cost is increasing), and to reduce marketing costs by facilitating
production and delivery scheduling and avoiding stockouts. Equilibrium inventory
behavior is the solution to a stochastic dynamic optimization problem. Pyndick
(2002) studies the short-run dynamics of commodity prices and inventories,
focusing on the behavior and role of volatility. He explained two principal ways
of how volatility affects prices, production, and inventories: (1) it directly affects
the marginal value of storage (the marginal convenience yield), i.e., the flow of
benefits from an extra unit of inventory; and (2) for a depletable resource like oil,
volatility affects the total marginal cost of production via the “option premium.”
His paper also shows how inventories adjust and affect prices in the short run.
Policy framework give impact to significantly improve price stability in
regulated market compared to unregulated one by providing a number of market
support measures. Increased volatility in commodity prices suggested that the
challenges associated with high levels of price volatility need to be addressed as a

11

priority. The case of Skim Milk Powder (SMP) policy reforms in European Union
(EU) present interesting empirical issues in quantifying increase of price
volatility. SMP market in EU are more regulated compared to world market, thus
they were less volatile as they are protected to the same degree from local and
global shocks (O’Connor and Keane, 2011). The study showed that the EU SMP
series is particularly well-modeled as an ARMA process as it displays normal
errors which are free from autocorrelation and ARCH. EU SMP market
experienced period of exceptional high levels of volatility in both world and
European prices in the long-term historical context. Thus, it is also reasonable to
assume that alternative specifications of these models other than basic
GARCH(1,1) model such as TGARCH (Threshold GARCH), AGARCH
(Asymmetric GARCH) may be more appropriate. These models are supposed to
be able to better captured the effects of government intervention to price such as
floor and ceiling price also government purchase to build up stocks which can
delay price recovery in world markets. If other trade restrictions exist, such as
export tax, the responses to prices dynamics can be explicitly captured by more
advanced GARCH model.
The driving forces behind commodity price volatility are traced back to
basic model of supply and demand. Such models are useful to understand the
causes of food price formation. However, special care is needed in order to
understand rice price formation. Supply and demand model is useful as starting
point but it is not enough to explain hoarding behavior and its subsequent impacts
on prices. Market structure plays important role in explaining rice price volatility
development in the short run and long run.
For this reason, Timmer (2009) use supply storage model to incorporate
the role of outside speculators. Furthermore he also quantified the impact of
financial factors and actors on commodity price formation using very short-run
prices and Granger causality analysis for a wide range of financial and commodity
markets, including rice. For rice, wheat, and corns, the study found that their price
formations are impacted by some exogenous variables such as weather shocks or
legislated mandates for biofuel usage; and also endogenous variables, such as
responses of producers and consumers to prices themselves, perhaps even policy
responses of governments to prices. For rice, one of the most prominent variables
to be included in the model is the insertion of export bans as a way to prevent
domestic food price inflation.
The challenge for rice is that it has been so volatile. The coefficient of
variation of world rice prices has often been doubles that of wheat or corn for
decades at a time. Understanding this volatility has been difficult because much of
it traces to the residual nature of the world rice market, as both importing and
exporting countries stabilize rice prices internally by using the world rice market
to dispose of surpluses or to meet deficits via imports. Thus supply and demand in
the world market are a direct result of political decisions in a number of Asian
countries. Rice is a very political commodity (Timmer and Falcon, 1975).
Another major contributor to the price dynamics are seasonality variables.
Most of rice producers in Indonesia are small farmers with very small land
ownership, around 0.3 Ha in average and scattered all over the land. They are still
using relatively simple rice cultivation and post-harvest techniques thus
sometimes the losses in post-harvest are relatively high.

12

Table 2 shows a compilation of growing season by Indonesian Center for
Food Crops Research and Development. On the average, rice can be harvested
four months after it is cultivated. Harvest period usually last until one month, the
same length as growing period. Maximu